[PATCHSET v5 0/12] Add support for async buffered reads

[PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

We technically support this already through io_uring, but it's
implemented with a thread backend to support cases where we would
block. This isn't ideal.

After a few prep patches, the core of this patchset is adding support
for async callbacks on page unlock. With this primitive, we can simply
retry the IO operation. With io_uring, this works a lot like poll based
retry for files that support it. If a page is currently locked and
needed, -EIOCBQUEUED is returned with a callback armed. The callers
callback is responsible for restarting the operation.

With this callback primitive, we can add support for
generic_file_buffered_read(), which is what most file systems end up
using for buffered reads. XFS/ext4/btrfs/bdev is wired up, but probably
trivial to add more.

The file flags support for this by setting FMODE_BUF_RASYNC, similar
to what we do for FMODE_NOWAIT. Open to suggestions here if this is
the preferred method or not.

In terms of results, I wrote a small test app that randomly reads 4G
of data in 4K chunks from a file hosted by ext4. The app uses a queue
depth of 32. If you want to test yourself, you can just use buffered=1
with ioengine=io_uring with fio. No application changes are needed to
use the more optimized buffered async read.

preadv for comparison:
	real    1m13.821s
	user    0m0.558s
	sys     0m11.125s
	CPU	~13%

Mainline:
	real    0m12.054s
	user    0m0.111s
	sys     0m5.659s
	CPU	~32% + ~50% == ~82%

This patchset:
	real    0m9.283s
	user    0m0.147s
	sys     0m4.619s
	CPU	~52%

The CPU numbers are just a rough estimate. For the mainline io_uring
run, this includes the app itself and all the threads doing IO on its
behalf (32% for the app, ~1.6% per worker and 32 of them). Context
switch rate is much smaller with the patchset, since we only have the
one task performing IO.

Also ran a simple fio based test case, varying the queue depth from 1
to 16, doubling every time:

[buf-test]
filename=/data/file
direct=0
ioengine=io_uring
norandommap
rw=randread
bs=4k
iodepth=${QD}
randseed=89
runtime=10s

QD/Test		Patchset IOPS		Mainline IOPS
1		9046			8294
2		19.8k			18.9k
4		39.2k			28.5k
8		64.4k			31.4k
16		65.7k			37.8k

Outside of my usual environment, so this is just running on a virtualized
NVMe device in qemu, using ext4 as the file system. NVMe isn't very
efficient virtualized, so we run out of steam at ~65K which is why we
flatline on the patched side (nvme_submit_cmd() eats ~75% of the test app
CPU). Before that happens, it's a linear increase. Not shown is context
switch rate, which is massively lower with the new code. The old thread
offload adds a blocking thread per pending IO, so context rate quickly
goes through the roof.

The goal here is efficiency. Async thread offload adds latency, and
it also adds noticable overhead on items such as adding pages to the
page cache. By allowing proper async buffered read support, we don't
have X threads hammering on the same inode page cache, we have just
the single app actually doing IO.

Been beating on this and it's solid for me, and I'm now pretty happy
with how it all turned out. Not aware of any missing bits/pieces or
code cleanups that need doing.

Series can also be found here:

https://git.kernel.dk/cgit/linux-block/log/?h=async-buffered.5

or pull from:

git://git.kernel.dk/linux-block async-buffered.5

 fs/block_dev.c            |   2 +-
 fs/btrfs/file.c           |   2 +-
 fs/ext4/file.c            |   2 +-
 fs/io_uring.c             | 130 ++++++++++++++++++++++++++++++++++++--
 fs/xfs/xfs_file.c         |   2 +-
 include/linux/blk_types.h |   3 +-
 include/linux/fs.h        |  10 ++-
 include/linux/pagemap.h   |  67 ++++++++++++++++++++
 mm/filemap.c              | 111 ++++++++++++++++++++------------
 9 files changed, 279 insertions(+), 50 deletions(-)

Changes since v5:
- Correct commit message, iocb->private -> iocb->ki_waitq
- Get rid of io_uring goto, use an iter read helper
Changes since v3:
- io_uring: don't retry if REQ_F_NOWAIT is set
- io_uring: alloc req->io if the request type didn't already
- Add iocb->ki_waitq instead of (ab)using iocb->private
Changes since v2:
- Get rid of unnecessary wait_page_async struct, just use wait_page_async
- Add another prep handler, adding wake_page_match()
- Use wake_page_match() in both callers
Changes since v1:
- Fix an issue with inline page locking
- Fix a potential race with __wait_on_page_locked_async()
- Fix a hang related to not setting page_match, thus missing a wakeup

-- 
Jens Axboe

[PATCH 01/12] block: read-ahead submission should imply no-wait as well

[Jens Axboe]

As read-ahead is opportunistic, don't block for request allocation.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 include/linux/blk_types.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index ccb895f911b1..c296463c15eb 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -374,7 +374,8 @@ enum req_flag_bits {
 #define REQ_INTEGRITY		(1ULL << __REQ_INTEGRITY)
 #define REQ_FUA			(1ULL << __REQ_FUA)
 #define REQ_PREFLUSH		(1ULL << __REQ_PREFLUSH)
-#define REQ_RAHEAD		(1ULL << __REQ_RAHEAD)
+#define REQ_RAHEAD		\
+	((1ULL << __REQ_RAHEAD) | (1ULL << __REQ_NOWAIT))
 #define REQ_BACKGROUND		(1ULL << __REQ_BACKGROUND)
 #define REQ_NOWAIT		(1ULL << __REQ_NOWAIT)
 #define REQ_CGROUP_PUNT		(1ULL << __REQ_CGROUP_PUNT)
-- 
2.26.2

[PATCH 02/12] mm: allow read-ahead with IOCB_NOWAIT set

[Jens Axboe]

The read-ahead shouldn't block, so allow it to be done even if
IOCB_NOWAIT is set in the kiocb.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 mm/filemap.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/mm/filemap.c b/mm/filemap.c
index 23a051a7ef0f..80747f1377d5 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -2031,8 +2031,6 @@ static ssize_t generic_file_buffered_read(struct kiocb *iocb,
 
 		page = find_get_page(mapping, index);
 		if (!page) {
-			if (iocb->ki_flags & IOCB_NOWAIT)
-				goto would_block;
 			page_cache_sync_readahead(mapping,
 					ra, filp,
 					index, last_index - index);
-- 
2.26.2

Re: [PATCH 02/12] mm: allow read-ahead with IOCB_NOWAIT set

[Johannes Weiner]

On Tue, May 26, 2020 at 01:51:13PM -0600, Jens Axboe wrote:
> The read-ahead shouldn't block, so allow it to be done even if
> IOCB_NOWAIT is set in the kiocb.
> 
> Signed-off-by: Jens Axboe <axboe@kernel.dk>

Acked-by: Johannes Weiner <hannes@cmpxchg.org>

Looks reasonable. Especially after patch 1 - although it seems that
even before that, IOCB_NOWAIT could have occasionally ended up in
page_cache_async_readahead(), which isn't too different from the sync
variant except for the range calculations, and may have blocked on IO
submission in the past.

[PATCH 03/12] mm: abstract out wake_page_match() from wake_page_function()

[Jens Axboe]

No functional changes in this patch, just in preparation for allowing
more callers.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 include/linux/pagemap.h | 37 +++++++++++++++++++++++++++++++++++++
 mm/filemap.c            | 35 ++++-------------------------------
 2 files changed, 41 insertions(+), 31 deletions(-)

diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
index a8f7bd8ea1c6..53d980f2208d 100644
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -456,6 +456,43 @@ static inline pgoff_t linear_page_index(struct vm_area_struct *vma,
 	return pgoff;
 }
 
+/* This has the same layout as wait_bit_key - see fs/cachefiles/rdwr.c */
+struct wait_page_key {
+	struct page *page;
+	int bit_nr;
+	int page_match;
+};
+
+struct wait_page_queue {
+	struct page *page;
+	int bit_nr;
+	wait_queue_entry_t wait;
+};
+
+static inline int wake_page_match(struct wait_page_queue *wait_page,
+				  struct wait_page_key *key)
+{
+	if (wait_page->page != key->page)
+	       return 0;
+	key->page_match = 1;
+
+	if (wait_page->bit_nr != key->bit_nr)
+		return 0;
+
+	/*
+	 * Stop walking if it's locked.
+	 * Is this safe if put_and_wait_on_page_locked() is in use?
+	 * Yes: the waker must hold a reference to this page, and if PG_locked
+	 * has now already been set by another task, that task must also hold
+	 * a reference to the *same usage* of this page; so there is no need
+	 * to walk on to wake even the put_and_wait_on_page_locked() callers.
+	 */
+	if (test_bit(key->bit_nr, &key->page->flags))
+		return -1;
+
+	return 1;
+}
+
 extern void __lock_page(struct page *page);
 extern int __lock_page_killable(struct page *page);
 extern int __lock_page_or_retry(struct page *page, struct mm_struct *mm,
diff --git a/mm/filemap.c b/mm/filemap.c
index 80747f1377d5..e891b5bee8fd 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -990,43 +990,16 @@ void __init pagecache_init(void)
 	page_writeback_init();
 }
 
-/* This has the same layout as wait_bit_key - see fs/cachefiles/rdwr.c */
-struct wait_page_key {
-	struct page *page;
-	int bit_nr;
-	int page_match;
-};
-
-struct wait_page_queue {
-	struct page *page;
-	int bit_nr;
-	wait_queue_entry_t wait;
-};
-
 static int wake_page_function(wait_queue_entry_t *wait, unsigned mode, int sync, void *arg)
 {
 	struct wait_page_key *key = arg;
 	struct wait_page_queue *wait_page
 		= container_of(wait, struct wait_page_queue, wait);
+	int ret;
 
-	if (wait_page->page != key->page)
-	       return 0;
-	key->page_match = 1;
-
-	if (wait_page->bit_nr != key->bit_nr)
-		return 0;
-
-	/*
-	 * Stop walking if it's locked.
-	 * Is this safe if put_and_wait_on_page_locked() is in use?
-	 * Yes: the waker must hold a reference to this page, and if PG_locked
-	 * has now already been set by another task, that task must also hold
-	 * a reference to the *same usage* of this page; so there is no need
-	 * to walk on to wake even the put_and_wait_on_page_locked() callers.
-	 */
-	if (test_bit(key->bit_nr, &key->page->flags))
-		return -1;
-
+	ret = wake_page_match(wait_page, key);
+	if (ret != 1)
+		return ret;
 	return autoremove_wake_function(wait, mode, sync, key);
 }
 
-- 
2.26.2

Re: [PATCH 03/12] mm: abstract out wake_page_match() from wake_page_function()

[Johannes Weiner]

On Tue, May 26, 2020 at 01:51:14PM -0600, Jens Axboe wrote:
> No functional changes in this patch, just in preparation for allowing
> more callers.
> 
> Signed-off-by: Jens Axboe <axboe@kernel.dk>

Acked-by: Johannes Weiner <hannes@cmpxchg.org>

Re: [PATCH 03/12] mm: abstract out wake_page_match() from wake_page_function()

[Matthew Wilcox]

On Tue, May 26, 2020 at 01:51:14PM -0600, Jens Axboe wrote:
> +++ b/include/linux/pagemap.h
> @@ -456,6 +456,43 @@ static inline pgoff_t linear_page_index(struct vm_area_struct *vma,
>  	return pgoff;
>  }
>  
> +/* This has the same layout as wait_bit_key - see fs/cachefiles/rdwr.c */
> +struct wait_page_key {
> +	struct page *page;
> +	int bit_nr;
> +	int page_match;
> +};

I know you only moved the struct, and the comment along with it, but now
that this struct is in pagemap.h, I think cachefiles should be updated
to use wait_page_key instead of wait_bit_key.  Dave?

[PATCH 04/12] mm: add support for async page locking

[Jens Axboe]

Normally waiting for a page to become unlocked, or locking the page,
requires waiting for IO to complete. Add support for lock_page_async()
and wait_on_page_locked_async(), which are callback based instead. This
allows a caller to get notified when a page becomes unlocked, rather
than wait for it.

We add a new iocb field, ki_waitq, to pass in the necessary data for this
to happen. We can unionize this with ki_cookie, since that is only used
for polled IO. Polled IO can never co-exist with async callbacks, as it is
(by definition) polled completions. struct wait_page_key is made public,
and we define struct wait_page_async as the interface between the caller
and the core.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 include/linux/fs.h      |  7 ++++++-
 include/linux/pagemap.h |  9 +++++++++
 mm/filemap.c            | 41 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 56 insertions(+), 1 deletion(-)

diff --git a/include/linux/fs.h b/include/linux/fs.h
index d3ebb49189df..ba1fff0e7bca 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -314,6 +314,8 @@ enum rw_hint {
 #define IOCB_SYNC		(1 << 5)
 #define IOCB_WRITE		(1 << 6)
 #define IOCB_NOWAIT		(1 << 7)
+/* iocb->ki_waitq is valid */
+#define IOCB_WAITQ		(1 << 8)
 
 struct kiocb {
 	struct file		*ki_filp;
@@ -327,7 +329,10 @@ struct kiocb {
 	int			ki_flags;
 	u16			ki_hint;
 	u16			ki_ioprio; /* See linux/ioprio.h */
-	unsigned int		ki_cookie; /* for ->iopoll */
+	union {
+		unsigned int		ki_cookie; /* for ->iopoll */
+		struct wait_page_queue	*ki_waitq; /* for async buffered IO */
+	};
 
 	randomized_struct_fields_end
 };
diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
index 53d980f2208d..d3e63c9c61ae 100644
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -495,6 +495,7 @@ static inline int wake_page_match(struct wait_page_queue *wait_page,
 
 extern void __lock_page(struct page *page);
 extern int __lock_page_killable(struct page *page);
+extern int __lock_page_async(struct page *page, struct wait_page_queue *wait);
 extern int __lock_page_or_retry(struct page *page, struct mm_struct *mm,
 				unsigned int flags);
 extern void unlock_page(struct page *page);
@@ -531,6 +532,14 @@ static inline int lock_page_killable(struct page *page)
 	return 0;
 }
 
+static inline int lock_page_async(struct page *page,
+				  struct wait_page_queue *wait)
+{
+	if (!trylock_page(page))
+		return __lock_page_async(page, wait);
+	return 0;
+}
+
 /*
  * lock_page_or_retry - Lock the page, unless this would block and the
  * caller indicated that it can handle a retry.
diff --git a/mm/filemap.c b/mm/filemap.c
index e891b5bee8fd..c746541b1d49 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1183,6 +1183,42 @@ int wait_on_page_bit_killable(struct page *page, int bit_nr)
 }
 EXPORT_SYMBOL(wait_on_page_bit_killable);
 
+static int __wait_on_page_locked_async(struct page *page,
+				       struct wait_page_queue *wait, bool set)
+{
+	struct wait_queue_head *q = page_waitqueue(page);
+	int ret = 0;
+
+	wait->page = page;
+	wait->bit_nr = PG_locked;
+
+	spin_lock_irq(&q->lock);
+	if (set)
+		ret = !trylock_page(page);
+	else
+		ret = PageLocked(page);
+	if (ret) {
+		__add_wait_queue_entry_tail(q, &wait->wait);
+		SetPageWaiters(page);
+		if (set)
+			ret = !trylock_page(page);
+		else
+			ret = PageLocked(page);
+		/*
+		 * If we were succesful now, we know we're still on the
+		 * waitqueue as we're still under the lock. This means it's
+		 * safe to remove and return success, we know the callback
+		 * isn't going to trigger.
+		 */
+		if (!ret)
+			__remove_wait_queue(q, &wait->wait);
+		else
+			ret = -EIOCBQUEUED;
+	}
+	spin_unlock_irq(&q->lock);
+	return ret;
+}
+
 /**
  * put_and_wait_on_page_locked - Drop a reference and wait for it to be unlocked
  * @page: The page to wait for.
@@ -1345,6 +1381,11 @@ int __lock_page_killable(struct page *__page)
 }
 EXPORT_SYMBOL_GPL(__lock_page_killable);
 
+int __lock_page_async(struct page *page, struct wait_page_queue *wait)
+{
+	return __wait_on_page_locked_async(page, wait, true);
+}
+
 /*
  * Return values:
  * 1 - page is locked; mmap_sem is still held.
-- 
2.26.2

Re: [PATCH 04/12] mm: add support for async page locking

[Johannes Weiner]

On Tue, May 26, 2020 at 01:51:15PM -0600, Jens Axboe wrote:
> Normally waiting for a page to become unlocked, or locking the page,
> requires waiting for IO to complete. Add support for lock_page_async()
> and wait_on_page_locked_async(), which are callback based instead. This

wait_on_page_locked_async() is actually in the next patch, requiring
some back and forth to review. I wonder if this and the next patch
could be merged to have the new API and callers introduced together?

> allows a caller to get notified when a page becomes unlocked, rather
> than wait for it.
> 
> We add a new iocb field, ki_waitq, to pass in the necessary data for this
> to happen. We can unionize this with ki_cookie, since that is only used
> for polled IO. Polled IO can never co-exist with async callbacks, as it is
> (by definition) polled completions. struct wait_page_key is made public,
> and we define struct wait_page_async as the interface between the caller
> and the core.
> 
> Signed-off-by: Jens Axboe <axboe@kernel.dk>

Acked-by: Johannes Weiner <hannes@cmpxchg.org>

Re: [PATCH 04/12] mm: add support for async page locking

[Jens Axboe]

On 5/26/20 3:59 PM, Johannes Weiner wrote:
> On Tue, May 26, 2020 at 01:51:15PM -0600, Jens Axboe wrote:
>> Normally waiting for a page to become unlocked, or locking the page,
>> requires waiting for IO to complete. Add support for lock_page_async()
>> and wait_on_page_locked_async(), which are callback based instead. This
> 
> wait_on_page_locked_async() is actually in the next patch, requiring
> some back and forth to review. I wonder if this and the next patch
> could be merged to have the new API and callers introduced together?

I'm fine with that, if that is preferable. Don't feel strongly about
that at all, just tried to do it as piecemeal as possible to make
it easier to review.

-- 
Jens Axboe

Re: [PATCH 04/12] mm: add support for async page locking

[Johannes Weiner]

On Tue, May 26, 2020 at 04:01:07PM -0600, Jens Axboe wrote:
> On 5/26/20 3:59 PM, Johannes Weiner wrote:
> > On Tue, May 26, 2020 at 01:51:15PM -0600, Jens Axboe wrote:
> >> Normally waiting for a page to become unlocked, or locking the page,
> >> requires waiting for IO to complete. Add support for lock_page_async()
> >> and wait_on_page_locked_async(), which are callback based instead. This
> > 
> > wait_on_page_locked_async() is actually in the next patch, requiring
> > some back and forth to review. I wonder if this and the next patch
> > could be merged to have the new API and callers introduced together?
> 
> I'm fine with that, if that is preferable. Don't feel strongly about
> that at all, just tried to do it as piecemeal as possible to make
> it easier to review.

Not worth sending a new iteration over, IMO.

Re: [PATCH 04/12] mm: add support for async page locking

[Matthew Wilcox]

On Tue, May 26, 2020 at 01:51:15PM -0600, Jens Axboe wrote:
> +static int __wait_on_page_locked_async(struct page *page,
> +				       struct wait_page_queue *wait, bool set)
> +{
> +	struct wait_queue_head *q = page_waitqueue(page);
> +	int ret = 0;
> +
> +	wait->page = page;
> +	wait->bit_nr = PG_locked;
> +
> +	spin_lock_irq(&q->lock);
> +	if (set)
> +		ret = !trylock_page(page);
> +	else
> +		ret = PageLocked(page);
> +	if (ret) {
> +		__add_wait_queue_entry_tail(q, &wait->wait);
> +		SetPageWaiters(page);
> +		if (set)
> +			ret = !trylock_page(page);
> +		else
> +			ret = PageLocked(page);

Between the callers and this function, we actually look at PG_lock three
times; once in the caller, then after taking the spinlock, then after
adding ourselves to the waitqueue.  I understand the first and third, but
is it really worth doing the second test?  It feels unlikely to succeed
and only saves us setting PageWaiters.

Re: [PATCH 04/12] mm: add support for async page locking

[Jens Axboe]

On 6/1/20 8:26 AM, Matthew Wilcox wrote:
> On Tue, May 26, 2020 at 01:51:15PM -0600, Jens Axboe wrote:
>> +static int __wait_on_page_locked_async(struct page *page,
>> +				       struct wait_page_queue *wait, bool set)
>> +{
>> +	struct wait_queue_head *q = page_waitqueue(page);
>> +	int ret = 0;
>> +
>> +	wait->page = page;
>> +	wait->bit_nr = PG_locked;
>> +
>> +	spin_lock_irq(&q->lock);
>> +	if (set)
>> +		ret = !trylock_page(page);
>> +	else
>> +		ret = PageLocked(page);
>> +	if (ret) {
>> +		__add_wait_queue_entry_tail(q, &wait->wait);
>> +		SetPageWaiters(page);
>> +		if (set)
>> +			ret = !trylock_page(page);
>> +		else
>> +			ret = PageLocked(page);
> 
> Between the callers and this function, we actually look at PG_lock three
> times; once in the caller, then after taking the spinlock, then after
> adding ourselves to the waitqueue.  I understand the first and third, but
> is it really worth doing the second test?  It feels unlikely to succeed
> and only saves us setting PageWaiters.

That's probably true, and we can skip the 2nd one. I'll make the change.

-- 
Jens Axboe

[PATCH 05/12] mm: support async buffered reads in generic_file_buffered_read()

[Jens Axboe]

Use the async page locking infrastructure, if IOCB_WAITQ is set in the
passed in iocb. The caller must expect an -EIOCBQUEUED return value,
which means that IO is started but not done yet. This is similar to how
O_DIRECT signals the same operation. Once the callback is received by
the caller for IO completion, the caller must retry the operation.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 mm/filemap.c | 33 ++++++++++++++++++++++++++-------
 1 file changed, 26 insertions(+), 7 deletions(-)

diff --git a/mm/filemap.c b/mm/filemap.c
index c746541b1d49..18022de7dc33 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1219,6 +1219,14 @@ static int __wait_on_page_locked_async(struct page *page,
 	return ret;
 }
 
+static int wait_on_page_locked_async(struct page *page,
+				     struct wait_page_queue *wait)
+{
+	if (!PageLocked(page))
+		return 0;
+	return __wait_on_page_locked_async(compound_head(page), wait, false);
+}
+
 /**
  * put_and_wait_on_page_locked - Drop a reference and wait for it to be unlocked
  * @page: The page to wait for.
@@ -2058,17 +2066,25 @@ static ssize_t generic_file_buffered_read(struct kiocb *iocb,
 					index, last_index - index);
 		}
 		if (!PageUptodate(page)) {
-			if (iocb->ki_flags & IOCB_NOWAIT) {
-				put_page(page);
-				goto would_block;
-			}
-
 			/*
 			 * See comment in do_read_cache_page on why
 			 * wait_on_page_locked is used to avoid unnecessarily
 			 * serialisations and why it's safe.
 			 */
-			error = wait_on_page_locked_killable(page);
+			if (iocb->ki_flags & IOCB_WAITQ) {
+				if (written) {
+					put_page(page);
+					goto out;
+				}
+				error = wait_on_page_locked_async(page,
+								iocb->ki_waitq);
+			} else {
+				if (iocb->ki_flags & IOCB_NOWAIT) {
+					put_page(page);
+					goto would_block;
+				}
+				error = wait_on_page_locked_killable(page);
+			}
 			if (unlikely(error))
 				goto readpage_error;
 			if (PageUptodate(page))
@@ -2156,7 +2172,10 @@ static ssize_t generic_file_buffered_read(struct kiocb *iocb,
 
 page_not_up_to_date:
 		/* Get exclusive access to the page ... */
-		error = lock_page_killable(page);
+		if (iocb->ki_flags & IOCB_WAITQ)
+			error = lock_page_async(page, iocb->ki_waitq);
+		else
+			error = lock_page_killable(page);
 		if (unlikely(error))
 			goto readpage_error;
 
-- 
2.26.2

Re: [PATCH 05/12] mm: support async buffered reads in generic_file_buffered_read()

[Johannes Weiner]

On Tue, May 26, 2020 at 01:51:16PM -0600, Jens Axboe wrote:
> Use the async page locking infrastructure, if IOCB_WAITQ is set in the
> passed in iocb. The caller must expect an -EIOCBQUEUED return value,
> which means that IO is started but not done yet. This is similar to how
> O_DIRECT signals the same operation. Once the callback is received by
> the caller for IO completion, the caller must retry the operation.
> 
> Signed-off-by: Jens Axboe <axboe@kernel.dk>

Acked-by: Johannes Weiner <hannes@cmpxchg.org>

[PATCH 06/12] fs: add FMODE_BUF_RASYNC

[Jens Axboe]

If set, this indicates that the file system supports IOCB_WAITQ for
buffered reads.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 include/linux/fs.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/include/linux/fs.h b/include/linux/fs.h
index ba1fff0e7bca..5ffc6d236b01 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -175,6 +175,9 @@ typedef int (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
 /* File does not contribute to nr_files count */
 #define FMODE_NOACCOUNT		((__force fmode_t)0x20000000)
 
+/* File supports async buffered reads */
+#define FMODE_BUF_RASYNC	((__force fmode_t)0x40000000)
+
 /*
  * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
  * that indicates that they should check the contents of the iovec are
-- 
2.26.2

[PATCH 07/12] ext4: flag as supporting buffered async reads

[Jens Axboe]

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/ext4/file.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/ext4/file.c b/fs/ext4/file.c
index 0d624250a62b..9f7d9bf427b4 100644
--- a/fs/ext4/file.c
+++ b/fs/ext4/file.c
@@ -826,7 +826,7 @@ static int ext4_file_open(struct inode * inode, struct file * filp)
 			return ret;
 	}
 
-	filp->f_mode |= FMODE_NOWAIT;
+	filp->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC;
 	return dquot_file_open(inode, filp);
 }
 
-- 
2.26.2

[PATCH 08/12] block: flag block devices as supporting IOCB_WAITQ

[Jens Axboe]

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/block_dev.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/block_dev.c b/fs/block_dev.c
index d1e08bba925a..980cfce01c9a 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -1851,7 +1851,7 @@ static int blkdev_open(struct inode * inode, struct file * filp)
 	 */
 	filp->f_flags |= O_LARGEFILE;
 
-	filp->f_mode |= FMODE_NOWAIT;
+	filp->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC;
 
 	if (filp->f_flags & O_NDELAY)
 		filp->f_mode |= FMODE_NDELAY;
-- 
2.26.2

[PATCH 09/12] xfs: flag files as supporting buffered async reads

[Jens Axboe]

XFS uses generic_file_read_iter(), which already supports this.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/xfs/xfs_file.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index 4b8bdecc3863..97f44fbf17f2 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -1080,7 +1080,7 @@ xfs_file_open(
 		return -EFBIG;
 	if (XFS_FORCED_SHUTDOWN(XFS_M(inode->i_sb)))
 		return -EIO;
-	file->f_mode |= FMODE_NOWAIT;
+	file->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC;
 	return 0;
 }
 
-- 
2.26.2

Re: [PATCH 09/12] xfs: flag files as supporting buffered async reads

[Darrick J. Wong]

On Tue, May 26, 2020 at 01:51:20PM -0600, Jens Axboe wrote:
> XFS uses generic_file_read_iter(), which already supports this.
> 
> Signed-off-by: Jens Axboe <axboe@kernel.dk>

Er... I guess that looks ok?  Assuming you've done enough qa on
io_uring to be able to tell if this breaks anything, since touching the
mm always feels murky to me:

Acked-by: Darrick J. Wong <darrick.wong@oracle.com>

--D

> ---
>  fs/xfs/xfs_file.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
> index 4b8bdecc3863..97f44fbf17f2 100644
> --- a/fs/xfs/xfs_file.c
> +++ b/fs/xfs/xfs_file.c
> @@ -1080,7 +1080,7 @@ xfs_file_open(
>  		return -EFBIG;
>  	if (XFS_FORCED_SHUTDOWN(XFS_M(inode->i_sb)))
>  		return -EIO;
> -	file->f_mode |= FMODE_NOWAIT;
> +	file->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC;
>  	return 0;
>  }
>  
> -- 
> 2.26.2
> 

Re: [PATCH 09/12] xfs: flag files as supporting buffered async reads

[Jens Axboe]

On 5/28/20 11:53 AM, Darrick J. Wong wrote:
> On Tue, May 26, 2020 at 01:51:20PM -0600, Jens Axboe wrote:
>> XFS uses generic_file_read_iter(), which already supports this.
>>
>> Signed-off-by: Jens Axboe <axboe@kernel.dk>
> 
> Er... I guess that looks ok?  Assuming you've done enough qa on
> io_uring to be able to tell if this breaks anything, since touching the
> mm always feels murky to me:
> 

The mm bits should be fine, haven't seen anything odd in testing.
And it's not like the mm changes are super complicated, I think
they turned out pretty clean and straight forward.

> Acked-by: Darrick J. Wong <darrick.wong@oracle.com>

Thanks!

-- 
Jens Axboe

[PATCH 10/12] btrfs: flag files as supporting buffered async reads

[Jens Axboe]

btrfs uses generic_file_read_iter(), which already supports this.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/btrfs/file.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 719e68ab552c..c933b6a1b4a8 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -3480,7 +3480,7 @@ static loff_t btrfs_file_llseek(struct file *file, loff_t offset, int whence)
 
 static int btrfs_file_open(struct inode *inode, struct file *filp)
 {
-	filp->f_mode |= FMODE_NOWAIT;
+	filp->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC;
 	return generic_file_open(inode, filp);
 }
 
-- 
2.26.2

Re: [PATCH 10/12] btrfs: flag files as supporting buffered async reads

[Chris Mason]

On 26 May 2020, at 15:51, Jens Axboe wrote:

> btrfs uses generic_file_read_iter(), which already supports this.
>
> Signed-off-by: Jens Axboe <axboe@kernel.dk>

Really looking forward to this!

Acked-by: Chris Mason <clm@fb.com>

[PATCH 11/12] mm: add kiocb_wait_page_queue_init() helper

[Jens Axboe]

Checks if the file supports it, and initializes the values that we need.
Caller passes in 'data' pointer, if any, and the callback function to
be used.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 include/linux/pagemap.h | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
index d3e63c9c61ae..8b65420410ee 100644
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -493,6 +493,27 @@ static inline int wake_page_match(struct wait_page_queue *wait_page,
 	return 1;
 }
 
+static inline int kiocb_wait_page_queue_init(struct kiocb *kiocb,
+					     struct wait_page_queue *wait,
+					     wait_queue_func_t func,
+					     void *data)
+{
+	/* Can't support async wakeup with polled IO */
+	if (kiocb->ki_flags & IOCB_HIPRI)
+		return -EINVAL;
+	if (kiocb->ki_filp->f_mode & FMODE_BUF_RASYNC) {
+		wait->wait.func = func;
+		wait->wait.private = data;
+		wait->wait.flags = 0;
+		INIT_LIST_HEAD(&wait->wait.entry);
+		kiocb->ki_flags |= IOCB_WAITQ;
+		kiocb->ki_waitq = wait;
+		return 0;
+	}
+
+	return -EOPNOTSUPP;
+}
+
 extern void __lock_page(struct page *page);
 extern int __lock_page_killable(struct page *page);
 extern int __lock_page_async(struct page *page, struct wait_page_queue *wait);
-- 
2.26.2

Re: [PATCH 11/12] mm: add kiocb_wait_page_queue_init() helper

[Johannes Weiner]

On Tue, May 26, 2020 at 01:51:22PM -0600, Jens Axboe wrote:
> Checks if the file supports it, and initializes the values that we need.
> Caller passes in 'data' pointer, if any, and the callback function to
> be used.
> 
> Signed-off-by: Jens Axboe <axboe@kernel.dk>

Acked-by: Johannes Weiner <hannes@cmpxchg.org>

[PATCH 12/12] io_uring: support true async buffered reads, if file provides it

[Jens Axboe]

If the file is flagged with FMODE_BUF_RASYNC, then we don't have to punt
the buffered read to an io-wq worker. Instead we can rely on page
unlocking callbacks to support retry based async IO. This is a lot more
efficient than doing async thread offload.

The retry is done similarly to how we handle poll based retry. From
the unlock callback, we simply queue the retry to a task_work based
handler.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c | 130 ++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 126 insertions(+), 4 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index e6865afa8467..95df63b0b2ce 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -79,6 +79,7 @@
 #include <linux/fs_struct.h>
 #include <linux/splice.h>
 #include <linux/task_work.h>
+#include <linux/pagemap.h>
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/io_uring.h>
@@ -498,6 +499,8 @@ struct io_async_rw {
 	struct iovec			*iov;
 	ssize_t				nr_segs;
 	ssize_t				size;
+	struct wait_page_queue		wpq;
+	struct callback_head		task_work;
 };
 
 struct io_async_ctx {
@@ -2568,6 +2571,119 @@ static int io_read_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe,
 	return 0;
 }
 
+static void io_async_buf_cancel(struct callback_head *cb)
+{
+	struct io_async_rw *rw;
+	struct io_ring_ctx *ctx;
+	struct io_kiocb *req;
+
+	rw = container_of(cb, struct io_async_rw, task_work);
+	req = rw->wpq.wait.private;
+	ctx = req->ctx;
+
+	spin_lock_irq(&ctx->completion_lock);
+	io_cqring_fill_event(req, -ECANCELED);
+	io_commit_cqring(ctx);
+	spin_unlock_irq(&ctx->completion_lock);
+
+	io_cqring_ev_posted(ctx);
+	req_set_fail_links(req);
+	io_double_put_req(req);
+}
+
+static void io_async_buf_retry(struct callback_head *cb)
+{
+	struct io_async_rw *rw;
+	struct io_ring_ctx *ctx;
+	struct io_kiocb *req;
+
+	rw = container_of(cb, struct io_async_rw, task_work);
+	req = rw->wpq.wait.private;
+	ctx = req->ctx;
+
+	__set_current_state(TASK_RUNNING);
+	mutex_lock(&ctx->uring_lock);
+	__io_queue_sqe(req, NULL);
+	mutex_unlock(&ctx->uring_lock);
+}
+
+static int io_async_buf_func(struct wait_queue_entry *wait, unsigned mode,
+			     int sync, void *arg)
+{
+	struct wait_page_queue *wpq;
+	struct io_kiocb *req = wait->private;
+	struct io_async_rw *rw = &req->io->rw;
+	struct wait_page_key *key = arg;
+	struct task_struct *tsk;
+	int ret;
+
+	wpq = container_of(wait, struct wait_page_queue, wait);
+
+	ret = wake_page_match(wpq, key);
+	if (ret != 1)
+		return ret;
+
+	list_del_init(&wait->entry);
+
+	init_task_work(&rw->task_work, io_async_buf_retry);
+	/* submit ref gets dropped, acquire a new one */
+	refcount_inc(&req->refs);
+	tsk = req->task;
+	ret = task_work_add(tsk, &rw->task_work, true);
+	if (unlikely(ret)) {
+		/* queue just for cancelation */
+		init_task_work(&rw->task_work, io_async_buf_cancel);
+		tsk = io_wq_get_task(req->ctx->io_wq);
+		task_work_add(tsk, &rw->task_work, true);
+	}
+	wake_up_process(tsk);
+	return 1;
+}
+
+static bool io_rw_should_retry(struct io_kiocb *req)
+{
+	struct kiocb *kiocb = &req->rw.kiocb;
+	int ret;
+
+	/* never retry for NOWAIT, we just complete with -EAGAIN */
+	if (req->flags & REQ_F_NOWAIT)
+		return false;
+
+	/* already tried, or we're doing O_DIRECT */
+	if (kiocb->ki_flags & (IOCB_DIRECT | IOCB_WAITQ))
+		return false;
+	/*
+	 * just use poll if we can, and don't attempt if the fs doesn't
+	 * support callback based unlocks
+	 */
+	if (file_can_poll(req->file) || !(req->file->f_mode & FMODE_BUF_RASYNC))
+		return false;
+
+	/*
+	 * If request type doesn't require req->io to defer in general,
+	 * we need to allocate it here
+	 */
+	if (!req->io && __io_alloc_async_ctx(req))
+		return false;
+
+	ret = kiocb_wait_page_queue_init(kiocb, &req->io->rw.wpq,
+						io_async_buf_func, req);
+	if (!ret) {
+		get_task_struct(current);
+		req->task = current;
+		return true;
+	}
+
+	return false;
+}
+
+static int io_iter_do_read(struct io_kiocb *req, struct iov_iter *iter)
+{
+	if (req->file->f_op->read_iter)
+		return call_read_iter(req->file, &req->rw.kiocb, iter);
+	return loop_rw_iter(READ, req->file, &req->rw.kiocb, iter);
+}
+
 static int io_read(struct io_kiocb *req, bool force_nonblock)
 {
 	struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
@@ -2601,10 +2717,7 @@ static int io_read(struct io_kiocb *req, bool force_nonblock)
 	if (!ret) {
 		ssize_t ret2;
 
-		if (req->file->f_op->read_iter)
-			ret2 = call_read_iter(req->file, kiocb, &iter);
-		else
-			ret2 = loop_rw_iter(READ, req->file, kiocb, &iter);
+		ret2 = io_iter_do_read(req, &iter);
 
 		/* Catch -EAGAIN return for forced non-blocking submission */
 		if (!force_nonblock || ret2 != -EAGAIN) {
@@ -2619,6 +2732,15 @@ static int io_read(struct io_kiocb *req, bool force_nonblock)
 			if (!(req->flags & REQ_F_NOWAIT) &&
 			    !file_can_poll(req->file))
 				req->flags |= REQ_F_MUST_PUNT;
+			/* if we can retry, do so with the callbacks armed */
+			if (io_rw_should_retry(req)) {
+				ret2 = io_iter_do_read(req, &iter);
+				if (ret2 != -EAGAIN) {
+					kiocb_done(kiocb, ret2);
+					goto out_free;
+				}
+			}
+			kiocb->ki_flags &= ~IOCB_WAITQ;
 			return -EAGAIN;
 		}
 	}
-- 
2.26.2

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 5/28/20 11:02 AM, Sedat Dilek wrote:
> On Tue, May 26, 2020 at 10:59 PM Jens Axboe <axboe@kernel.dk> wrote:
>>
>> We technically support this already through io_uring, but it's
>> implemented with a thread backend to support cases where we would
>> block. This isn't ideal.
>>
>> After a few prep patches, the core of this patchset is adding support
>> for async callbacks on page unlock. With this primitive, we can simply
>> retry the IO operation. With io_uring, this works a lot like poll based
>> retry for files that support it. If a page is currently locked and
>> needed, -EIOCBQUEUED is returned with a callback armed. The callers
>> callback is responsible for restarting the operation.
>>
>> With this callback primitive, we can add support for
>> generic_file_buffered_read(), which is what most file systems end up
>> using for buffered reads. XFS/ext4/btrfs/bdev is wired up, but probably
>> trivial to add more.
>>
>> The file flags support for this by setting FMODE_BUF_RASYNC, similar
>> to what we do for FMODE_NOWAIT. Open to suggestions here if this is
>> the preferred method or not.
>>
>> In terms of results, I wrote a small test app that randomly reads 4G
>> of data in 4K chunks from a file hosted by ext4. The app uses a queue
>> depth of 32. If you want to test yourself, you can just use buffered=1
>> with ioengine=io_uring with fio. No application changes are needed to
>> use the more optimized buffered async read.
>>
>> preadv for comparison:
>>         real    1m13.821s
>>         user    0m0.558s
>>         sys     0m11.125s
>>         CPU     ~13%
>>
>> Mainline:
>>         real    0m12.054s
>>         user    0m0.111s
>>         sys     0m5.659s
>>         CPU     ~32% + ~50% == ~82%
>>
>> This patchset:
>>         real    0m9.283s
>>         user    0m0.147s
>>         sys     0m4.619s
>>         CPU     ~52%
>>
>> The CPU numbers are just a rough estimate. For the mainline io_uring
>> run, this includes the app itself and all the threads doing IO on its
>> behalf (32% for the app, ~1.6% per worker and 32 of them). Context
>> switch rate is much smaller with the patchset, since we only have the
>> one task performing IO.
>>
>> Also ran a simple fio based test case, varying the queue depth from 1
>> to 16, doubling every time:
>>
>> [buf-test]
>> filename=/data/file
>> direct=0
>> ioengine=io_uring
>> norandommap
>> rw=randread
>> bs=4k
>> iodepth=${QD}
>> randseed=89
>> runtime=10s
>>
>> QD/Test         Patchset IOPS           Mainline IOPS
>> 1               9046                    8294
>> 2               19.8k                   18.9k
>> 4               39.2k                   28.5k
>> 8               64.4k                   31.4k
>> 16              65.7k                   37.8k
>>
>> Outside of my usual environment, so this is just running on a virtualized
>> NVMe device in qemu, using ext4 as the file system. NVMe isn't very
>> efficient virtualized, so we run out of steam at ~65K which is why we
>> flatline on the patched side (nvme_submit_cmd() eats ~75% of the test app
>> CPU). Before that happens, it's a linear increase. Not shown is context
>> switch rate, which is massively lower with the new code. The old thread
>> offload adds a blocking thread per pending IO, so context rate quickly
>> goes through the roof.
>>
>> The goal here is efficiency. Async thread offload adds latency, and
>> it also adds noticable overhead on items such as adding pages to the
>> page cache. By allowing proper async buffered read support, we don't
>> have X threads hammering on the same inode page cache, we have just
>> the single app actually doing IO.
>>
>> Been beating on this and it's solid for me, and I'm now pretty happy
>> with how it all turned out. Not aware of any missing bits/pieces or
>> code cleanups that need doing.
>>
>> Series can also be found here:
>>
>> https://git.kernel.dk/cgit/linux-block/log/?h=async-buffered.5
>>
>> or pull from:
>>
>> git://git.kernel.dk/linux-block async-buffered.5
>>
> 
> Hi Jens,
> 
> I have pulled linux-block.git#async-buffered.5 on top of Linux v5.7-rc7.
> 
> From first feelings:
> The booting into the system (until sddm display-login-manager) took a
> bit longer.
> The same after login and booting into KDE/Plasma.

There is no difference for "regular" use cases, only io_uring with
buffered reads will behave differently. So I don't think you have longer
boot times due to this.

> I am building/linking with LLVM/Clang/LLD v10.0.1-rc1 on Debian/testing AMD64.
> 
> Here I have an internal HDD (SATA) and my Debian-system is on an
> external HDD connected via USB-3.0.
> Primarily, I use Ext4-FS.
> 
> As said above is the "emotional" side, but I need some technical instructions.
> 
> How can I see Async Buffer Reads is active on a Ext4-FS-formatted partition?

You can't see that. It'll always be available on ext4 with this series,
and you can watch io_uring instances to see if anyone is using it.

> Do I need a special boot-parameter (GRUB line)?
> 
> Do I need to activate some cool variables via sysfs?
> 
> Do I need to pass an option via fstab entry?

No to all of these, you don't need anything to activate it. You need the
program to use io_uring to do buffered reads.

> Are any Async Buffer Reads related linux-kconfig options not set?
> Which make sense?

No kconfig options are needed.

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

On Thu, May 28, 2020 at 7:06 PM Jens Axboe <axboe@kernel.dk> wrote:
>
> On 5/28/20 11:02 AM, Sedat Dilek wrote:
> > On Tue, May 26, 2020 at 10:59 PM Jens Axboe <axboe@kernel.dk> wrote:
> >>
> >> We technically support this already through io_uring, but it's
> >> implemented with a thread backend to support cases where we would
> >> block. This isn't ideal.
> >>
> >> After a few prep patches, the core of this patchset is adding support
> >> for async callbacks on page unlock. With this primitive, we can simply
> >> retry the IO operation. With io_uring, this works a lot like poll based
> >> retry for files that support it. If a page is currently locked and
> >> needed, -EIOCBQUEUED is returned with a callback armed. The callers
> >> callback is responsible for restarting the operation.
> >>
> >> With this callback primitive, we can add support for
> >> generic_file_buffered_read(), which is what most file systems end up
> >> using for buffered reads. XFS/ext4/btrfs/bdev is wired up, but probably
> >> trivial to add more.
> >>
> >> The file flags support for this by setting FMODE_BUF_RASYNC, similar
> >> to what we do for FMODE_NOWAIT. Open to suggestions here if this is
> >> the preferred method or not.
> >>
> >> In terms of results, I wrote a small test app that randomly reads 4G
> >> of data in 4K chunks from a file hosted by ext4. The app uses a queue
> >> depth of 32. If you want to test yourself, you can just use buffered=1
> >> with ioengine=io_uring with fio. No application changes are needed to
> >> use the more optimized buffered async read.
> >>
> >> preadv for comparison:
> >>         real    1m13.821s
> >>         user    0m0.558s
> >>         sys     0m11.125s
> >>         CPU     ~13%
> >>
> >> Mainline:
> >>         real    0m12.054s
> >>         user    0m0.111s
> >>         sys     0m5.659s
> >>         CPU     ~32% + ~50% == ~82%
> >>
> >> This patchset:
> >>         real    0m9.283s
> >>         user    0m0.147s
> >>         sys     0m4.619s
> >>         CPU     ~52%
> >>
> >> The CPU numbers are just a rough estimate. For the mainline io_uring
> >> run, this includes the app itself and all the threads doing IO on its
> >> behalf (32% for the app, ~1.6% per worker and 32 of them). Context
> >> switch rate is much smaller with the patchset, since we only have the
> >> one task performing IO.
> >>
> >> Also ran a simple fio based test case, varying the queue depth from 1
> >> to 16, doubling every time:
> >>
> >> [buf-test]
> >> filename=/data/file
> >> direct=0
> >> ioengine=io_uring
> >> norandommap
> >> rw=randread
> >> bs=4k
> >> iodepth=${QD}
> >> randseed=89
> >> runtime=10s
> >>
> >> QD/Test         Patchset IOPS           Mainline IOPS
> >> 1               9046                    8294
> >> 2               19.8k                   18.9k
> >> 4               39.2k                   28.5k
> >> 8               64.4k                   31.4k
> >> 16              65.7k                   37.8k
> >>
> >> Outside of my usual environment, so this is just running on a virtualized
> >> NVMe device in qemu, using ext4 as the file system. NVMe isn't very
> >> efficient virtualized, so we run out of steam at ~65K which is why we
> >> flatline on the patched side (nvme_submit_cmd() eats ~75% of the test app
> >> CPU). Before that happens, it's a linear increase. Not shown is context
> >> switch rate, which is massively lower with the new code. The old thread
> >> offload adds a blocking thread per pending IO, so context rate quickly
> >> goes through the roof.
> >>
> >> The goal here is efficiency. Async thread offload adds latency, and
> >> it also adds noticable overhead on items such as adding pages to the
> >> page cache. By allowing proper async buffered read support, we don't
> >> have X threads hammering on the same inode page cache, we have just
> >> the single app actually doing IO.
> >>
> >> Been beating on this and it's solid for me, and I'm now pretty happy
> >> with how it all turned out. Not aware of any missing bits/pieces or
> >> code cleanups that need doing.
> >>
> >> Series can also be found here:
> >>
> >> https://git.kernel.dk/cgit/linux-block/log/?h=async-buffered.5
> >>
> >> or pull from:
> >>
> >> git://git.kernel.dk/linux-block async-buffered.5
> >>
> >
> > Hi Jens,
> >
> > I have pulled linux-block.git#async-buffered.5 on top of Linux v5.7-rc7.
> >
> > From first feelings:
> > The booting into the system (until sddm display-login-manager) took a
> > bit longer.
> > The same after login and booting into KDE/Plasma.
>
> There is no difference for "regular" use cases, only io_uring with
> buffered reads will behave differently. So I don't think you have longer
> boot times due to this.
>
> > I am building/linking with LLVM/Clang/LLD v10.0.1-rc1 on Debian/testing AMD64.
> >
> > Here I have an internal HDD (SATA) and my Debian-system is on an
> > external HDD connected via USB-3.0.
> > Primarily, I use Ext4-FS.
> >
> > As said above is the "emotional" side, but I need some technical instructions.
> >
> > How can I see Async Buffer Reads is active on a Ext4-FS-formatted partition?
>
> You can't see that. It'll always be available on ext4 with this series,
> and you can watch io_uring instances to see if anyone is using it.
>

Thanks for answering my questions.

How can I "watch io_uring instances"?

FIO?
Debian has fio version 3.19-2 in its apt repositories.
Version OK?

- Sedat -

> > Do I need a special boot-parameter (GRUB line)?
> >
> > Do I need to activate some cool variables via sysfs?
> >
> > Do I need to pass an option via fstab entry?
>
> No to all of these, you don't need anything to activate it. You need the
> program to use io_uring to do buffered reads.
>
> > Are any Async Buffer Reads related linux-kconfig options not set?
> > Which make sense?
>
> No kconfig options are needed.
>
> --
> Jens Axboe
>

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 5/28/20 11:12 AM, Sedat Dilek wrote:
> On Thu, May 28, 2020 at 7:06 PM Jens Axboe <axboe@kernel.dk> wrote:
>>
>> On 5/28/20 11:02 AM, Sedat Dilek wrote:
>>> On Tue, May 26, 2020 at 10:59 PM Jens Axboe <axboe@kernel.dk> wrote:
>>>>
>>>> We technically support this already through io_uring, but it's
>>>> implemented with a thread backend to support cases where we would
>>>> block. This isn't ideal.
>>>>
>>>> After a few prep patches, the core of this patchset is adding support
>>>> for async callbacks on page unlock. With this primitive, we can simply
>>>> retry the IO operation. With io_uring, this works a lot like poll based
>>>> retry for files that support it. If a page is currently locked and
>>>> needed, -EIOCBQUEUED is returned with a callback armed. The callers
>>>> callback is responsible for restarting the operation.
>>>>
>>>> With this callback primitive, we can add support for
>>>> generic_file_buffered_read(), which is what most file systems end up
>>>> using for buffered reads. XFS/ext4/btrfs/bdev is wired up, but probably
>>>> trivial to add more.
>>>>
>>>> The file flags support for this by setting FMODE_BUF_RASYNC, similar
>>>> to what we do for FMODE_NOWAIT. Open to suggestions here if this is
>>>> the preferred method or not.
>>>>
>>>> In terms of results, I wrote a small test app that randomly reads 4G
>>>> of data in 4K chunks from a file hosted by ext4. The app uses a queue
>>>> depth of 32. If you want to test yourself, you can just use buffered=1
>>>> with ioengine=io_uring with fio. No application changes are needed to
>>>> use the more optimized buffered async read.
>>>>
>>>> preadv for comparison:
>>>>         real    1m13.821s
>>>>         user    0m0.558s
>>>>         sys     0m11.125s
>>>>         CPU     ~13%
>>>>
>>>> Mainline:
>>>>         real    0m12.054s
>>>>         user    0m0.111s
>>>>         sys     0m5.659s
>>>>         CPU     ~32% + ~50% == ~82%
>>>>
>>>> This patchset:
>>>>         real    0m9.283s
>>>>         user    0m0.147s
>>>>         sys     0m4.619s
>>>>         CPU     ~52%
>>>>
>>>> The CPU numbers are just a rough estimate. For the mainline io_uring
>>>> run, this includes the app itself and all the threads doing IO on its
>>>> behalf (32% for the app, ~1.6% per worker and 32 of them). Context
>>>> switch rate is much smaller with the patchset, since we only have the
>>>> one task performing IO.
>>>>
>>>> Also ran a simple fio based test case, varying the queue depth from 1
>>>> to 16, doubling every time:
>>>>
>>>> [buf-test]
>>>> filename=/data/file
>>>> direct=0
>>>> ioengine=io_uring
>>>> norandommap
>>>> rw=randread
>>>> bs=4k
>>>> iodepth=${QD}
>>>> randseed=89
>>>> runtime=10s
>>>>
>>>> QD/Test         Patchset IOPS           Mainline IOPS
>>>> 1               9046                    8294
>>>> 2               19.8k                   18.9k
>>>> 4               39.2k                   28.5k
>>>> 8               64.4k                   31.4k
>>>> 16              65.7k                   37.8k
>>>>
>>>> Outside of my usual environment, so this is just running on a virtualized
>>>> NVMe device in qemu, using ext4 as the file system. NVMe isn't very
>>>> efficient virtualized, so we run out of steam at ~65K which is why we
>>>> flatline on the patched side (nvme_submit_cmd() eats ~75% of the test app
>>>> CPU). Before that happens, it's a linear increase. Not shown is context
>>>> switch rate, which is massively lower with the new code. The old thread
>>>> offload adds a blocking thread per pending IO, so context rate quickly
>>>> goes through the roof.
>>>>
>>>> The goal here is efficiency. Async thread offload adds latency, and
>>>> it also adds noticable overhead on items such as adding pages to the
>>>> page cache. By allowing proper async buffered read support, we don't
>>>> have X threads hammering on the same inode page cache, we have just
>>>> the single app actually doing IO.
>>>>
>>>> Been beating on this and it's solid for me, and I'm now pretty happy
>>>> with how it all turned out. Not aware of any missing bits/pieces or
>>>> code cleanups that need doing.
>>>>
>>>> Series can also be found here:
>>>>
>>>> https://git.kernel.dk/cgit/linux-block/log/?h=async-buffered.5
>>>>
>>>> or pull from:
>>>>
>>>> git://git.kernel.dk/linux-block async-buffered.5
>>>>
>>>
>>> Hi Jens,
>>>
>>> I have pulled linux-block.git#async-buffered.5 on top of Linux v5.7-rc7.
>>>
>>> From first feelings:
>>> The booting into the system (until sddm display-login-manager) took a
>>> bit longer.
>>> The same after login and booting into KDE/Plasma.
>>
>> There is no difference for "regular" use cases, only io_uring with
>> buffered reads will behave differently. So I don't think you have longer
>> boot times due to this.
>>
>>> I am building/linking with LLVM/Clang/LLD v10.0.1-rc1 on Debian/testing AMD64.
>>>
>>> Here I have an internal HDD (SATA) and my Debian-system is on an
>>> external HDD connected via USB-3.0.
>>> Primarily, I use Ext4-FS.
>>>
>>> As said above is the "emotional" side, but I need some technical instructions.
>>>
>>> How can I see Async Buffer Reads is active on a Ext4-FS-formatted partition?
>>
>> You can't see that. It'll always be available on ext4 with this series,
>> and you can watch io_uring instances to see if anyone is using it.
>>
> 
> Thanks for answering my questions.
> 
> How can I "watch io_uring instances"?

You can enable io_uring tracing:

# echo 1 > /sys/kernel/debug/tracing/events/io_uring/io_uring_create/enable
# tail /sys/kernel/debug/tracing/trace

and see if you get any events for setup. Generally you can also look for
the existence of io_wq_manager processes, these will exist for an
io_uring instance.

> FIO?
> Debian has fio version 3.19-2 in its apt repositories.
> Version OK?

Yeah that should work.

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

On Thu, May 28, 2020 at 7:14 PM Jens Axboe <axboe@kernel.dk> wrote:
>
> On 5/28/20 11:12 AM, Sedat Dilek wrote:
> > On Thu, May 28, 2020 at 7:06 PM Jens Axboe <axboe@kernel.dk> wrote:
> >>
> >> On 5/28/20 11:02 AM, Sedat Dilek wrote:
> >>> On Tue, May 26, 2020 at 10:59 PM Jens Axboe <axboe@kernel.dk> wrote:
> >>>>
> >>>> We technically support this already through io_uring, but it's
> >>>> implemented with a thread backend to support cases where we would
> >>>> block. This isn't ideal.
> >>>>
> >>>> After a few prep patches, the core of this patchset is adding support
> >>>> for async callbacks on page unlock. With this primitive, we can simply
> >>>> retry the IO operation. With io_uring, this works a lot like poll based
> >>>> retry for files that support it. If a page is currently locked and
> >>>> needed, -EIOCBQUEUED is returned with a callback armed. The callers
> >>>> callback is responsible for restarting the operation.
> >>>>
> >>>> With this callback primitive, we can add support for
> >>>> generic_file_buffered_read(), which is what most file systems end up
> >>>> using for buffered reads. XFS/ext4/btrfs/bdev is wired up, but probably
> >>>> trivial to add more.
> >>>>
> >>>> The file flags support for this by setting FMODE_BUF_RASYNC, similar
> >>>> to what we do for FMODE_NOWAIT. Open to suggestions here if this is
> >>>> the preferred method or not.
> >>>>
> >>>> In terms of results, I wrote a small test app that randomly reads 4G
> >>>> of data in 4K chunks from a file hosted by ext4. The app uses a queue
> >>>> depth of 32. If you want to test yourself, you can just use buffered=1
> >>>> with ioengine=io_uring with fio. No application changes are needed to
> >>>> use the more optimized buffered async read.
> >>>>
> >>>> preadv for comparison:
> >>>>         real    1m13.821s
> >>>>         user    0m0.558s
> >>>>         sys     0m11.125s
> >>>>         CPU     ~13%
> >>>>
> >>>> Mainline:
> >>>>         real    0m12.054s
> >>>>         user    0m0.111s
> >>>>         sys     0m5.659s
> >>>>         CPU     ~32% + ~50% == ~82%
> >>>>
> >>>> This patchset:
> >>>>         real    0m9.283s
> >>>>         user    0m0.147s
> >>>>         sys     0m4.619s
> >>>>         CPU     ~52%
> >>>>
> >>>> The CPU numbers are just a rough estimate. For the mainline io_uring
> >>>> run, this includes the app itself and all the threads doing IO on its
> >>>> behalf (32% for the app, ~1.6% per worker and 32 of them). Context
> >>>> switch rate is much smaller with the patchset, since we only have the
> >>>> one task performing IO.
> >>>>
> >>>> Also ran a simple fio based test case, varying the queue depth from 1
> >>>> to 16, doubling every time:
> >>>>
> >>>> [buf-test]
> >>>> filename=/data/file
> >>>> direct=0
> >>>> ioengine=io_uring
> >>>> norandommap
> >>>> rw=randread
> >>>> bs=4k
> >>>> iodepth=${QD}
> >>>> randseed=89
> >>>> runtime=10s
> >>>>
> >>>> QD/Test         Patchset IOPS           Mainline IOPS
> >>>> 1               9046                    8294
> >>>> 2               19.8k                   18.9k
> >>>> 4               39.2k                   28.5k
> >>>> 8               64.4k                   31.4k
> >>>> 16              65.7k                   37.8k
> >>>>
> >>>> Outside of my usual environment, so this is just running on a virtualized
> >>>> NVMe device in qemu, using ext4 as the file system. NVMe isn't very
> >>>> efficient virtualized, so we run out of steam at ~65K which is why we
> >>>> flatline on the patched side (nvme_submit_cmd() eats ~75% of the test app
> >>>> CPU). Before that happens, it's a linear increase. Not shown is context
> >>>> switch rate, which is massively lower with the new code. The old thread
> >>>> offload adds a blocking thread per pending IO, so context rate quickly
> >>>> goes through the roof.
> >>>>
> >>>> The goal here is efficiency. Async thread offload adds latency, and
> >>>> it also adds noticable overhead on items such as adding pages to the
> >>>> page cache. By allowing proper async buffered read support, we don't
> >>>> have X threads hammering on the same inode page cache, we have just
> >>>> the single app actually doing IO.
> >>>>
> >>>> Been beating on this and it's solid for me, and I'm now pretty happy
> >>>> with how it all turned out. Not aware of any missing bits/pieces or
> >>>> code cleanups that need doing.
> >>>>
> >>>> Series can also be found here:
> >>>>
> >>>> https://git.kernel.dk/cgit/linux-block/log/?h=async-buffered.5
> >>>>
> >>>> or pull from:
> >>>>
> >>>> git://git.kernel.dk/linux-block async-buffered.5
> >>>>
> >>>
> >>> Hi Jens,
> >>>
> >>> I have pulled linux-block.git#async-buffered.5 on top of Linux v5.7-rc7.
> >>>
> >>> From first feelings:
> >>> The booting into the system (until sddm display-login-manager) took a
> >>> bit longer.
> >>> The same after login and booting into KDE/Plasma.
> >>
> >> There is no difference for "regular" use cases, only io_uring with
> >> buffered reads will behave differently. So I don't think you have longer
> >> boot times due to this.
> >>
> >>> I am building/linking with LLVM/Clang/LLD v10.0.1-rc1 on Debian/testing AMD64.
> >>>
> >>> Here I have an internal HDD (SATA) and my Debian-system is on an
> >>> external HDD connected via USB-3.0.
> >>> Primarily, I use Ext4-FS.
> >>>
> >>> As said above is the "emotional" side, but I need some technical instructions.
> >>>
> >>> How can I see Async Buffer Reads is active on a Ext4-FS-formatted partition?
> >>
> >> You can't see that. It'll always be available on ext4 with this series,
> >> and you can watch io_uring instances to see if anyone is using it.
> >>
> >
> > Thanks for answering my questions.
> >
> > How can I "watch io_uring instances"?
>
> You can enable io_uring tracing:
>
> # echo 1 > /sys/kernel/debug/tracing/events/io_uring/io_uring_create/enable
> # tail /sys/kernel/debug/tracing/trace
>
> and see if you get any events for setup. Generally you can also look for
> the existence of io_wq_manager processes, these will exist for an
> io_uring instance.
>
> > FIO?
> > Debian has fio version 3.19-2 in its apt repositories.
> > Version OK?
>
> Yeah that should work.
>

I did:

# echo 1 > /sys/kernel/debug/tracing/events/io_uring/io_uring_create/enable
# cat /sys/kernel/debug/tracing/events/io_uring/io_uring_create/enable
1

# cat ./buf-test-dileks-min
[buf-test-dileks-min]
filename=/path/to/iso-image-file
buffered=1
ioengine=io_uring

# fio --showcmd ./buf-test-dileks-min
fio --name=buf-test-dileks-min --buffered=1 --ioengine=io_uring
--filename=filename=/path/to/iso-image-file

# fio ./buf-test-dileks-min

# ps -ef | egrep 'f[i]o|i[o]_wq'
root        6695    6066 24 20:13 pts/2    00:00:00 fio ./buf-test-dileks-min
root        6701    6695 22 20:13 ?        00:00:00 fio ./buf-test-dileks-min
root        6702       2  0 20:13 ?        00:00:00 [io_wq_manager]
root        6703       2  0 20:13 ?        00:00:00 [io_wqe_worker-0]

# LC_ALL=C tail -f /sys/kernel/debug/tracing/trace
...
# entries-in-buffer/entries-written: 16/16   #P:4
#
#                              _-----=> irqs-off
#                             / _----=> need-resched
#                            | / _---=> hardirq/softirq
#                            || / _--=> preempt-depth
#                            ||| /     delay
#           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
#              | |       |   ||||       |         |
...
             fio-6701  [001] ....  6775.117015: io_uring_create: ring
00000000ef052188, fd 5 sq size 1, cq size 2, flags 0

Looks like this works.

Thanks Jens.

- Sedat -

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

[ ... ]

> > Hi Jens,
> >
> > I have pulled linux-block.git#async-buffered.5 on top of Linux v5.7-rc7.
> >
> > From first feelings:
> > The booting into the system (until sddm display-login-manager) took a
> > bit longer.
> > The same after login and booting into KDE/Plasma.
>
> There is no difference for "regular" use cases, only io_uring with
> buffered reads will behave differently. So I don't think you have longer
> boot times due to this.
>

Yupp, you are right.

The previous Linux v5.7-rc7 without your patchset shows the same symptoms.

I did some debugging and optimizing with systemd-analyze boot and time.

I optimized systemd-journald.service and systemd-journal-flush.service...

# cat /etc/systemd/journald.conf.d/00-journal-size.conf
[Journal]
SystemMaxUse=50M

...and reduced the time spent flushing systemd's journal from ~30s
down to 1,6s...

# journalctl -b --unit systemd-journald.service
-- Logs begin at Fri 2020-05-29 00:58:37 CEST, end at Fri 2020-05-29
11:42:18 CEST. --
Mai 29 11:34:52 iniza systemd-journald[281]: Journal started
Mai 29 11:34:52 iniza systemd-journald[281]: Runtime Journal
(/run/log/journal/566abbcb226b405db834b17a26fe4727) is 8.0M, max
78.5M, 70.5M free.
Mai 29 11:34:53 iniza systemd-journald[281]: Time spent on flushing to
/var/log/journal/566abbcb226b405db834b17a26fe4727 is 1.656233s for 765
entries.
Mai 29 11:34:53 iniza systemd-journald[281]: System Journal
(/var/log/journal/566abbcb226b405db834b17a26fe4727) is 56.2M, max
50.0M, 0B free.

Unfortunately, I upgraded some user-space stuff like udisks2 and
libblockdev packages.
Downgrading did not help and disabling the systemd-unit also.

As I saw stallings with e2scrub_reap.service and swap partition
(partly seen in the boot-process and noted the UUID 3f8e).
I disabled e2scrub_reap.service and deactivated swap partition in /etc/fstab.

Doing all above together did not help.

Finally, I checked the health of the HDD where my root-fs is.
smartmontools says everything is OK.

I have not checked the status of the Ext4-FS where my root-fs is.
Such things I do with a linux-live-system - as a Debianist I admit I
use an ArchLinux ISO on USB-stick :-).

Unsure, if I will contact the systemd (and mabye udisks) Debian folks
to hear their opinion.

Thanks Jens and your patchset.
I don't know when I last run systemd-analyze & stuff and investigated
so deeply :-).

A lot of Hygge (I love to write wrong Huegge - see English hugs) to you Jens.

- Sedat -

[1] http://ftp.halifax.rwth-aachen.de/archlinux/iso/

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

On Fri, May 29, 2020 at 12:02 PM Sedat Dilek <sedat.dilek@gmail.com> wrote:

[ ... ]

> As I saw stallings with e2scrub_reap.service and swap partition
> (partly seen in the boot-process and noted the UUID 3f8e).
> I disabled e2scrub_reap.service and deactivated swap partition in /etc/fstab.
>

I switched over from using a swap-partition to a swap-file.

The boot-slowdown is "gone" in sense of 1m30s stalls or better in
sense of boot-time is shorter.

# cat systemd-analyze-time.txt
Startup finished in 6.903s (kernel) + 1min 13.501s (userspace) = 1min 20.404s
graphical.target reached after 1min 13.481s in userspace

# cat systemd-analyze-time_swapfile.txt
Startup finished in 6.721s (kernel) + 1min 9.470s (userspace) = 1min 16.192s
graphical.target reached after 1min 9.451s in userspace

# cat systemd-analyze-blame.txt | head -20
35.943s udisks2.service
32.559s accounts-daemon.service
27.925s smartmontools.service
26.561s NetworkManager.service
24.543s dev-sdc2.device
24.478s polkit.service
20.426s NetworkManager-wait-online.service
19.785s avahi-daemon.service
19.586s switcheroo-control.service
19.185s rtkit-daemon.service
18.661s wpa_supplicant.service
18.269s systemd-logind.service
17.627s rsyslog.service
16.312s gpm.service
14.842s e2scrub_reap.service
14.387s packagekit.service
12.017s ModemManager.service
10.584s alsa-restore.service
 8.407s atd.service
 6.025s exim4.service

# cat systemd-analyze-blame_swapfile.txt | head -20
29.571s udisks2.service
26.383s accounts-daemon.service
24.564s smartmontools.service
20.735s NetworkManager.service
19.254s NetworkManager-wait-online.service
18.675s polkit.service
15.523s dev-sdc2.device
14.152s avahi-daemon.service
14.006s switcheroo-control.service
13.800s rtkit-daemon.service
13.662s packagekit.service
13.353s wpa_supplicant.service
13.178s rsyslog.service
12.788s systemd-logind.service
12.313s e2scrub_reap.service
11.105s ModemManager.service
11.003s gpm.service
10.018s networking.service
 6.608s apparmor.service
 5.858s exim4.service

Thanks.

Time to experience with ZRAM :-).

- Sedat -

LINK: https://wiki.debian.org/Swap
LINK: https://help.ubuntu.com/community/SwapFaq

mount -t auto /dev/sdb1 /mnt/sandisk

fallocate -l 8g /mnt/sandisk/swapfile

8 x 1024 x 1024 = 8388608

dd if=/dev/zero bs=1024 count=8388608 of=/mnt/sandisk/swapfile
8388608+0 Datensätze ein
8388608+0 Datensätze aus
8589934592 bytes (8,6 GB, 8,0 GiB) copied, 176,511 s, 48,7 MB/s

# chmod 600 /mnt/sandisk/swapfile

# ll /mnt/sandisk/swapfile
-rw------- 1 root root 8,0G Mai 29 12:23 /mnt/sandisk/swapfile

# mkswap /mnt/sandisk/swapfile
Setting up swapspace version 1, size = 8 GiB (8589930496 bytes)
no label, UUID=d3b72e81-c0fc-49fa-9704-cbbaba3822fc

# swapon /mnt/sandisk/swapfile

# free -h
              total        used        free      shared  buff/cache   available
Mem:          7,7Gi       1,6Gi       4,8Gi       167Mi       1,3Gi       5,7Gi
Swap:         8,0Gi          0B       8,0Gi

- EOT -

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

> Time to experience with ZRAM :-).

I switched over from swap-file to zramswap.

And I can definitely say, my last kernel w/o your patchset does not
show the symptoms.

# cat systemd-analyze-time_5.7.0-rc7-2-amd64-clang_2nd-try.txt
Startup finished in 6.129s (kernel) + 44.192s (userspace) = 50.322s
graphical.target reached after 44.168s in userspace

# cat systemd-analyze-blame_5.7.0-rc7-2-amd64-clang_2nd-try.txt
24.050s udisks2.service
23.711s accounts-daemon.service
18.615s dev-sdc2.device
17.119s polkit.service
16.980s avahi-daemon.service
16.879s NetworkManager.service
16.112s rtkit-daemon.service
15.126s switcheroo-control.service
15.117s wpa_supplicant.service
15.105s systemd-logind.service
14.475s NetworkManager-wait-online.service
14.258s smartmontools.service
13.161s zramswap.service
 9.522s rsyslog.service
 8.337s gpm.service
 6.026s packagekit.service
 5.871s ModemManager.service
 5.746s networking.service
 5.383s e2scrub_reap.service
 3.960s systemd-udevd.service
 3.396s apparmor.service
 3.231s exim4.service
 2.795s systemd-journal-flush.service
 2.359s alsa-restore.service
 2.186s systemd-rfkill.service
 1.878s atd.service
 1.164s keyboard-setup.service
 1.098s bluetooth.service
 1.089s systemd-tmpfiles-setup.service
 1.021s pppd-dns.service
  968ms systemd-backlight@backlight:intel_backlight.service
  964ms upower.service
  937ms binfmt-support.service
  873ms systemd-modules-load.service
  849ms systemd-sysusers.service
  845ms systemd-journald.service
  683ms systemd-timesyncd.service
  676ms modprobe@drm.service
  641ms systemd-udev-trigger.service
  620ms dev-hugepages.mount
  618ms dev-mqueue.mount
  618ms sys-kernel-debug.mount
  617ms sys-kernel-tracing.mount
  501ms ifupdown-wait-online.service
  434ms systemd-sysctl.service
  419ms systemd-random-seed.service
  413ms systemd-tmpfiles-setup-dev.service
  405ms user@1000.service
  389ms systemd-remount-fs.service
  383ms console-setup.service
  301ms kmod-static-nodes.service
  181ms proc-sys-fs-binfmt_misc.mount
  174ms systemd-update-utmp.service
   85ms systemd-user-sessions.service
   22ms user-runtime-dir@1000.service
   19ms systemd-update-utmp-runlevel.service
    5ms ifupdown-pre.service
    4ms sys-fs-fuse-connections.mount

[ /etc/zramswap.conf ]

ZRAM_SIZE_PERCENT=20
# ZSTD support for ZRAM and ZSWAP requires Linux >= 5.7-rc+. -dileks
ZRAM_COMPRESSION_ALGO=zstd

[ /etc/fstab ]

# /etc/fstab: static file system information.
#
# Use 'blkid' to print the universally unique identifier for a
# device; this may be used with UUID= as a more robust way to name devices
# that works even if disks are added and removed. See fstab(5).
#
# <file system> || <mount point> || <type> || <options> || <dump> || <pass>
#
# Root-FS (here: /dev/sdc2)
UUID=<myUUID>       /       ext4    errors=remount-ro       0       1
# SWAP (here: /dev/zram0)
# Zram-based swap (compressed RAM block devices), for details see
zramswap.service and zramswap.conf

# dmesg | egrep 'zram|zswap'
[    1.041958] zswap: loaded using pool zstd/zbud
[   29.569355] zram: Added device: zram0
[   29.581631] zram0: detected capacity change from 0 to 1647824896
[   30.562279] Adding 1609200k swap on /dev/zram0.  Priority:100
extents:1 across:1609200k SSFS

# cat /sys/devices/virtual/block/zram0/comp_algorithm
lzo lzo-rle lz4 lz4hc [zstd]

# swapon --show
NAME       TYPE      SIZE USED PRIO
/dev/zram0 partition 1,5G   0B  100

# cat /proc/swaps
Filename                                Type            Size    Used    Priority
/dev/zram0                              partition       1609200 0       100

If you have any ideas let me know.

Ah, I see there is async-buffered.6.

- Sedat -

[1] https://aur.archlinux.org/packages/zramswap/
[2] https://aur.archlinux.org/cgit/aur.git/tree/zramswap.conf?h=zramswap
[3] https://aur.archlinux.org/cgit/aur.git/tree/zramswap.service?h=zramswap

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

Here are the numbers with your patchset:

# cat systemd-analyze-time_5.7.0-rc7-4-amd64-clang_2nd-try.txt
Startup finished in 7.229s (kernel) + 1min 18.304s (userspace) = 1min 25.534s
graphical.target reached after 1min 18.286s in userspace

# cat systemd-analyze-blame_5.7.0-rc7-4-amd64-clang_2nd-try.txt
53.727s udisks2.service
47.043s accounts-daemon.service
40.252s NetworkManager.service
38.916s polkit.service
38.730s smartmontools.service
36.970s zramswap.service
35.556s dev-sdc2.device
30.307s switcheroo-control.service
29.179s wpa_supplicant.service
29.051s avahi-daemon.service
28.817s rtkit-daemon.service
27.697s systemd-logind.service
23.913s rsyslog.service
22.616s NetworkManager-wait-online.service
22.357s e2scrub_reap.service
19.579s gpm.service
14.879s ModemManager.service
14.126s packagekit.service
14.047s networking.service
 6.137s alsa-restore.service
 5.330s exim4.service
 4.331s systemd-udevd.service
 3.954s apparmor.service
 2.606s atd.service
 2.342s modprobe@drm.service
 2.046s systemd-journal-flush.service
 1.519s bluetooth.service
 1.459s systemd-journald.service
 1.386s systemd-udev-trigger.service
 1.271s systemd-modules-load.service
 1.210s keyboard-setup.service
 1.136s systemd-sysusers.service
  930ms upower.service
  896ms pppd-dns.service
  826ms systemd-tmpfiles-setup.service
  807ms dev-hugepages.mount
  807ms dev-mqueue.mount
  806ms sys-kernel-debug.mount
  806ms sys-kernel-tracing.mount
  712ms ifupdown-wait-online.service
  617ms systemd-remount-fs.service
  588ms systemd-timesyncd.service
  509ms binfmt-support.service
  506ms systemd-backlight@backlight:intel_backlight.service
  497ms systemd-random-seed.service
  490ms systemd-rfkill.service
  376ms user@1000.service
  343ms systemd-tmpfiles-setup-dev.service
  305ms console-setup.service
  303ms systemd-update-utmp.service
  300ms systemd-user-sessions.service
  295ms kmod-static-nodes.service
  267ms systemd-sysctl.service
   71ms proc-sys-fs-binfmt_misc.mount
   19ms user-runtime-dir@1000.service
   13ms systemd-update-utmp-runlevel.service
    5ms sys-fs-fuse-connections.mount
    4ms ifupdown-pre.service

[ diffconfig ]

$ scripts/diffconfig /boot/config-5.7.0-rc7-2-amd64-clang
/boot/config-5.7.0-rc7-4-amd64-clang
 BUILD_SALT "5.7.0-rc7-2-amd64-clang" -> "5.7.0-rc7-4-amd64-clang"
+BLK_INLINE_ENCRYPTION n

- Sedat -


- Sedat -

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 5/30/20 12:57 PM, Sedat Dilek wrote:
> Here are the numbers with your patchset:
> 
> # cat systemd-analyze-time_5.7.0-rc7-4-amd64-clang_2nd-try.txt
> Startup finished in 7.229s (kernel) + 1min 18.304s (userspace) = 1min 25.534s
> graphical.target reached after 1min 18.286s in userspace

Can you see if this makes a difference?

diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index c296463c15eb..ccb895f911b1 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -374,8 +374,7 @@ enum req_flag_bits {
 #define REQ_INTEGRITY		(1ULL << __REQ_INTEGRITY)
 #define REQ_FUA			(1ULL << __REQ_FUA)
 #define REQ_PREFLUSH		(1ULL << __REQ_PREFLUSH)
-#define REQ_RAHEAD		\
-	((1ULL << __REQ_RAHEAD) | (1ULL << __REQ_NOWAIT))
+#define REQ_RAHEAD		(1ULL << __REQ_RAHEAD)
 #define REQ_BACKGROUND		(1ULL << __REQ_BACKGROUND)
 #define REQ_NOWAIT		(1ULL << __REQ_NOWAIT)
 #define REQ_CGROUP_PUNT		(1ULL << __REQ_CGROUP_PUNT)

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

On Sun, May 31, 2020 at 9:04 AM Sedat Dilek <sedat.dilek@gmail.com> wrote:
>
> On Sun, May 31, 2020 at 3:57 AM Jens Axboe <axboe@kernel.dk> wrote:
> >
> > On 5/30/20 12:57 PM, Sedat Dilek wrote:
> > > Here are the numbers with your patchset:
> > >
> > > # cat systemd-analyze-time_5.7.0-rc7-4-amd64-clang_2nd-try.txt
> > > Startup finished in 7.229s (kernel) + 1min 18.304s (userspace) = 1min 25.534s
> > > graphical.target reached after 1min 18.286s in userspace
> >
> > Can you see if this makes a difference?
> >
> > diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
> > index c296463c15eb..ccb895f911b1 100644
> > --- a/include/linux/blk_types.h
> > +++ b/include/linux/blk_types.h
> > @@ -374,8 +374,7 @@ enum req_flag_bits {
> >  #define REQ_INTEGRITY          (1ULL << __REQ_INTEGRITY)
> >  #define REQ_FUA                        (1ULL << __REQ_FUA)
> >  #define REQ_PREFLUSH           (1ULL << __REQ_PREFLUSH)
> > -#define REQ_RAHEAD             \
> > -       ((1ULL << __REQ_RAHEAD) | (1ULL << __REQ_NOWAIT))
> > +#define REQ_RAHEAD             (1ULL << __REQ_RAHEAD)
> >  #define REQ_BACKGROUND         (1ULL << __REQ_BACKGROUND)
> >  #define REQ_NOWAIT             (1ULL << __REQ_NOWAIT)
> >  #define REQ_CGROUP_PUNT                (1ULL << __REQ_CGROUP_PUNT)
> >
>
> Looks good!
>
> With your patch I now get...
>
> # cat systemd-analyze-time_5.7.0-rc7-6-amd64-clang.txt
> Startup finished in 6.199s (kernel) + 45.143s (userspace) = 51.343s
> graphical.target reached after 45.123s in userspace
>
> # cat systemd-analyze-blame_5.7.0-rc7-6-amd64-clang.txt
> 25.029s udisks2.service
> 24.582s accounts-daemon.service
> 19.257s dev-sdc2.device
> 18.016s polkit.service
> 17.857s avahi-daemon.service
> 17.730s NetworkManager.service
> 17.433s rtkit-daemon.service
> 16.558s switcheroo-control.service
> 16.547s wpa_supplicant.service
> 16.538s systemd-logind.service
> 16.078s smartmontools.service
> 14.982s fwupd-refresh.service
> 14.580s NetworkManager-wait-online.service
> 13.630s zramswap.service
> 11.286s fwupd.service
>  8.153s rsyslog.service
>  6.858s gpm.service
>  6.835s e2scrub_reap.service
>  6.449s ModemManager.service
>  6.439s networking.service
>  5.881s packagekit.service
>  3.696s systemd-udevd.service
>  3.322s apparmor.service
>  3.277s exim4.service
>  2.910s alsa-restore.service
>  1.611s systemd-tmpfiles-setup.service
>  1.540s bluetooth.service
>  1.448s systemd-journal-flush.service
>  1.353s keyboard-setup.service
>  1.322s atd.service
>  1.239s systemd-modules-load.service
>  1.216s binfmt-support.service
>  1.060s modprobe@drm.service
>   994ms systemd-journald.service
>   983ms upower.service
>   937ms systemd-sysusers.service
>   914ms ifupdown-wait-online.service
>   904ms pppd-dns.service
>   710ms systemd-udev-trigger.service
>   666ms dev-hugepages.mount
>   650ms dev-mqueue.mount
>   649ms sys-kernel-debug.mount
>   647ms sys-kernel-tracing.mount
>   607ms console-setup.service
>   590ms systemd-tmpfiles-setup-dev.service
>   570ms systemd-timesyncd.service
>   508ms systemd-random-seed.service
>   461ms systemd-backlight@backlight:intel_backlight.service
>   433ms user@1000.service
>   422ms systemd-remount-fs.service
>   365ms systemd-sysctl.service
>   279ms kmod-static-nodes.service
>   246ms proc-sys-fs-binfmt_misc.mount
>   202ms systemd-rfkill.service
>   130ms systemd-user-sessions.service
>   104ms systemd-update-utmp.service
>    25ms user-runtime-dir@1000.service
>    14ms systemd-update-utmp-runlevel.service
>     6ms sys-fs-fuse-connections.mount
>     4ms ifupdown-pre.service
>
> # cat systemd-analyze-critical-chain_5.7.0-rc7-6-amd64-clang.txt
> The time when unit became active or started is printed after the "@" character.
> The time the unit took to start is printed after the "+" character.
>
> graphical.target @45.123s
> └─multi-user.target @45.122s
>   └─exim4.service @41.842s +3.277s
>     └─network-online.target @41.840s
>       └─NetworkManager-wait-online.service @27.259s +14.580s
>         └─NetworkManager.service @9.522s +17.730s
>           └─dbus.service @9.514s
>             └─basic.target @9.328s
>               └─sockets.target @9.327s
>                 └─dbus.socket @9.327s
>                   └─sysinit.target @9.202s
>
> └─systemd-backlight@backlight:intel_backlight.service @21.927s +461ms
>                       └─system-systemd\x2dbacklight.slice @21.925s
>                         └─system.slice @2.529s
>                           └─-.slice @2.529s
>
> Will you send a separate patch on this?
>
> Feel free to add:
> Reported-by: Sedat Dilek <sedat.dilek@gmail.com>
> Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
>
> My kernel-config and dmesg-output are attached.
>
> Thanks Jens!
>

2nd-try (reboot into system with activated WLAN before startup)...

# systemd-analyze --no-pager time
Startup finished in 6.044s (kernel) + 41.370s (userspace) = 47.414s
graphical.target reached after 40.431s in userspace

# systemd-analyze --no-pager blame | head -20
24.674s udisks2.service
24.348s accounts-daemon.service
19.504s dev-sdc2.device
18.677s polkit.service
18.283s avahi-daemon.service
18.231s NetworkManager.service
17.918s rtkit-daemon.service
16.948s switcheroo-control.service
16.932s wpa_supplicant.service
16.928s systemd-logind.service
15.335s smartmontools.service
14.063s zramswap.service
11.507s NetworkManager-wait-online.service
11.279s rsyslog.service
 7.490s gpm.service
 6.556s packagekit.service
 6.506s networking.service
 6.446s e2scrub_reap.service
 5.404s ModemManager.service
 3.329s systemd-udevd.service

# systemd-analyze critical-chain
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.

graphical.target @40.431s
└─multi-user.target @40.431s
  └─exim4.service @38.726s +1.701s
    └─network-online.target @38.723s
      └─NetworkManager-wait-online.service @27.215s +11.507s
        └─NetworkManager.service @8.978s +18.231s
          └─dbus.service @8.971s
            └─basic.target @8.862s
              └─sockets.target @8.861s
                └─dbus.socket @8.861s
                  └─sysinit.target @8.723s

└─systemd-backlight@backlight:intel_backlight.service @20.555s +812ms
                      └─system-systemd\x2dbacklight.slice @19.665s
                        └─system.slice @2.473s
                          └─-.slice @2.473s

Nice!

- Sedat -

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

Hi Jens,

with Linux v5.7 final I switched to linux-block.git/for-next and reverted...

"block: read-ahead submission should imply no-wait as well"

...and see no boot-slowdowns.

Regards,
- Sedat -

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

[-- Attachment #1: Type: text/plain, Size: 355 bytes --]

On 6/1/20 7:35 AM, Sedat Dilek wrote:
> Hi Jens,
> 
> with Linux v5.7 final I switched to linux-block.git/for-next and reverted...
> 
> "block: read-ahead submission should imply no-wait as well"
> 
> ...and see no boot-slowdowns.

Can you try with these patches applied instead? Or pull my async-readahead
branch from the same location.

-- 
Jens Axboe


[-- Attachment #2: 0006-Revert-block-read-ahead-submission-should-imply-no-w.patch --]
[-- Type: text/x-patch, Size: 1014 bytes --]

From 297f4d794780f7f55180fb0de6692bd1a1d81c58 Mon Sep 17 00:00:00 2001
From: Jens Axboe <axboe@kernel.dk>
Date: Sun, 31 May 2020 21:31:08 -0600
Subject: [PATCH 6/6] Revert "block: read-ahead submission should imply no-wait
 as well"

This reverts commit 381000ebbf4fed96c206571a04511c0667e38e18.
---
 include/linux/blk_types.h | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index c296463c15eb..ccb895f911b1 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -374,8 +374,7 @@ enum req_flag_bits {
 #define REQ_INTEGRITY		(1ULL << __REQ_INTEGRITY)
 #define REQ_FUA			(1ULL << __REQ_FUA)
 #define REQ_PREFLUSH		(1ULL << __REQ_PREFLUSH)
-#define REQ_RAHEAD		\
-	((1ULL << __REQ_RAHEAD) | (1ULL << __REQ_NOWAIT))
+#define REQ_RAHEAD		(1ULL << __REQ_RAHEAD)
 #define REQ_BACKGROUND		(1ULL << __REQ_BACKGROUND)
 #define REQ_NOWAIT		(1ULL << __REQ_NOWAIT)
 #define REQ_CGROUP_PUNT		(1ULL << __REQ_CGROUP_PUNT)
-- 
2.26.2


[-- Attachment #3: 0005-fs-make-mpage_readpages-take-a-struct-kiocb-argument.patch --]
[-- Type: text/x-patch, Size: 10724 bytes --]

From ffbf3c623a4835c388296d52d82ee7abe75535cf Mon Sep 17 00:00:00 2001
From: Jens Axboe <axboe@kernel.dk>
Date: Sun, 31 May 2020 20:53:26 -0600
Subject: [PATCH 5/6] fs: make mpage_readpages() take a struct kiocb argument

The callers already have it, so just pass it in. Have mpage_readpages()
set REQ_NOWAIT, if IOCB_NOWAIT is set in the iocb.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/block_dev.c        |  2 +-
 fs/exfat/inode.c      |  2 +-
 fs/ext2/inode.c       |  2 +-
 fs/fat/inode.c        |  2 +-
 fs/gfs2/aops.c        |  2 +-
 fs/hpfs/file.c        |  2 +-
 fs/isofs/inode.c      |  2 +-
 fs/jfs/inode.c        |  2 +-
 fs/mpage.c            | 17 +++++++++++++----
 fs/nilfs2/inode.c     |  2 +-
 fs/ocfs2/aops.c       |  2 +-
 fs/omfs/file.c        |  2 +-
 fs/qnx6/inode.c       |  2 +-
 fs/reiserfs/inode.c   |  2 +-
 fs/udf/inode.c        |  2 +-
 include/linux/mpage.h |  5 +++--
 16 files changed, 30 insertions(+), 20 deletions(-)

diff --git a/fs/block_dev.c b/fs/block_dev.c
index 87a540b42fa7..bbc1bd8078d6 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -617,7 +617,7 @@ static int blkdev_readpage(struct file * file, struct page * page)
 static int blkdev_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, blkdev_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, blkdev_get_block);
 }
 
 static int blkdev_write_begin(struct file *file, struct address_space *mapping,
diff --git a/fs/exfat/inode.c b/fs/exfat/inode.c
index 9905307dfba2..a7757edf5534 100644
--- a/fs/exfat/inode.c
+++ b/fs/exfat/inode.c
@@ -375,7 +375,7 @@ static int exfat_readpage(struct file *file, struct page *page)
 static int exfat_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned int nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, exfat_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, exfat_get_block);
 }
 
 static int exfat_writepage(struct page *page, struct writeback_control *wbc)
diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c
index 3843900d6251..706fc2628a3a 100644
--- a/fs/ext2/inode.c
+++ b/fs/ext2/inode.c
@@ -881,7 +881,7 @@ static int
 ext2_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, ext2_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, ext2_get_block);
 }
 
 static int
diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index ab9dfcbf1bd3..3c47485d7f83 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -213,7 +213,7 @@ static int fat_readpage(struct file *file, struct page *page)
 static int fat_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			 struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, fat_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, fat_get_block);
 }
 
 static void fat_write_failed(struct address_space *mapping, loff_t to)
diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c
index 37d18f82e5a1..3e26ca463ca3 100644
--- a/fs/gfs2/aops.c
+++ b/fs/gfs2/aops.c
@@ -608,7 +608,7 @@ static int gfs2_readpages(struct kiocb *kiocb, struct address_space *mapping,
 	if (unlikely(ret))
 		goto out_uninit;
 	if (!gfs2_is_stuffed(ip))
-		ret = mpage_readpages(mapping, pages, nr_pages, gfs2_block_map);
+		ret = mpage_readpages(mapping, kiocb, pages, nr_pages, gfs2_block_map);
 	gfs2_glock_dq(&gh);
 out_uninit:
 	gfs2_holder_uninit(&gh);
diff --git a/fs/hpfs/file.c b/fs/hpfs/file.c
index 5b9c537d011c..21df02964219 100644
--- a/fs/hpfs/file.c
+++ b/fs/hpfs/file.c
@@ -128,7 +128,7 @@ static int hpfs_writepage(struct page *page, struct writeback_control *wbc)
 static int hpfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			  struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, hpfs_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, hpfs_get_block);
 }
 
 static int hpfs_writepages(struct address_space *mapping,
diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c
index 071da59b7266..060e8dfbe1a2 100644
--- a/fs/isofs/inode.c
+++ b/fs/isofs/inode.c
@@ -1186,7 +1186,7 @@ static int isofs_readpage(struct file *file, struct page *page)
 static int isofs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, isofs_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, isofs_get_block);
 }
 
 static sector_t _isofs_bmap(struct address_space *mapping, sector_t block)
diff --git a/fs/jfs/inode.c b/fs/jfs/inode.c
index 503e2e5cb79d..569e4f4123dd 100644
--- a/fs/jfs/inode.c
+++ b/fs/jfs/inode.c
@@ -299,7 +299,7 @@ static int jfs_readpage(struct file *file, struct page *page)
 static int jfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, jfs_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, jfs_get_block);
 }
 
 static void jfs_write_failed(struct address_space *mapping, loff_t to)
diff --git a/fs/mpage.c b/fs/mpage.c
index ccba3c4c4479..9148d1eae8ff 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -138,6 +138,7 @@ struct mpage_readpage_args {
 	struct page *page;
 	unsigned int nr_pages;
 	bool is_readahead;
+	bool nowait;
 	sector_t last_block_in_bio;
 	struct buffer_head map_bh;
 	unsigned long first_logical_block;
@@ -182,6 +183,8 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 		op_flags = 0;
 		gfp = mapping_gfp_constraint(page->mapping, GFP_KERNEL);
 	}
+	if (args->nowait)
+		op_flags |= REQ_NOWAIT;
 
 	if (page_has_buffers(page))
 		goto confused;
@@ -382,12 +385,14 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
  * This all causes the disk requests to be issued in the correct order.
  */
 int
-mpage_readpages(struct address_space *mapping, struct list_head *pages,
-				unsigned nr_pages, get_block_t get_block)
+mpage_readpages(struct address_space *mapping, struct kiocb *kiocb,
+		struct list_head *pages, unsigned nr_pages,
+		get_block_t get_block)
 {
 	struct mpage_readpage_args args = {
 		.get_block = get_block,
 		.is_readahead = true,
+		.nowait = kiocb->ki_flags & IOCB_NOWAIT,
 	};
 	unsigned page_idx;
 
@@ -406,8 +411,12 @@ mpage_readpages(struct address_space *mapping, struct list_head *pages,
 		put_page(page);
 	}
 	BUG_ON(!list_empty(pages));
-	if (args.bio)
-		mpage_bio_submit(REQ_OP_READ, REQ_RAHEAD, args.bio);
+	if (args.bio) {
+		int op = REQ_RAHEAD;
+		if (args.nowait)
+			op |= REQ_NOWAIT;
+		mpage_bio_submit(REQ_OP_READ, op, args.bio);
+	}
 	return 0;
 }
 EXPORT_SYMBOL(mpage_readpages);
diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c
index a981c9404fbc..3cbc8b7284a4 100644
--- a/fs/nilfs2/inode.c
+++ b/fs/nilfs2/inode.c
@@ -156,7 +156,7 @@ static int nilfs_readpage(struct file *file, struct page *page)
 static int nilfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			   struct list_head *pages, unsigned int nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, nilfs_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, nilfs_get_block);
 }
 
 static int nilfs_writepages(struct address_space *mapping,
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index aa54b0d0d4a1..50ed84f53b20 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -388,7 +388,7 @@ static int ocfs2_readpages(struct kiocb *kiocb, struct address_space *mapping,
 	if (start >= i_size_read(inode))
 		goto out_unlock;
 
-	err = mpage_readpages(mapping, pages, nr_pages, ocfs2_get_block);
+	err = mpage_readpages(mapping, kiocb, pages, nr_pages, ocfs2_get_block);
 
 out_unlock:
 	up_read(&oi->ip_alloc_sem);
diff --git a/fs/omfs/file.c b/fs/omfs/file.c
index 195bb390ba24..384e8c99fc4f 100644
--- a/fs/omfs/file.c
+++ b/fs/omfs/file.c
@@ -292,7 +292,7 @@ static int omfs_readpage(struct file *file, struct page *page)
 static int omfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, omfs_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, omfs_get_block);
 }
 
 static int omfs_writepage(struct page *page, struct writeback_control *wbc)
diff --git a/fs/qnx6/inode.c b/fs/qnx6/inode.c
index c964fd63d4a5..ef603636ca2d 100644
--- a/fs/qnx6/inode.c
+++ b/fs/qnx6/inode.c
@@ -102,7 +102,7 @@ static int qnx6_readpage(struct file *file, struct page *page)
 static int qnx6_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		   struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, qnx6_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, qnx6_get_block);
 }
 
 /*
diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
index 01d5ccdaa5b4..50c93c526557 100644
--- a/fs/reiserfs/inode.c
+++ b/fs/reiserfs/inode.c
@@ -1164,7 +1164,7 @@ static int
 reiserfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		   struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, reiserfs_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, reiserfs_get_block);
 }
 
 /*
diff --git a/fs/udf/inode.c b/fs/udf/inode.c
index c541ca95f851..c45d4da5e707 100644
--- a/fs/udf/inode.c
+++ b/fs/udf/inode.c
@@ -198,7 +198,7 @@ static int udf_readpage(struct file *file, struct page *page)
 static int udf_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			struct list_head *pages, unsigned nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, udf_get_block);
+	return mpage_readpages(mapping, kiocb, pages, nr_pages, udf_get_block);
 }
 
 static int udf_write_begin(struct file *file, struct address_space *mapping,
diff --git a/include/linux/mpage.h b/include/linux/mpage.h
index 001f1fcf9836..f708c7a442ae 100644
--- a/include/linux/mpage.h
+++ b/include/linux/mpage.h
@@ -14,8 +14,9 @@
 
 struct writeback_control;
 
-int mpage_readpages(struct address_space *mapping, struct list_head *pages,
-				unsigned nr_pages, get_block_t get_block);
+int mpage_readpages(struct address_space *mapping, struct kiocb *kiocb,
+			struct list_head *pages, unsigned nr_pages,
+			get_block_t get_block);
 int mpage_readpage(struct page *page, get_block_t get_block);
 int mpage_writepages(struct address_space *mapping,
 		struct writeback_control *wbc, get_block_t get_block);
-- 
2.26.2


[-- Attachment #4: 0004-iomap-set-REQ_NOWAIT-on-bio-if-IOCB_NOWAIT-is-set-in.patch --]
[-- Type: text/x-patch, Size: 3773 bytes --]

From d17311d78f5d5ffa73cbdd5cbf925d0a4188807d Mon Sep 17 00:00:00 2001
From: Jens Axboe <axboe@kernel.dk>
Date: Sun, 31 May 2020 20:44:18 -0600
Subject: [PATCH 4/6] iomap: set REQ_NOWAIT on bio if IOCB_NOWAIT is set in
 kiocb

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/iomap/buffered-io.c | 9 +++++++--
 fs/xfs/xfs_aops.c      | 3 ++-
 fs/zonefs/super.c      | 5 +++--
 include/linux/iomap.h  | 5 +++--
 4 files changed, 15 insertions(+), 7 deletions(-)

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index 89e21961d1ad..4558b455182f 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -215,6 +215,7 @@ struct iomap_readpage_ctx {
 	struct page		*cur_page;
 	bool			cur_page_in_bio;
 	bool			is_readahead;
+	bool			nowait;
 	struct bio		*bio;
 	struct list_head	*pages;
 };
@@ -321,6 +322,8 @@ iomap_readpage_actor(struct inode *inode, loff_t pos, loff_t length, void *data,
 		ctx->bio->bi_opf = REQ_OP_READ;
 		if (ctx->is_readahead)
 			ctx->bio->bi_opf |= REQ_RAHEAD;
+		if (ctx->nowait)
+			ctx->bio->bi_opf |= REQ_NOWAIT;
 		ctx->bio->bi_iter.bi_sector = sector;
 		bio_set_dev(ctx->bio, iomap->bdev);
 		ctx->bio->bi_end_io = iomap_read_end_io;
@@ -432,12 +435,14 @@ iomap_readpages_actor(struct inode *inode, loff_t pos, loff_t length,
 }
 
 int
-iomap_readpages(struct address_space *mapping, struct list_head *pages,
-		unsigned nr_pages, const struct iomap_ops *ops)
+iomap_readpages(struct address_space *mapping, struct kiocb *kiocb,
+		struct list_head *pages, unsigned nr_pages,
+		const struct iomap_ops *ops)
 {
 	struct iomap_readpage_ctx ctx = {
 		.pages		= pages,
 		.is_readahead	= true,
+		.nowait		= kiocb->ki_flags & IOCB_NOWAIT,
 	};
 	loff_t pos = page_offset(list_entry(pages->prev, struct page, lru));
 	loff_t last = page_offset(list_entry(pages->next, struct page, lru));
diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index bff9a2374ece..bc8ea7cd1441 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -628,7 +628,8 @@ xfs_vm_readpages(
 	struct list_head	*pages,
 	unsigned		nr_pages)
 {
-	return iomap_readpages(mapping, pages, nr_pages, &xfs_read_iomap_ops);
+	return iomap_readpages(mapping, kiocb, pages, nr_pages,
+				&xfs_read_iomap_ops);
 }
 
 static int
diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
index 11812841a5fa..06b4f621a14a 100644
--- a/fs/zonefs/super.c
+++ b/fs/zonefs/super.c
@@ -79,10 +79,11 @@ static int zonefs_readpage(struct file *unused, struct page *page)
 	return iomap_readpage(page, &zonefs_iomap_ops);
 }
 
-static int zonefs_readpages(struct kiocb *unused, struct address_space *mapping,
+static int zonefs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			    struct list_head *pages, unsigned int nr_pages)
 {
-	return iomap_readpages(mapping, pages, nr_pages, &zonefs_iomap_ops);
+	return iomap_readpages(mapping, kiocb, pages, nr_pages,
+				&zonefs_iomap_ops);
 }
 
 /*
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index 8b09463dae0d..6b5d802d7227 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -155,8 +155,9 @@ loff_t iomap_apply(struct inode *inode, loff_t pos, loff_t length,
 ssize_t iomap_file_buffered_write(struct kiocb *iocb, struct iov_iter *from,
 		const struct iomap_ops *ops);
 int iomap_readpage(struct page *page, const struct iomap_ops *ops);
-int iomap_readpages(struct address_space *mapping, struct list_head *pages,
-		unsigned nr_pages, const struct iomap_ops *ops);
+int iomap_readpages(struct address_space *mapping, struct kiocb *kiocb,
+		struct list_head *pages, unsigned nr_pages,
+		const struct iomap_ops *ops);
 int iomap_set_page_dirty(struct page *page);
 int iomap_is_partially_uptodate(struct page *page, unsigned long from,
 		unsigned long count);
-- 
2.26.2


[-- Attachment #5: 0003-mm-make-generic_file_buffered_read-use-iocb-read-ahe.patch --]
[-- Type: text/x-patch, Size: 1189 bytes --]

From 3a7284cf4ef7c762ac21ffb9644c04346f0f4a59 Mon Sep 17 00:00:00 2001
From: Jens Axboe <axboe@kernel.dk>
Date: Sun, 31 May 2020 20:26:34 -0600
Subject: [PATCH 3/6] mm: make generic_file_buffered_read() use iocb read-ahead
 helpers

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 mm/filemap.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/mm/filemap.c b/mm/filemap.c
index abdd3f32c932..eea40b5894fb 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -2053,17 +2053,16 @@ static ssize_t generic_file_buffered_read(struct kiocb *iocb,
 
 		page = find_get_page(mapping, index);
 		if (!page) {
-			page_cache_sync_readahead(mapping,
-					ra, filp,
-					index, last_index - index);
+			__page_cache_sync_readahead(mapping, ra, iocb, index,
+							last_index - index);
 			page = find_get_page(mapping, index);
 			if (unlikely(page == NULL))
 				goto no_cached_page;
 		}
 		if (PageReadahead(page)) {
-			page_cache_async_readahead(mapping,
-					ra, filp, page,
-					index, last_index - index);
+			__page_cache_async_readahead(mapping, ra, iocb, page,
+							index,
+							last_index - index);
 		}
 		if (!PageUptodate(page)) {
 			/*
-- 
2.26.2


[-- Attachment #6: 0002-mm-provide-read-ahead-helpers-that-take-a-struct-kio.patch --]
[-- Type: text/x-patch, Size: 4030 bytes --]

From 72dfe63dd44979d182be741d2446d364546c5df5 Mon Sep 17 00:00:00 2001
From: Jens Axboe <axboe@kernel.dk>
Date: Sun, 31 May 2020 20:25:04 -0600
Subject: [PATCH 2/6] mm: provide read-ahead helpers that take a struct kiocb

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 mm/internal.h  | 10 +++++++
 mm/readahead.c | 79 +++++++++++++++++++++++++++++++-------------------
 2 files changed, 59 insertions(+), 30 deletions(-)

diff --git a/mm/internal.h b/mm/internal.h
index 1d051cbadf1a..c486d675af41 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -63,6 +63,16 @@ static inline unsigned long ra_submit(struct file_ra_state *ra,
 					ra->start, ra->size, ra->async_size);
 }
 
+void __page_cache_sync_readahead(struct address_space *mapping,
+				 struct file_ra_state *ra,
+				 struct kiocb *kiocb,  pgoff_t offset,
+				 unsigned long req_size);
+
+void __page_cache_async_readahead(struct address_space *mapping,
+				  struct file_ra_state *ra, struct kiocb *kiocb,
+				  struct page *page, pgoff_t offset,
+				  unsigned long req_size);
+
 /**
  * page_evictable - test whether a page is evictable
  * @page: the page to test
diff --git a/mm/readahead.c b/mm/readahead.c
index 657206f6318d..54a41dae4fea 100644
--- a/mm/readahead.c
+++ b/mm/readahead.c
@@ -493,6 +493,29 @@ ondemand_readahead(struct address_space *mapping,
 	return ra_submit(ra, mapping, kiocb);
 }
 
+void __page_cache_sync_readahead(struct address_space *mapping,
+				 struct file_ra_state *ra, struct kiocb *kiocb,
+				 pgoff_t offset, unsigned long req_size)
+{
+	struct file *filp = kiocb->ki_filp;
+
+	/* no read-ahead */
+	if (!ra->ra_pages)
+		return;
+
+	if (blk_cgroup_congested())
+		return;
+
+	/* be dumb */
+	if (filp && (filp->f_mode & FMODE_RANDOM)) {
+		force_page_cache_readahead(mapping, kiocb, offset, req_size);
+		return;
+	}
+
+	/* do read-ahead */
+	ondemand_readahead(mapping, ra, kiocb, false, offset, req_size);
+}
+
 /**
  * page_cache_sync_readahead - generic file readahead
  * @mapping: address_space which holds the pagecache and I/O vectors
@@ -513,23 +536,40 @@ void page_cache_sync_readahead(struct address_space *mapping,
 {
 	struct kiocb kiocb = { .ki_filp = filp, };
 
+	__page_cache_sync_readahead(mapping, ra, &kiocb, offset, req_size);
+}
+EXPORT_SYMBOL_GPL(page_cache_sync_readahead);
+
+void
+__page_cache_async_readahead(struct address_space *mapping,
+			     struct file_ra_state *ra, struct kiocb *kiocb,
+			     struct page *page, pgoff_t offset,
+			     unsigned long req_size)
+{
 	/* no read-ahead */
 	if (!ra->ra_pages)
 		return;
 
-	if (blk_cgroup_congested())
+	/*
+	 * Same bit is used for PG_readahead and PG_reclaim.
+	 */
+	if (PageWriteback(page))
 		return;
 
-	/* be dumb */
-	if (filp && (filp->f_mode & FMODE_RANDOM)) {
-		force_page_cache_readahead(mapping, &kiocb, offset, req_size);
+	ClearPageReadahead(page);
+
+	/*
+	 * Defer asynchronous read-ahead on IO congestion.
+	 */
+	if (inode_read_congested(mapping->host))
+		return;
+
+	if (blk_cgroup_congested())
 		return;
-	}
 
 	/* do read-ahead */
-	ondemand_readahead(mapping, ra, &kiocb, false, offset, req_size);
+	ondemand_readahead(mapping, ra, kiocb, true, offset, req_size);
 }
-EXPORT_SYMBOL_GPL(page_cache_sync_readahead);
 
 /**
  * page_cache_async_readahead - file readahead for marked pages
@@ -554,29 +594,8 @@ page_cache_async_readahead(struct address_space *mapping,
 {
 	struct kiocb kiocb = { .ki_filp = filp, };
 
-	/* no read-ahead */
-	if (!ra->ra_pages)
-		return;
-
-	/*
-	 * Same bit is used for PG_readahead and PG_reclaim.
-	 */
-	if (PageWriteback(page))
-		return;
-
-	ClearPageReadahead(page);
-
-	/*
-	 * Defer asynchronous read-ahead on IO congestion.
-	 */
-	if (inode_read_congested(mapping->host))
-		return;
-
-	if (blk_cgroup_congested())
-		return;
-
-	/* do read-ahead */
-	ondemand_readahead(mapping, ra, &kiocb, true, offset, req_size);
+	__page_cache_async_readahead(mapping, ra, &kiocb, page, offset,
+					req_size);
 }
 EXPORT_SYMBOL_GPL(page_cache_async_readahead);
 
-- 
2.26.2


[-- Attachment #7: 0001-fs-make-aops-readpages-take-kiocb-argument.patch --]
[-- Type: text/x-patch, Size: 28302 bytes --]

From 707258c22fea35a5915ff0373231fb07a8a1c9eb Mon Sep 17 00:00:00 2001
From: Jens Axboe <axboe@kernel.dk>
Date: Sun, 31 May 2020 19:35:33 -0600
Subject: [PATCH 1/6] fs: make aops->readpages() take kiocb argument

Instead of passing in a file, pass in the kiocb instead. This still
provides access to the file through kiocb->ki_filp, but also provides
access to more information related to the IO. We'll be using that to
ensure that ->readpages() understands IOCB_NOWAIT.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/9p/vfs_addr.c       |  3 ++-
 fs/afs/file.c          |  8 ++++----
 fs/block_dev.c         |  2 +-
 fs/btrfs/inode.c       |  2 +-
 fs/ceph/addr.c         |  3 ++-
 fs/cifs/file.c         |  3 ++-
 fs/erofs/data.c        |  2 +-
 fs/erofs/zdata.c       |  2 +-
 fs/exfat/inode.c       |  2 +-
 fs/ext2/inode.c        |  2 +-
 fs/ext4/inode.c        |  2 +-
 fs/f2fs/data.c         |  2 +-
 fs/fat/inode.c         |  2 +-
 fs/fuse/file.c         |  6 +++---
 fs/gfs2/aops.c         |  2 +-
 fs/hpfs/file.c         |  2 +-
 fs/isofs/inode.c       |  2 +-
 fs/jfs/inode.c         |  2 +-
 fs/nfs/read.c          |  3 ++-
 fs/nilfs2/inode.c      |  2 +-
 fs/ocfs2/aops.c        |  2 +-
 fs/omfs/file.c         |  2 +-
 fs/qnx6/inode.c        |  2 +-
 fs/reiserfs/inode.c    |  2 +-
 fs/udf/inode.c         |  2 +-
 fs/xfs/xfs_aops.c      |  2 +-
 fs/zonefs/super.c      |  2 +-
 include/linux/fs.h     |  2 +-
 include/linux/mm.h     |  6 +++---
 include/linux/nfs_fs.h |  2 +-
 mm/fadvise.c           |  7 +++++--
 mm/filemap.c           |  3 ++-
 mm/internal.h          |  6 +++---
 mm/readahead.c         | 39 ++++++++++++++++++++++-----------------
 34 files changed, 73 insertions(+), 60 deletions(-)

diff --git a/fs/9p/vfs_addr.c b/fs/9p/vfs_addr.c
index cce9ace651a2..f2980829bb9f 100644
--- a/fs/9p/vfs_addr.c
+++ b/fs/9p/vfs_addr.c
@@ -95,9 +95,10 @@ static int v9fs_vfs_readpage(struct file *filp, struct page *page)
  *
  */
 
-static int v9fs_vfs_readpages(struct file *filp, struct address_space *mapping,
+static int v9fs_vfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			     struct list_head *pages, unsigned nr_pages)
 {
+	struct file *filp = kiocb->ki_filp;
 	int ret = 0;
 	struct inode *inode;
 
diff --git a/fs/afs/file.c b/fs/afs/file.c
index 8415733f7bc1..447b6524c2e9 100644
--- a/fs/afs/file.c
+++ b/fs/afs/file.c
@@ -22,7 +22,7 @@ static void afs_invalidatepage(struct page *page, unsigned int offset,
 			       unsigned int length);
 static int afs_releasepage(struct page *page, gfp_t gfp_flags);
 
-static int afs_readpages(struct file *filp, struct address_space *mapping,
+static int afs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			 struct list_head *pages, unsigned nr_pages);
 
 const struct file_operations afs_file_operations = {
@@ -538,10 +538,10 @@ static int afs_readpages_one(struct file *file, struct address_space *mapping,
 /*
  * read a set of pages
  */
-static int afs_readpages(struct file *file, struct address_space *mapping,
+static int afs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			 struct list_head *pages, unsigned nr_pages)
 {
-	struct key *key = afs_file_key(file);
+	struct key *key = afs_file_key(kiocb->ki_filp);
 	struct afs_vnode *vnode;
 	int ret = 0;
 
@@ -589,7 +589,7 @@ static int afs_readpages(struct file *file, struct address_space *mapping,
 	}
 
 	while (!list_empty(pages)) {
-		ret = afs_readpages_one(file, mapping, pages);
+		ret = afs_readpages_one(kiocb->ki_filp, mapping, pages);
 		if (ret < 0)
 			break;
 	}
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 980cfce01c9a..87a540b42fa7 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -614,7 +614,7 @@ static int blkdev_readpage(struct file * file, struct page * page)
 	return block_read_full_page(page, blkdev_get_block);
 }
 
-static int blkdev_readpages(struct file *file, struct address_space *mapping,
+static int blkdev_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			struct list_head *pages, unsigned nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, blkdev_get_block);
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 320d1062068d..1569d29fdfc4 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -8294,7 +8294,7 @@ static int btrfs_writepages(struct address_space *mapping,
 }
 
 static int
-btrfs_readpages(struct file *file, struct address_space *mapping,
+btrfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned nr_pages)
 {
 	return extent_readpages(mapping, pages, nr_pages);
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index 6f4678d98df7..687714a29e89 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -477,9 +477,10 @@ static int start_read(struct inode *inode, struct ceph_rw_context *rw_ctx,
  * Read multiple pages.  Leave pages we don't read + unlock in page_list;
  * the caller (VM) cleans them up.
  */
-static int ceph_readpages(struct file *file, struct address_space *mapping,
+static int ceph_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			  struct list_head *page_list, unsigned nr_pages)
 {
+	struct file *file = kiocb->ki_filp;
 	struct inode *inode = file_inode(file);
 	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
 	struct ceph_file_info *fi = file->private_data;
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 75ddce8ef456..5eec4f40811c 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -4344,11 +4344,12 @@ readpages_get_pages(struct address_space *mapping, struct list_head *page_list,
 	return rc;
 }
 
-static int cifs_readpages(struct file *file, struct address_space *mapping,
+static int cifs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 	struct list_head *page_list, unsigned num_pages)
 {
 	int rc;
 	struct list_head tmplist;
+	struct file *file = kiocb->ki_filp;
 	struct cifsFileInfo *open_file = file->private_data;
 	struct cifs_sb_info *cifs_sb = CIFS_FILE_SB(file);
 	struct TCP_Server_Info *server;
diff --git a/fs/erofs/data.c b/fs/erofs/data.c
index fc3a8d8064f8..853b905a15c4 100644
--- a/fs/erofs/data.c
+++ b/fs/erofs/data.c
@@ -280,7 +280,7 @@ static int erofs_raw_access_readpage(struct file *file, struct page *page)
 	return 0;
 }
 
-static int erofs_raw_access_readpages(struct file *filp,
+static int erofs_raw_access_readpages(struct kiocb *kiocb,
 				      struct address_space *mapping,
 				      struct list_head *pages,
 				      unsigned int nr_pages)
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index c4b6c9aa87ec..bc8c02b088ce 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -1305,7 +1305,7 @@ static bool should_decompress_synchronously(struct erofs_sb_info *sbi,
 	return nr <= sbi->max_sync_decompress_pages;
 }
 
-static int z_erofs_readpages(struct file *filp, struct address_space *mapping,
+static int z_erofs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			     struct list_head *pages, unsigned int nr_pages)
 {
 	struct inode *const inode = mapping->host;
diff --git a/fs/exfat/inode.c b/fs/exfat/inode.c
index 06887492f54b..9905307dfba2 100644
--- a/fs/exfat/inode.c
+++ b/fs/exfat/inode.c
@@ -372,7 +372,7 @@ static int exfat_readpage(struct file *file, struct page *page)
 	return mpage_readpage(page, exfat_get_block);
 }
 
-static int exfat_readpages(struct file *file, struct address_space *mapping,
+static int exfat_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned int nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, exfat_get_block);
diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c
index c885cf7d724b..3843900d6251 100644
--- a/fs/ext2/inode.c
+++ b/fs/ext2/inode.c
@@ -878,7 +878,7 @@ static int ext2_readpage(struct file *file, struct page *page)
 }
 
 static int
-ext2_readpages(struct file *file, struct address_space *mapping,
+ext2_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, ext2_get_block);
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 2a4aae6acdcb..6297804d1ac2 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -3231,7 +3231,7 @@ static int ext4_readpage(struct file *file, struct page *page)
 }
 
 static int
-ext4_readpages(struct file *file, struct address_space *mapping,
+ext4_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned nr_pages)
 {
 	struct inode *inode = mapping->host;
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index cdf2f626bea7..7330aa1e2e39 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -2304,7 +2304,7 @@ static int f2fs_read_data_page(struct file *file, struct page *page)
 	return ret;
 }
 
-static int f2fs_read_data_pages(struct file *file,
+static int f2fs_read_data_pages(struct kiocb *kiocb,
 			struct address_space *mapping,
 			struct list_head *pages, unsigned nr_pages)
 {
diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index 71946da84388..ab9dfcbf1bd3 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -210,7 +210,7 @@ static int fat_readpage(struct file *file, struct page *page)
 	return mpage_readpage(page, fat_get_block);
 }
 
-static int fat_readpages(struct file *file, struct address_space *mapping,
+static int fat_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			 struct list_head *pages, unsigned nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, fat_get_block);
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index 9d67b830fb7a..ebd40681ee94 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -962,7 +962,7 @@ static int fuse_readpages_fill(void *_data, struct page *page)
 	return 0;
 }
 
-static int fuse_readpages(struct file *file, struct address_space *mapping,
+static int fuse_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			  struct list_head *pages, unsigned nr_pages)
 {
 	struct inode *inode = mapping->host;
@@ -974,7 +974,7 @@ static int fuse_readpages(struct file *file, struct address_space *mapping,
 	if (is_bad_inode(inode))
 		goto out;
 
-	data.file = file;
+	data.file = kiocb->ki_filp;
 	data.inode = inode;
 	data.nr_pages = nr_pages;
 	data.max_pages = min_t(unsigned int, nr_pages, fc->max_pages);
@@ -987,7 +987,7 @@ static int fuse_readpages(struct file *file, struct address_space *mapping,
 	err = read_cache_pages(mapping, pages, fuse_readpages_fill, &data);
 	if (!err) {
 		if (data.ia->ap.num_pages)
-			fuse_send_readpages(data.ia, file);
+			fuse_send_readpages(data.ia, kiocb->ki_filp);
 		else
 			fuse_io_free(data.ia);
 	}
diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c
index 786c1ce8f030..37d18f82e5a1 100644
--- a/fs/gfs2/aops.c
+++ b/fs/gfs2/aops.c
@@ -594,7 +594,7 @@ int gfs2_internal_read(struct gfs2_inode *ip, char *buf, loff_t *pos,
  * 4. gfs2_block_map() is relied upon to set BH_Boundary in the right places.
  */
 
-static int gfs2_readpages(struct file *file, struct address_space *mapping,
+static int gfs2_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			  struct list_head *pages, unsigned nr_pages)
 {
 	struct inode *inode = mapping->host;
diff --git a/fs/hpfs/file.c b/fs/hpfs/file.c
index b36abf9cb345..5b9c537d011c 100644
--- a/fs/hpfs/file.c
+++ b/fs/hpfs/file.c
@@ -125,7 +125,7 @@ static int hpfs_writepage(struct page *page, struct writeback_control *wbc)
 	return block_write_full_page(page, hpfs_get_block, wbc);
 }
 
-static int hpfs_readpages(struct file *file, struct address_space *mapping,
+static int hpfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			  struct list_head *pages, unsigned nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, hpfs_get_block);
diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c
index 276107cdaaf1..071da59b7266 100644
--- a/fs/isofs/inode.c
+++ b/fs/isofs/inode.c
@@ -1183,7 +1183,7 @@ static int isofs_readpage(struct file *file, struct page *page)
 	return mpage_readpage(page, isofs_get_block);
 }
 
-static int isofs_readpages(struct file *file, struct address_space *mapping,
+static int isofs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			struct list_head *pages, unsigned nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, isofs_get_block);
diff --git a/fs/jfs/inode.c b/fs/jfs/inode.c
index 9486afcdac76..503e2e5cb79d 100644
--- a/fs/jfs/inode.c
+++ b/fs/jfs/inode.c
@@ -296,7 +296,7 @@ static int jfs_readpage(struct file *file, struct page *page)
 	return mpage_readpage(page, jfs_get_block);
 }
 
-static int jfs_readpages(struct file *file, struct address_space *mapping,
+static int jfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, jfs_get_block);
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index 13b22e898116..4719f6b67b5c 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -402,9 +402,10 @@ readpage_async_filler(void *data, struct page *page)
 	return error;
 }
 
-int nfs_readpages(struct file *filp, struct address_space *mapping,
+int nfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned nr_pages)
 {
+	struct file *filp = kiocb->ki_filp;
 	struct nfs_pageio_descriptor pgio;
 	struct nfs_pgio_mirror *pgm;
 	struct nfs_readdesc desc = {
diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c
index 671085512e0f..a981c9404fbc 100644
--- a/fs/nilfs2/inode.c
+++ b/fs/nilfs2/inode.c
@@ -153,7 +153,7 @@ static int nilfs_readpage(struct file *file, struct page *page)
  * @pages - the pages to be read
  * @nr_pages - number of pages to be read
  */
-static int nilfs_readpages(struct file *file, struct address_space *mapping,
+static int nilfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			   struct list_head *pages, unsigned int nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, nilfs_get_block);
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index 3a67a6518ddf..aa54b0d0d4a1 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -350,7 +350,7 @@ static int ocfs2_readpage(struct file *file, struct page *page)
  * grow out to a tree. If need be, detecting boundary extents could
  * trivially be added in a future version of ocfs2_get_block().
  */
-static int ocfs2_readpages(struct file *filp, struct address_space *mapping,
+static int ocfs2_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			   struct list_head *pages, unsigned nr_pages)
 {
 	int ret, err = -EIO;
diff --git a/fs/omfs/file.c b/fs/omfs/file.c
index d640b9388238..195bb390ba24 100644
--- a/fs/omfs/file.c
+++ b/fs/omfs/file.c
@@ -289,7 +289,7 @@ static int omfs_readpage(struct file *file, struct page *page)
 	return block_read_full_page(page, omfs_get_block);
 }
 
-static int omfs_readpages(struct file *file, struct address_space *mapping,
+static int omfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		struct list_head *pages, unsigned nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, omfs_get_block);
diff --git a/fs/qnx6/inode.c b/fs/qnx6/inode.c
index 345db56c98fd..c964fd63d4a5 100644
--- a/fs/qnx6/inode.c
+++ b/fs/qnx6/inode.c
@@ -99,7 +99,7 @@ static int qnx6_readpage(struct file *file, struct page *page)
 	return mpage_readpage(page, qnx6_get_block);
 }
 
-static int qnx6_readpages(struct file *file, struct address_space *mapping,
+static int qnx6_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		   struct list_head *pages, unsigned nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, qnx6_get_block);
diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
index 6419e6dacc39..01d5ccdaa5b4 100644
--- a/fs/reiserfs/inode.c
+++ b/fs/reiserfs/inode.c
@@ -1161,7 +1161,7 @@ int reiserfs_get_block(struct inode *inode, sector_t block,
 }
 
 static int
-reiserfs_readpages(struct file *file, struct address_space *mapping,
+reiserfs_readpages(struct kiocb *kiocb, struct address_space *mapping,
 		   struct list_head *pages, unsigned nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, reiserfs_get_block);
diff --git a/fs/udf/inode.c b/fs/udf/inode.c
index e875bc5668ee..c541ca95f851 100644
--- a/fs/udf/inode.c
+++ b/fs/udf/inode.c
@@ -195,7 +195,7 @@ static int udf_readpage(struct file *file, struct page *page)
 	return mpage_readpage(page, udf_get_block);
 }
 
-static int udf_readpages(struct file *file, struct address_space *mapping,
+static int udf_readpages(struct kiocb *kiocb, struct address_space *mapping,
 			struct list_head *pages, unsigned nr_pages)
 {
 	return mpage_readpages(mapping, pages, nr_pages, udf_get_block);
diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index 9d9cebf18726..bff9a2374ece 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -623,7 +623,7 @@ xfs_vm_readpage(
 
 STATIC int
 xfs_vm_readpages(
-	struct file		*unused,
+	struct kiocb		*kiocb,
 	struct address_space	*mapping,
 	struct list_head	*pages,
 	unsigned		nr_pages)
diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
index 25afcf55aa41..11812841a5fa 100644
--- a/fs/zonefs/super.c
+++ b/fs/zonefs/super.c
@@ -79,7 +79,7 @@ static int zonefs_readpage(struct file *unused, struct page *page)
 	return iomap_readpage(page, &zonefs_iomap_ops);
 }
 
-static int zonefs_readpages(struct file *unused, struct address_space *mapping,
+static int zonefs_readpages(struct kiocb *unused, struct address_space *mapping,
 			    struct list_head *pages, unsigned int nr_pages)
 {
 	return iomap_readpages(mapping, pages, nr_pages, &zonefs_iomap_ops);
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 5ffc6d236b01..cb5adf72041b 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -381,7 +381,7 @@ struct address_space_operations {
 	 * Reads in the requested pages. Unlike ->readpage(), this is
 	 * PURELY used for read-ahead!.
 	 */
-	int (*readpages)(struct file *filp, struct address_space *mapping,
+	int (*readpages)(struct kiocb *kiocb, struct address_space *mapping,
 			struct list_head *pages, unsigned nr_pages);
 
 	int (*write_begin)(struct file *, struct address_space *mapping,
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 5a323422d783..7c805a32fdc7 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2604,18 +2604,18 @@ void task_dirty_inc(struct task_struct *tsk);
 /* readahead.c */
 #define VM_READAHEAD_PAGES	(SZ_128K / PAGE_SIZE)
 
-int force_page_cache_readahead(struct address_space *mapping, struct file *filp,
+int force_page_cache_readahead(struct address_space *mapping, struct kiocb *kiocb,
 			pgoff_t offset, unsigned long nr_to_read);
 
 void page_cache_sync_readahead(struct address_space *mapping,
 			       struct file_ra_state *ra,
-			       struct file *filp,
+			       struct file *file,
 			       pgoff_t offset,
 			       unsigned long size);
 
 void page_cache_async_readahead(struct address_space *mapping,
 				struct file_ra_state *ra,
-				struct file *filp,
+				struct file *file,
 				struct page *pg,
 				pgoff_t offset,
 				unsigned long size);
diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h
index 73eda45f1cfd..27c62a08f499 100644
--- a/include/linux/nfs_fs.h
+++ b/include/linux/nfs_fs.h
@@ -550,7 +550,7 @@ nfs_have_writebacks(struct inode *inode)
  * linux/fs/nfs/read.c
  */
 extern int  nfs_readpage(struct file *, struct page *);
-extern int  nfs_readpages(struct file *, struct address_space *,
+extern int  nfs_readpages(struct kiocb *, struct address_space *,
 		struct list_head *, unsigned);
 extern int  nfs_readpage_async(struct nfs_open_context *, struct inode *,
 			       struct page *);
diff --git a/mm/fadvise.c b/mm/fadvise.c
index 4f17c83db575..58654fd13486 100644
--- a/mm/fadvise.c
+++ b/mm/fadvise.c
@@ -92,7 +92,9 @@ int generic_fadvise(struct file *file, loff_t offset, loff_t len, int advice)
 		file->f_mode &= ~FMODE_RANDOM;
 		spin_unlock(&file->f_lock);
 		break;
-	case POSIX_FADV_WILLNEED:
+	case POSIX_FADV_WILLNEED: {
+		struct kiocb kiocb = { .ki_filp = file, };
+
 		/* First and last PARTIAL page! */
 		start_index = offset >> PAGE_SHIFT;
 		end_index = endbyte >> PAGE_SHIFT;
@@ -106,8 +108,9 @@ int generic_fadvise(struct file *file, loff_t offset, loff_t len, int advice)
 		 * Ignore return value because fadvise() shall return
 		 * success even if filesystem can't retrieve a hint,
 		 */
-		force_page_cache_readahead(mapping, file, start_index, nrpages);
+		force_page_cache_readahead(mapping, &kiocb, start_index, nrpages);
 		break;
+		}
 	case POSIX_FADV_NOREUSE:
 		break;
 	case POSIX_FADV_DONTNEED:
diff --git a/mm/filemap.c b/mm/filemap.c
index 18022de7dc33..abdd3f32c932 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -2403,6 +2403,7 @@ static struct file *do_sync_mmap_readahead(struct vm_fault *vmf)
 	struct address_space *mapping = file->f_mapping;
 	struct file *fpin = NULL;
 	pgoff_t offset = vmf->pgoff;
+	struct kiocb kiocb = { .ki_filp = file, };
 
 	/* If we don't want any read-ahead, don't bother */
 	if (vmf->vma->vm_flags & VM_RAND_READ)
@@ -2435,7 +2436,7 @@ static struct file *do_sync_mmap_readahead(struct vm_fault *vmf)
 	ra->start = max_t(long, 0, offset - ra->ra_pages / 2);
 	ra->size = ra->ra_pages;
 	ra->async_size = ra->ra_pages / 4;
-	ra_submit(ra, mapping, file);
+	ra_submit(ra, mapping, &kiocb);
 	return fpin;
 }
 
diff --git a/mm/internal.h b/mm/internal.h
index b5634e78f01d..1d051cbadf1a 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -50,16 +50,16 @@ void unmap_page_range(struct mmu_gather *tlb,
 			     struct zap_details *details);
 
 extern unsigned int __do_page_cache_readahead(struct address_space *mapping,
-		struct file *filp, pgoff_t offset, unsigned long nr_to_read,
+		struct kiocb *kiocb, pgoff_t offset, unsigned long nr_to_read,
 		unsigned long lookahead_size);
 
 /*
  * Submit IO for the read-ahead request in file_ra_state.
  */
 static inline unsigned long ra_submit(struct file_ra_state *ra,
-		struct address_space *mapping, struct file *filp)
+		struct address_space *mapping, struct kiocb *kiocb)
 {
-	return __do_page_cache_readahead(mapping, filp,
+	return __do_page_cache_readahead(mapping, kiocb,
 					ra->start, ra->size, ra->async_size);
 }
 
diff --git a/mm/readahead.c b/mm/readahead.c
index 2fe72cd29b47..657206f6318d 100644
--- a/mm/readahead.c
+++ b/mm/readahead.c
@@ -113,7 +113,7 @@ int read_cache_pages(struct address_space *mapping, struct list_head *pages,
 
 EXPORT_SYMBOL(read_cache_pages);
 
-static int read_pages(struct address_space *mapping, struct file *filp,
+static int read_pages(struct address_space *mapping, struct kiocb *kiocb,
 		struct list_head *pages, unsigned int nr_pages, gfp_t gfp)
 {
 	struct blk_plug plug;
@@ -123,7 +123,7 @@ static int read_pages(struct address_space *mapping, struct file *filp,
 	blk_start_plug(&plug);
 
 	if (mapping->a_ops->readpages) {
-		ret = mapping->a_ops->readpages(filp, mapping, pages, nr_pages);
+		ret = mapping->a_ops->readpages(kiocb, mapping, pages, nr_pages);
 		/* Clean up the remaining pages */
 		put_pages_list(pages);
 		goto out;
@@ -133,7 +133,7 @@ static int read_pages(struct address_space *mapping, struct file *filp,
 		struct page *page = lru_to_page(pages);
 		list_del(&page->lru);
 		if (!add_to_page_cache_lru(page, mapping, page->index, gfp))
-			mapping->a_ops->readpage(filp, page);
+			mapping->a_ops->readpage(kiocb->ki_filp, page);
 		put_page(page);
 	}
 	ret = 0;
@@ -153,7 +153,7 @@ static int read_pages(struct address_space *mapping, struct file *filp,
  * Returns the number of pages requested, or the maximum amount of I/O allowed.
  */
 unsigned int __do_page_cache_readahead(struct address_space *mapping,
-		struct file *filp, pgoff_t offset, unsigned long nr_to_read,
+		struct kiocb *kiocb, pgoff_t offset, unsigned long nr_to_read,
 		unsigned long lookahead_size)
 {
 	struct inode *inode = mapping->host;
@@ -187,7 +187,7 @@ unsigned int __do_page_cache_readahead(struct address_space *mapping,
 			 * batch.
 			 */
 			if (nr_pages)
-				read_pages(mapping, filp, &page_pool, nr_pages,
+				read_pages(mapping, kiocb, &page_pool, nr_pages,
 						gfp_mask);
 			nr_pages = 0;
 			continue;
@@ -209,7 +209,7 @@ unsigned int __do_page_cache_readahead(struct address_space *mapping,
 	 * will then handle the error.
 	 */
 	if (nr_pages)
-		read_pages(mapping, filp, &page_pool, nr_pages, gfp_mask);
+		read_pages(mapping, kiocb, &page_pool, nr_pages, gfp_mask);
 	BUG_ON(!list_empty(&page_pool));
 out:
 	return nr_pages;
@@ -219,11 +219,12 @@ unsigned int __do_page_cache_readahead(struct address_space *mapping,
  * Chunk the readahead into 2 megabyte units, so that we don't pin too much
  * memory at once.
  */
-int force_page_cache_readahead(struct address_space *mapping, struct file *filp,
-			       pgoff_t offset, unsigned long nr_to_read)
+int force_page_cache_readahead(struct address_space *mapping,
+			       struct kiocb *kiocb, pgoff_t offset,
+			       unsigned long nr_to_read)
 {
 	struct backing_dev_info *bdi = inode_to_bdi(mapping->host);
-	struct file_ra_state *ra = &filp->f_ra;
+	struct file_ra_state *ra = &kiocb->ki_filp->f_ra;
 	unsigned long max_pages;
 
 	if (unlikely(!mapping->a_ops->readpage && !mapping->a_ops->readpages))
@@ -240,7 +241,7 @@ int force_page_cache_readahead(struct address_space *mapping, struct file *filp,
 
 		if (this_chunk > nr_to_read)
 			this_chunk = nr_to_read;
-		__do_page_cache_readahead(mapping, filp, offset, this_chunk, 0);
+		__do_page_cache_readahead(mapping, kiocb, offset, this_chunk, 0);
 
 		offset += this_chunk;
 		nr_to_read -= this_chunk;
@@ -380,7 +381,7 @@ static int try_context_readahead(struct address_space *mapping,
  */
 static unsigned long
 ondemand_readahead(struct address_space *mapping,
-		   struct file_ra_state *ra, struct file *filp,
+		   struct file_ra_state *ra, struct kiocb *kiocb,
 		   bool hit_readahead_marker, pgoff_t offset,
 		   unsigned long req_size)
 {
@@ -464,7 +465,7 @@ ondemand_readahead(struct address_space *mapping,
 	 * standalone, small random read
 	 * Read as is, and do not pollute the readahead state.
 	 */
-	return __do_page_cache_readahead(mapping, filp, offset, req_size, 0);
+	return __do_page_cache_readahead(mapping, kiocb, offset, req_size, 0);
 
 initial_readahead:
 	ra->start = offset;
@@ -489,14 +490,14 @@ ondemand_readahead(struct address_space *mapping,
 		}
 	}
 
-	return ra_submit(ra, mapping, filp);
+	return ra_submit(ra, mapping, kiocb);
 }
 
 /**
  * page_cache_sync_readahead - generic file readahead
  * @mapping: address_space which holds the pagecache and I/O vectors
  * @ra: file_ra_state which holds the readahead state
- * @filp: passed on to ->readpage() and ->readpages()
+ * @kiocb: passed on to ->readpage() and ->readpages()
  * @offset: start offset into @mapping, in pagecache page-sized units
  * @req_size: hint: total size of the read which the caller is performing in
  *            pagecache pages
@@ -510,6 +511,8 @@ void page_cache_sync_readahead(struct address_space *mapping,
 			       struct file_ra_state *ra, struct file *filp,
 			       pgoff_t offset, unsigned long req_size)
 {
+	struct kiocb kiocb = { .ki_filp = filp, };
+
 	/* no read-ahead */
 	if (!ra->ra_pages)
 		return;
@@ -519,12 +522,12 @@ void page_cache_sync_readahead(struct address_space *mapping,
 
 	/* be dumb */
 	if (filp && (filp->f_mode & FMODE_RANDOM)) {
-		force_page_cache_readahead(mapping, filp, offset, req_size);
+		force_page_cache_readahead(mapping, &kiocb, offset, req_size);
 		return;
 	}
 
 	/* do read-ahead */
-	ondemand_readahead(mapping, ra, filp, false, offset, req_size);
+	ondemand_readahead(mapping, ra, &kiocb, false, offset, req_size);
 }
 EXPORT_SYMBOL_GPL(page_cache_sync_readahead);
 
@@ -549,6 +552,8 @@ page_cache_async_readahead(struct address_space *mapping,
 			   struct page *page, pgoff_t offset,
 			   unsigned long req_size)
 {
+	struct kiocb kiocb = { .ki_filp = filp, };
+
 	/* no read-ahead */
 	if (!ra->ra_pages)
 		return;
@@ -571,7 +576,7 @@ page_cache_async_readahead(struct address_space *mapping,
 		return;
 
 	/* do read-ahead */
-	ondemand_readahead(mapping, ra, filp, true, offset, req_size);
+	ondemand_readahead(mapping, ra, &kiocb, true, offset, req_size);
 }
 EXPORT_SYMBOL_GPL(page_cache_async_readahead);
 
-- 
2.26.2

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

On Mon, Jun 1, 2020 at 4:04 PM Jens Axboe <axboe@kernel.dk> wrote:
>
> On 6/1/20 7:35 AM, Sedat Dilek wrote:
> > Hi Jens,
> >
> > with Linux v5.7 final I switched to linux-block.git/for-next and reverted...
> >
> > "block: read-ahead submission should imply no-wait as well"
> >
> > ...and see no boot-slowdowns.
>
> Can you try with these patches applied instead? Or pull my async-readahead
> branch from the same location.
>

Yes, I can do that.
I pulled from linux-block.git#async-readahead and will report later.

Any specific testing desired by you?

- Sedat -

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/1/20 8:13 AM, Sedat Dilek wrote:
> On Mon, Jun 1, 2020 at 4:04 PM Jens Axboe <axboe@kernel.dk> wrote:
>>
>> On 6/1/20 7:35 AM, Sedat Dilek wrote:
>>> Hi Jens,
>>>
>>> with Linux v5.7 final I switched to linux-block.git/for-next and reverted...
>>>
>>> "block: read-ahead submission should imply no-wait as well"
>>>
>>> ...and see no boot-slowdowns.
>>
>> Can you try with these patches applied instead? Or pull my async-readahead
>> branch from the same location.
>>
> 
> Yes, I can do that.
> I pulled from linux-block.git#async-readahead and will report later.
> 
> Any specific testing desired by you?

Just do your boot timing test and see if it works, thanks.

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/1/20 8:14 AM, Jens Axboe wrote:
> On 6/1/20 8:13 AM, Sedat Dilek wrote:
>> On Mon, Jun 1, 2020 at 4:04 PM Jens Axboe <axboe@kernel.dk> wrote:
>>>
>>> On 6/1/20 7:35 AM, Sedat Dilek wrote:
>>>> Hi Jens,
>>>>
>>>> with Linux v5.7 final I switched to linux-block.git/for-next and reverted...
>>>>
>>>> "block: read-ahead submission should imply no-wait as well"
>>>>
>>>> ...and see no boot-slowdowns.
>>>
>>> Can you try with these patches applied instead? Or pull my async-readahead
>>> branch from the same location.
>>>
>>
>> Yes, I can do that.
>> I pulled from linux-block.git#async-readahead and will report later.
>>
>> Any specific testing desired by you?
> 
> Just do your boot timing test and see if it works, thanks.

Actually, can you just re-test with the current async-buffered.6 branch?
I think the major surgery should wait for 5.9, we can do this a bit
easier without having to touch everything around us.

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

On Mon, Jun 1, 2020 at 4:35 PM Jens Axboe <axboe@kernel.dk> wrote:
>
> On 6/1/20 8:14 AM, Jens Axboe wrote:
> > On 6/1/20 8:13 AM, Sedat Dilek wrote:
> >> On Mon, Jun 1, 2020 at 4:04 PM Jens Axboe <axboe@kernel.dk> wrote:
> >>>
> >>> On 6/1/20 7:35 AM, Sedat Dilek wrote:
> >>>> Hi Jens,
> >>>>
> >>>> with Linux v5.7 final I switched to linux-block.git/for-next and reverted...
> >>>>
> >>>> "block: read-ahead submission should imply no-wait as well"
> >>>>
> >>>> ...and see no boot-slowdowns.
> >>>
> >>> Can you try with these patches applied instead? Or pull my async-readahead
> >>> branch from the same location.
> >>>
> >>
> >> Yes, I can do that.
> >> I pulled from linux-block.git#async-readahead and will report later.
> >>
> >> Any specific testing desired by you?
> >
> > Just do your boot timing test and see if it works, thanks.
>
> Actually, can you just re-test with the current async-buffered.6 branch?
> I think the major surgery should wait for 5.9, we can do this a bit
> easier without having to touch everything around us.
>

With linux-block.git#async-readahead:

  mycompiler -Wp,-MD,kernel/.sys.o.d -nostdinc -isystem
/home/dileks/src/llvm-toolchain/install/lib/clang/10.0.1rc1/include
-I./arch/x86/include -I./arch/x86/include/generated  -I./include
-I./arch/x86/include/uapi -I./arch/x86/include/generated/uapi
-I./include/uapi -I./include/generated/uapi -include
./include/linux/kconfig.h -include ./include/linux/compiler_types.h
-D__KERNEL__ -Qunused-arguments -Wall -Wundef
-Werror=strict-prototypes -Wno-trigraphs -fno-strict-aliasing
-fno-common -fshort-wchar -fno-PIE
-Werror=implicit-function-declaration -Werror=implicit-int
-Wno-format-security -std=gnu89 -no-integrated-as
-Werror=unknown-warning-option -mno-sse -mno-mmx -mno-sse2 -mno-3dnow
-mno-avx -m64 -mno-80387 -mstack-alignment=8 -mtune=generic
-mno-red-zone -mcmodel=kernel -Wno-sign-compare
-fno-asynchronous-unwind-tables -mretpoline-external-thunk
-fno-delete-null-pointer-checks -Wno-address-of-packed-member -O2
-Wframe-larger-than=2048 -fstack-protector-strong
-Wno-format-invalid-specifier -Wno-gnu -mno-global-merge
-Wno-unused-const-variable -g -gz=zlib -pg -mfentry -DCC_USING_FENTRY
-Wdeclaration-after-statement -Wvla -Wno-pointer-sign
-Wno-array-bounds -fno-strict-overflow -fno-merge-all-constants
-fno-stack-check -Werror=date-time -Werror=incompatible-pointer-types
-fmacro-prefix-map=./= -fcf-protection=none -Wno-initializer-overrides
-Wno-format -Wno-sign-compare -Wno-format-zero-length
-Wno-tautological-constant-out-of-range-compare
-DKBUILD_MODFILE='"kernel/sys"' -DKBUILD_BASENAME='"sys"'
-DKBUILD_MODNAME='"sys"' -c -o kernel/sys.o kernel/sys.c
fs/9p/vfs_addr.c:112:4: error: use of undeclared identifier 'filp'
                        filp->private_data);
                        ^
1 error generated.
make[5]: *** [scripts/Makefile.build:267: fs/9p/vfs_addr.o] Error 1
make[4]: *** [scripts/Makefile.build:488: fs/9p] Error 2
make[3]: *** [Makefile:1735: fs] Error 2
make[3]: *** Waiting for unfinished jobs....

I guess block.git#async-buffered.6 needs the same revert of "block:
read-ahead submission should imply no-wait as well".

- Sedat -

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/1/20 8:43 AM, Sedat Dilek wrote:
> On Mon, Jun 1, 2020 at 4:35 PM Jens Axboe <axboe@kernel.dk> wrote:
>>
>> On 6/1/20 8:14 AM, Jens Axboe wrote:
>>> On 6/1/20 8:13 AM, Sedat Dilek wrote:
>>>> On Mon, Jun 1, 2020 at 4:04 PM Jens Axboe <axboe@kernel.dk> wrote:
>>>>>
>>>>> On 6/1/20 7:35 AM, Sedat Dilek wrote:
>>>>>> Hi Jens,
>>>>>>
>>>>>> with Linux v5.7 final I switched to linux-block.git/for-next and reverted...
>>>>>>
>>>>>> "block: read-ahead submission should imply no-wait as well"
>>>>>>
>>>>>> ...and see no boot-slowdowns.
>>>>>
>>>>> Can you try with these patches applied instead? Or pull my async-readahead
>>>>> branch from the same location.
>>>>>
>>>>
>>>> Yes, I can do that.
>>>> I pulled from linux-block.git#async-readahead and will report later.
>>>>
>>>> Any specific testing desired by you?
>>>
>>> Just do your boot timing test and see if it works, thanks.
>>
>> Actually, can you just re-test with the current async-buffered.6 branch?
>> I think the major surgery should wait for 5.9, we can do this a bit
>> easier without having to touch everything around us.
>>
> 
> With linux-block.git#async-readahead:
> 
>   mycompiler -Wp,-MD,kernel/.sys.o.d -nostdinc -isystem
> /home/dileks/src/llvm-toolchain/install/lib/clang/10.0.1rc1/include
> -I./arch/x86/include -I./arch/x86/include/generated  -I./include
> -I./arch/x86/include/uapi -I./arch/x86/include/generated/uapi
> -I./include/uapi -I./include/generated/uapi -include
> ./include/linux/kconfig.h -include ./include/linux/compiler_types.h
> -D__KERNEL__ -Qunused-arguments -Wall -Wundef
> -Werror=strict-prototypes -Wno-trigraphs -fno-strict-aliasing
> -fno-common -fshort-wchar -fno-PIE
> -Werror=implicit-function-declaration -Werror=implicit-int
> -Wno-format-security -std=gnu89 -no-integrated-as
> -Werror=unknown-warning-option -mno-sse -mno-mmx -mno-sse2 -mno-3dnow
> -mno-avx -m64 -mno-80387 -mstack-alignment=8 -mtune=generic
> -mno-red-zone -mcmodel=kernel -Wno-sign-compare
> -fno-asynchronous-unwind-tables -mretpoline-external-thunk
> -fno-delete-null-pointer-checks -Wno-address-of-packed-member -O2
> -Wframe-larger-than=2048 -fstack-protector-strong
> -Wno-format-invalid-specifier -Wno-gnu -mno-global-merge
> -Wno-unused-const-variable -g -gz=zlib -pg -mfentry -DCC_USING_FENTRY
> -Wdeclaration-after-statement -Wvla -Wno-pointer-sign
> -Wno-array-bounds -fno-strict-overflow -fno-merge-all-constants
> -fno-stack-check -Werror=date-time -Werror=incompatible-pointer-types
> -fmacro-prefix-map=./= -fcf-protection=none -Wno-initializer-overrides
> -Wno-format -Wno-sign-compare -Wno-format-zero-length
> -Wno-tautological-constant-out-of-range-compare
> -DKBUILD_MODFILE='"kernel/sys"' -DKBUILD_BASENAME='"sys"'
> -DKBUILD_MODNAME='"sys"' -c -o kernel/sys.o kernel/sys.c
> fs/9p/vfs_addr.c:112:4: error: use of undeclared identifier 'filp'
>                         filp->private_data);
>                         ^
> 1 error generated.
> make[5]: *** [scripts/Makefile.build:267: fs/9p/vfs_addr.o] Error 1
> make[4]: *** [scripts/Makefile.build:488: fs/9p] Error 2
> make[3]: *** [Makefile:1735: fs] Error 2
> make[3]: *** Waiting for unfinished jobs....
> 
> I guess block.git#async-buffered.6 needs the same revert of "block:
> read-ahead submission should imply no-wait as well".

Sorry, forgot to push out the updated version. But as per previous
email, I think that major work should wait. Just try the updated
async-buffered.6 branch instead.

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

On Mon, Jun 1, 2020 at 4:46 PM Jens Axboe <axboe@kernel.dk> wrote:
>
> On 6/1/20 8:43 AM, Sedat Dilek wrote:
> > On Mon, Jun 1, 2020 at 4:35 PM Jens Axboe <axboe@kernel.dk> wrote:
> >>
> >> On 6/1/20 8:14 AM, Jens Axboe wrote:
> >>> On 6/1/20 8:13 AM, Sedat Dilek wrote:
> >>>> On Mon, Jun 1, 2020 at 4:04 PM Jens Axboe <axboe@kernel.dk> wrote:
> >>>>>
> >>>>> On 6/1/20 7:35 AM, Sedat Dilek wrote:
> >>>>>> Hi Jens,
> >>>>>>
> >>>>>> with Linux v5.7 final I switched to linux-block.git/for-next and reverted...
> >>>>>>
> >>>>>> "block: read-ahead submission should imply no-wait as well"
> >>>>>>
> >>>>>> ...and see no boot-slowdowns.
> >>>>>
> >>>>> Can you try with these patches applied instead? Or pull my async-readahead
> >>>>> branch from the same location.
> >>>>>
> >>>>
> >>>> Yes, I can do that.
> >>>> I pulled from linux-block.git#async-readahead and will report later.
> >>>>
> >>>> Any specific testing desired by you?
> >>>
> >>> Just do your boot timing test and see if it works, thanks.
> >>
> >> Actually, can you just re-test with the current async-buffered.6 branch?
> >> I think the major surgery should wait for 5.9, we can do this a bit
> >> easier without having to touch everything around us.
> >>
> >
> > With linux-block.git#async-readahead:
> >
> >   mycompiler -Wp,-MD,kernel/.sys.o.d -nostdinc -isystem
> > /home/dileks/src/llvm-toolchain/install/lib/clang/10.0.1rc1/include
> > -I./arch/x86/include -I./arch/x86/include/generated  -I./include
> > -I./arch/x86/include/uapi -I./arch/x86/include/generated/uapi
> > -I./include/uapi -I./include/generated/uapi -include
> > ./include/linux/kconfig.h -include ./include/linux/compiler_types.h
> > -D__KERNEL__ -Qunused-arguments -Wall -Wundef
> > -Werror=strict-prototypes -Wno-trigraphs -fno-strict-aliasing
> > -fno-common -fshort-wchar -fno-PIE
> > -Werror=implicit-function-declaration -Werror=implicit-int
> > -Wno-format-security -std=gnu89 -no-integrated-as
> > -Werror=unknown-warning-option -mno-sse -mno-mmx -mno-sse2 -mno-3dnow
> > -mno-avx -m64 -mno-80387 -mstack-alignment=8 -mtune=generic
> > -mno-red-zone -mcmodel=kernel -Wno-sign-compare
> > -fno-asynchronous-unwind-tables -mretpoline-external-thunk
> > -fno-delete-null-pointer-checks -Wno-address-of-packed-member -O2
> > -Wframe-larger-than=2048 -fstack-protector-strong
> > -Wno-format-invalid-specifier -Wno-gnu -mno-global-merge
> > -Wno-unused-const-variable -g -gz=zlib -pg -mfentry -DCC_USING_FENTRY
> > -Wdeclaration-after-statement -Wvla -Wno-pointer-sign
> > -Wno-array-bounds -fno-strict-overflow -fno-merge-all-constants
> > -fno-stack-check -Werror=date-time -Werror=incompatible-pointer-types
> > -fmacro-prefix-map=./= -fcf-protection=none -Wno-initializer-overrides
> > -Wno-format -Wno-sign-compare -Wno-format-zero-length
> > -Wno-tautological-constant-out-of-range-compare
> > -DKBUILD_MODFILE='"kernel/sys"' -DKBUILD_BASENAME='"sys"'
> > -DKBUILD_MODNAME='"sys"' -c -o kernel/sys.o kernel/sys.c
> > fs/9p/vfs_addr.c:112:4: error: use of undeclared identifier 'filp'
> >                         filp->private_data);
> >                         ^
> > 1 error generated.
> > make[5]: *** [scripts/Makefile.build:267: fs/9p/vfs_addr.o] Error 1
> > make[4]: *** [scripts/Makefile.build:488: fs/9p] Error 2
> > make[3]: *** [Makefile:1735: fs] Error 2
> > make[3]: *** Waiting for unfinished jobs....
> >
> > I guess block.git#async-buffered.6 needs the same revert of "block:
> > read-ahead submission should imply no-wait as well".
>
> Sorry, forgot to push out the updated version. But as per previous
> email, I think that major work should wait. Just try the updated
> async-buffered.6 branch instead.
>

Hehe, I was looking at git.k.o URLs.

I just pulled from
<https://git.kernel.dk/cgit/linux-block/log/?h=async-buffered.6>.

- Sedat -

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

On Mon, Jun 1, 2020 at 4:46 PM Jens Axboe <axboe@kernel.dk> wrote:
>
> On 6/1/20 8:43 AM, Sedat Dilek wrote:
> > On Mon, Jun 1, 2020 at 4:35 PM Jens Axboe <axboe@kernel.dk> wrote:
> >>
> >> On 6/1/20 8:14 AM, Jens Axboe wrote:
> >>> On 6/1/20 8:13 AM, Sedat Dilek wrote:
> >>>> On Mon, Jun 1, 2020 at 4:04 PM Jens Axboe <axboe@kernel.dk> wrote:
> >>>>>
> >>>>> On 6/1/20 7:35 AM, Sedat Dilek wrote:
> >>>>>> Hi Jens,
> >>>>>>
> >>>>>> with Linux v5.7 final I switched to linux-block.git/for-next and reverted...
> >>>>>>
> >>>>>> "block: read-ahead submission should imply no-wait as well"
> >>>>>>
> >>>>>> ...and see no boot-slowdowns.
> >>>>>
> >>>>> Can you try with these patches applied instead? Or pull my async-readahead
> >>>>> branch from the same location.
> >>>>>
> >>>>
> >>>> Yes, I can do that.
> >>>> I pulled from linux-block.git#async-readahead and will report later.
> >>>>
> >>>> Any specific testing desired by you?
> >>>
> >>> Just do your boot timing test and see if it works, thanks.
> >>
> >> Actually, can you just re-test with the current async-buffered.6 branch?
> >> I think the major surgery should wait for 5.9, we can do this a bit
> >> easier without having to touch everything around us.
> >>
> >
> > With linux-block.git#async-readahead:
> >
> >   mycompiler -Wp,-MD,kernel/.sys.o.d -nostdinc -isystem
> > /home/dileks/src/llvm-toolchain/install/lib/clang/10.0.1rc1/include
> > -I./arch/x86/include -I./arch/x86/include/generated  -I./include
> > -I./arch/x86/include/uapi -I./arch/x86/include/generated/uapi
> > -I./include/uapi -I./include/generated/uapi -include
> > ./include/linux/kconfig.h -include ./include/linux/compiler_types.h
> > -D__KERNEL__ -Qunused-arguments -Wall -Wundef
> > -Werror=strict-prototypes -Wno-trigraphs -fno-strict-aliasing
> > -fno-common -fshort-wchar -fno-PIE
> > -Werror=implicit-function-declaration -Werror=implicit-int
> > -Wno-format-security -std=gnu89 -no-integrated-as
> > -Werror=unknown-warning-option -mno-sse -mno-mmx -mno-sse2 -mno-3dnow
> > -mno-avx -m64 -mno-80387 -mstack-alignment=8 -mtune=generic
> > -mno-red-zone -mcmodel=kernel -Wno-sign-compare
> > -fno-asynchronous-unwind-tables -mretpoline-external-thunk
> > -fno-delete-null-pointer-checks -Wno-address-of-packed-member -O2
> > -Wframe-larger-than=2048 -fstack-protector-strong
> > -Wno-format-invalid-specifier -Wno-gnu -mno-global-merge
> > -Wno-unused-const-variable -g -gz=zlib -pg -mfentry -DCC_USING_FENTRY
> > -Wdeclaration-after-statement -Wvla -Wno-pointer-sign
> > -Wno-array-bounds -fno-strict-overflow -fno-merge-all-constants
> > -fno-stack-check -Werror=date-time -Werror=incompatible-pointer-types
> > -fmacro-prefix-map=./= -fcf-protection=none -Wno-initializer-overrides
> > -Wno-format -Wno-sign-compare -Wno-format-zero-length
> > -Wno-tautological-constant-out-of-range-compare
> > -DKBUILD_MODFILE='"kernel/sys"' -DKBUILD_BASENAME='"sys"'
> > -DKBUILD_MODNAME='"sys"' -c -o kernel/sys.o kernel/sys.c
> > fs/9p/vfs_addr.c:112:4: error: use of undeclared identifier 'filp'
> >                         filp->private_data);
> >                         ^
> > 1 error generated.
> > make[5]: *** [scripts/Makefile.build:267: fs/9p/vfs_addr.o] Error 1
> > make[4]: *** [scripts/Makefile.build:488: fs/9p] Error 2
> > make[3]: *** [Makefile:1735: fs] Error 2
> > make[3]: *** Waiting for unfinished jobs....
> >
> > I guess block.git#async-buffered.6 needs the same revert of "block:
> > read-ahead submission should imply no-wait as well".
>
> Sorry, forgot to push out the updated version. But as per previous
> email, I think that major work should wait. Just try the updated
> async-buffered.6 branch instead.
>
> --
> Jens Axboe
>

Looks good.

$ cat systemd-analyze_time_5.7.0-2-amd64-clang.txt
Startup finished in 6.054s (kernel) + 42.626s (userspace) = 48.680s
graphical.target reached after 42.605s in userspace

$ cat systemd-analyze_blame_5.7.0-2-amd64-clang.txt
25.777s udisks2.service
24.527s accounts-daemon.service
18.228s polkit.service
17.593s avahi-daemon.service
17.554s NetworkManager.service
17.526s rtkit-daemon.service
17.417s dev-sdc2.device
17.200s switcheroo-control.service
17.186s wpa_supplicant.service
17.174s systemd-logind.service
15.714s smartmontools.service
14.255s zramswap.service
13.495s NetworkManager-wait-online.service
11.092s rsyslog.service
 8.900s networking.service
 7.817s gpm.service
 6.571s ModemManager.service
 6.403s packagekit.service
 5.701s e2scrub_reap.service
 3.714s alsa-restore.service
 3.509s systemd-udevd.service
 3.286s apparmor.service
 3.028s atd.service
 2.361s exim4.service
 1.532s systemd-tmpfiles-setup.service
 1.213s systemd-journal-flush.service
 1.163s keyboard-setup.service
 1.011s bluetooth.service
 1.008s binfmt-support.service
  995ms systemd-modules-load.service
  987ms pppd-dns.service
  919ms systemd-journald.service
  875ms upower.service
  856ms systemd-sysusers.service
  750ms systemd-udev-trigger.service
  728ms dev-hugepages.mount
  725ms dev-mqueue.mount
  724ms sys-kernel-debug.mount
  723ms sys-kernel-tracing.mount
  648ms modprobe@drm.service
  640ms ifupdown-wait-online.service
  609ms systemd-timesyncd.service
  579ms systemd-tmpfiles-setup-dev.service
  521ms systemd-backlight@backlight:intel_backlight.service
  477ms systemd-remount-fs.service
  454ms systemd-sysctl.service
  441ms systemd-random-seed.service
  389ms console-setup.service
  306ms systemd-rfkill.service
  266ms kmod-static-nodes.service
  244ms user@1000.service
  150ms proc-sys-fs-binfmt_misc.mount
  131ms systemd-update-utmp.service
   98ms systemd-user-sessions.service
   26ms user-runtime-dir@1000.service
   15ms systemd-update-utmp-runlevel.service
    6ms sys-fs-fuse-connections.mount
    3ms ifupdown-pre.service

$ cat systemd-analyze_critical-chain_5.7.0-2-amd64-clang.txt
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.

graphical.target @42.605s
└─multi-user.target @42.604s
  └─exim4.service @40.240s +2.361s
    └─network-online.target @40.237s
      └─NetworkManager-wait-online.service @26.740s +13.495s
        └─NetworkManager.service @9.177s +17.554s
          └─dbus.service @9.170s
            └─basic.target @9.081s
              └─sockets.target @9.080s
                └─dbus.socket @9.080s
                  └─sysinit.target @8.814s

└─systemd-backlight@backlight:intel_backlight.service @23.030s +521ms
                      └─system-systemd\x2dbacklight.slice @22.251s
                        └─system.slice @2.429s
                          └─-.slice @2.429s

Hope this helps you.

- Sedat -

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/3/20 6:59 PM, Andres Freund wrote:
> Hi,
> 
> I was trying to benchmark the benefits of this for the io_uring using
> postgres I am working on. The initial results where quite promising
> (reducing cpu usage significantly, to lower than non-uring sync io). But
> unfortunately trying another workload triggered both panics and before
> that seemingly returned wrong data.
> 
> I first saw that problem with b360d424ce02, which was
> linux-block/async-buffered.6 at the time. After hitting the issue, I
> updated to the current linux-block/async-buffered.6, but the problem
> persists.
> 
> The workload that triggers the bug within a few seconds is postgres
> doing a parallel sequential scan of a large table (and aggregating the
> data, but that shouldn't matter). In the triggering case that boils down
> to 9 processes sequentially reading a number of 1GB files (we chunk
> tables internally into smaller files). Each process will read a 512kB
> chunk of the file on its own, and then claim the next 512kB from a
> shared memory location. Most of the IO will be READV requests, reading
> 16 * 8kB into postgres' buffer pool (which may or may not be neighboring
> 8kB pages).

I'll try and reproduce this, any chance you have a test case that can
be run so I don't have to write one from scratch? The more detailed
instructions the better.

> The io submissions in this case will all be done to the same io_uring,
> targeting MAP_HUGETLB | MAP_SHARED | MAP_ANONYMOUS memory. The data is
> on xfs.
> 
> The benchmark starts with dropping caches, which helpfully is visible in
> dmesg... That takes a few seconds, so the actual buffered io_uring load
> starts soon after.
> 
> [  125.526092] tee (7775): drop_caches: 3
> [  146.264327] kill_fasync: bad magic number in fasync_struct!
> 
> After this, but before the oops, I see some userspace memory or
> non-io-uring pipe reads being corrupted.
> 
> 
> [  146.264327] kill_fasync: bad magic number in fasync_struct!
> [  146.285175] kill_fasync: bad magic number in fasync_struct!
> [  146.290793] kill_fasync: bad magic number in fasync_struct!
> [  146.285175] kill_fasync: bad magic number in fasync_struct!
> [  146.290793] kill_fasync: bad magic number in fasync_struct!
> [  157.071979] BUG: kernel NULL pointer dereference, address: 0000000000000020
> [  157.078945] #PF: supervisor read access in kernel mode
> [  157.084082] #PF: error_code(0x0000) - not-present page
> [  157.089225] PGD 0 P4D 0
> [  157.091763] Oops: 0000 [#1] SMP NOPTI
> [  157.095429] CPU: 3 PID: 7756 Comm: postgres Not tainted 5.7.0-rc7-andres-00133-gc8707bf69395 #41
> [  157.104208] Hardware name: Supermicro SYS-7049A-T/X11DAi-N, BIOS 3.2 11/13/2019
> [  157.111518] RIP: 0010:xfs_log_commit_cil+0x356/0x7f0
> [  157.116478] Code: 00 00 48 89 7e 08 49 89 76 30 48 89 11 48 89 4a 08 8b 4c 24 2c 48 89 85 c0 00 00 00 48 89 85 c8 00 00 00 49 8b 46 20 45 31 c9 <8b> 70 20 85 f6 0f 84 d9 02 00 00 45 31 c0 85 c9 7e 4f 41 8b 46 2c
> [  157.135226] RSP: 0018:ffffc90021fefb00 EFLAGS: 00010246
> [  157.140452] RAX: 0000000000000000 RBX: ffff8897a9ea80b0 RCX: 0000000000000458
> [  157.147582] RDX: ffff8897a9ea80c0 RSI: ffff88afc39a1200 RDI: ffff88afb1ae2c18
> [  157.154715] RBP: ffff8897a9ea8000 R08: ffffc90021fefb30 R09: 0000000000000000
> [  157.161848] R10: 0000000000000084 R11: 0000000000000012 R12: ffff8897721bd400
> [  157.168983] R13: ffff88afb1ae2c00 R14: ffff88afb19860c0 R15: ffff8897a9ea80a0
> [  157.176115] FS:  00007ffbbe9a1980(0000) GS:ffff8897e0cc0000(0000) knlGS:0000000000000000
> [  157.184199] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [  157.189946] CR2: 0000000000000020 CR3: 000000179c0ac005 CR4: 00000000007606e0
> [  157.197081] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [  157.204212] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [  157.211343] PKRU: 55555554
> [  157.214048] Call Trace:
> [  157.216498]  __xfs_trans_commit+0xa1/0x340
> [  157.220596]  xfs_create+0x4cc/0x5e0
> [  157.224087]  xfs_generic_create+0x241/0x310
> [  157.228274]  path_openat+0xdb7/0x1020
> [  157.231940]  do_filp_open+0x91/0x100
> [  157.235518]  ? __sys_recvfrom+0xe4/0x170
> [  157.239444]  ? _cond_resched+0x19/0x30
> [  157.243196]  do_sys_openat2+0x215/0x2d0
> [  157.247037]  do_sys_open+0x44/0x80
> [  157.250443]  do_syscall_64+0x48/0x130
> [  157.254108]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> [  157.259161] RIP: 0033:0x7ffbc0ded307
> [  157.262738] Code: 25 00 00 41 00 3d 00 00 41 00 74 47 64 8b 04 25 18 00 00 00 85 c0 75 6b 44 89 e2 48 89 ee bf 9c ff ff ff b8 01 01 00 00 0f 05 <48> 3d 00 f0 ff ff 0f 87 95 00 00 00 48 8b 4c 24 28 64 48 33 0c 25
> [  157.281485] RSP: 002b:00007fff520a4940 EFLAGS: 00000246 ORIG_RAX: 0000000000000101
> [  157.289052] RAX: ffffffffffffffda RBX: 00005561dc02eca0 RCX: 00007ffbc0ded307
> [  157.296185] RDX: 0000000000000241 RSI: 00005561dc013470 RDI: 00000000ffffff9c
> [  157.303317] RBP: 00005561dc013470 R08: 0000000000000004 R09: 0000000000000001
> [  157.310447] R10: 00000000000001b6 R11: 0000000000000246 R12: 0000000000000241
> [  157.317575] R13: 00005561dc02eca0 R14: 0000000000000001 R15: 0000000000000000
> [  157.324705] Modules linked in: isst_if_common squashfs nls_iso8859_1 nls_cp437 snd_hda_codec_realtek vfat snd_hda_codec_generic fat iwlmvm ledtrig_audio snd_hda_codec_hdmi x86_pkg_temp_thermal intel_powerclamp snd_hda_intel snd_intel_dspcfg iwlwifi efi_pstore btusb snd_hda_codec btrtl btbcm efivars snd_hwdep btintel snd_hda_core iTCO_wdt iTCO_vendor_support mei_me mei loop coretemp hid_logitech_hidpp hid_logitech_dj hid_lenovo amdgpu gpu_sched i40e ixgbe ast drm_vram_helper drm_ttm_helper ttm mdio xhci_pci xhci_hcd
> [  157.370249] CR2: 0000000000000020
> 
> There's a lot more of these later, some interspersed (possibly related
> to grabbing the output via serial->bmc->network->ipmitool). I've
> attached the whole dmesg and .config.
> 
> What would be helpful for debugging? Should I try the v5 branch instead?
> 
> I'll try setting up a VM + passing through NVME to be able to test this
> without fear... In one instance I did see some minor corruption on
> another device & fs (ext4 on dm-crypt on nvme). It's all backed up,
> but...

I have a known issue with request starvation, wonder if that could be it.
I'm going to rebase the branch on top of the aops->readahead() changes
shortly, and fix that issue. Hopefully that's what's plaguing your run
here, but if not, I'll hunt that one down.

Thanks for testing!

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Andres Freund]

Hi,

On 2020-06-03 19:04:17 -0600, Jens Axboe wrote:
> > The workload that triggers the bug within a few seconds is postgres
> > doing a parallel sequential scan of a large table (and aggregating the
> > data, but that shouldn't matter). In the triggering case that boils down
> > to 9 processes sequentially reading a number of 1GB files (we chunk
> > tables internally into smaller files). Each process will read a 512kB
> > chunk of the file on its own, and then claim the next 512kB from a
> > shared memory location. Most of the IO will be READV requests, reading
> > 16 * 8kB into postgres' buffer pool (which may or may not be neighboring
> > 8kB pages).
> 
> I'll try and reproduce this, any chance you have a test case that can
> be run so I don't have to write one from scratch? The more detailed
> instructions the better.

It shouldn't be too hard to write you a detailed script for reproducing
the issue. But it'd not be an all that minimal reproducer, unless it
also triggers on smaller scale (it's a 130GB database that triggers the
problem reliably, and small tables don't seem to do so reliably).

I'll try to write that up after I set up kvm / repro there.

One thing I forgot in the earlier email: I ran the benchmark using 'perf
stat -a -e ...'. I'm fairly, but not absolutely, certain that it also
triggered without that. I don't think it's related, but I thought I
better mention it.


> I have a known issue with request starvation, wonder if that could be it.
> I'm going to rebase the branch on top of the aops->readahead() changes
> shortly, and fix that issue. Hopefully that's what's plaguing your run
> here, but if not, I'll hunt that one down.

FWIW, I had iostat -xm /dev/nvme1n1 1 running during this. Shortly
before the crash I see:

Device            r/s     rMB/s   rrqm/s  %rrqm r_await rareq-sz     w/s     wMB/s   wrqm/s  %wrqm w_await wareq-sz     d/s     dMB/s   drqm/s  %drqm d_await dareq-sz     f/s f_await  aqu-sz  %util
nvme1n1       6221.00    956.09  3428.00  35.53    0.24   157.38    0.00      0.00     0.00   0.00    0.00     0.00    0.00      0.00     0.00   0.00    0.00     0.00    0.00    0.00    1.48  99.00

Device            r/s     rMB/s   rrqm/s  %rrqm r_await rareq-sz     w/s     wMB/s   wrqm/s  %wrqm w_await wareq-sz     d/s     dMB/s   drqm/s  %drqm d_await dareq-sz     f/s f_await  aqu-sz  %util
nvme1n1       6456.00    978.83  3439.00  34.75    0.21   155.25    0.00      0.00     0.00   0.00    0.00     0.00    0.00      0.00     0.00   0.00    0.00     0.00    0.00    0.00    1.38  98.70

It's maybe also worth noting that in this workload the results are
*worse* than when using 5.7-rc7 io_uring. So perhaps request starvation
isn't the worst guess...

Greetings,

Andres Freund

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Andres Freund]

Hi,

On 2020-06-03 18:30:45 -0700, Andres Freund wrote:
> > I'll try and reproduce this, any chance you have a test case that can
> > be run so I don't have to write one from scratch? The more detailed
> > instructions the better.
>
> It shouldn't be too hard to write you a detailed script for reproducing
> the issue. But it'd not be an all that minimal reproducer, unless it
> also triggers on smaller scale (it's a 130GB database that triggers the
> problem reliably, and small tables don't seem to do so reliably).
>
> I'll try to write that up after I set up kvm / repro there.

Sorry, took a bit longer. Ran into some issues with my aio branch when
running on 5.6. Not yet sure if that's my fault, or io_uring's.


Here's a recipe that repros reliably for me. For me it triggers reliably
at this size (-s 1000 ends up around ~16GB).


# Build aio branch of postgres
sudo mkdir /srv/src
sudo chown andres. /srv/src
cd /srv/src
git clone -b aio https://github.com/anarazel/postgres.git pg
cd pg
./configure --enable-depend --with-liburing --prefix=/home/andres/bin/
make -j48 -s install

# check that target directory is in PATH
type psql
psql is /home/andres/bin/bin/psql

in one shell:
# create database directory, start postgres
echo 1200 | sudo tee /proc/sys/vm/nr_hugepages
initdb /srv/pguringcrash
postgres -D /srv/pguringcrash/ -c bgwriter_delay=10ms -c bgwriter_lru_maxpages=1000 -c wal_level=minimal -c max_wal_senders=0 -c wal_buffers=128MB -c max_wal_size=100GB -c shared_buffers=2GB -c max_parallel_workers_per_gather=8 -c huge_pages=on

in the other shell:
# fill with test data (skipping index creation etc)
pgbench -i -s 1000 postgres -IdtGv -q

# and trigger a parallel sequential scan, using 8 additional workers
echo 3 | sudo tee /proc/sys/vm/drop_caches && psql -c 'SET max_parallel_workers_per_gather=8' -c 'SELECT SUM(abalance) FROM pgbench_accounts ;'


For me that faily reliably triggers the issue within a few seconds.

[  323.312854] BUG: unable to handle page fault for address: 00007f0f1428c000
[  323.319879] #PF: supervisor read access in kernel mode
[  323.321391] #PF: error_code(0x0001) - permissions violation
[  323.323019] PGD 102e760067 P4D 102e760067 PUD 1037878067 PMD 8000000fc7e000e7
[  323.325065] Oops: 0001 [#1] SMP NOPTI
[  323.326248] CPU: 1 PID: 3145 Comm: postgres Not tainted 5.7.0-rc7-andres-00133-gc8707bf69395 #43
[  323.328653] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
[  323.330365] RIP: 0010:__wake_up_common+0x53/0x140
[  323.331276] Code: 41 f6 01 04 0f 85 a3 00 00 00 48 8b 43 08 48 8d 78 e8 48 8d 43 08 48 89 c1 48 89 04 24 48 8d 47 18 48 39 c1 0f 84 b8 00 00 00 <48> 8b 47 18 89 54 24 14 31 ed 4c 8d 60 e8 8b 1f f6 c3 04 75 52 48
[  323.334419] RSP: 0018:ffffc90000dcbdc8 EFLAGS: 00010086
[  323.335409] RAX: 00007f0f1428c000 RBX: ffff889038b189f8 RCX: ffff889038b18a00
[  323.336723] RDX: 0000000000000001 RSI: 0000000000000003 RDI: 00007f0f1428bfe8
[  323.338035] RBP: 0000000000000246 R08: 0000000000000001 R09: ffffc90000dcbe18
[  323.339331] R10: ffff889036a5e750 R11: 0000000000000000 R12: 0000000000000001
[  323.340312] R13: 0000000000000003 R14: 0000000000000000 R15: 0000000000000001
[  323.341204] FS:  00007f0f6f271740(0000) GS:ffff88903f840000(0000) knlGS:0000000000000000
[  323.342217] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  323.342950] CR2: 00007f0f1428c000 CR3: 0000001037ed0005 CR4: 0000000000760ee0
[  323.343836] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  323.344724] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  323.345613] PKRU: 55555554
[  323.346017] Call Trace:
[  323.346435]  __wake_up_common_lock+0x7c/0xc0
[  323.347015]  ep_free+0x30/0xd0
[  323.347468]  ep_eventpoll_release+0x16/0x20
[  323.348040]  __fput+0xda/0x240
[  323.348498]  task_work_run+0x62/0x90
[  323.349015]  exit_to_usermode_loop+0xbd/0xe0
[  323.349573]  do_syscall_64+0xf2/0x130
[  323.349987]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[  323.350523] RIP: 0033:0x7f0f6f5a5f13
[  323.350906] Code: 00 00 f7 d8 64 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 64 8b 04 25 18 00 00 00 85 c0 75 14 b8 03 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 45 c3 0f 1f 40 00 48 83 ec 18 89 7c 24 0c e8
[  323.352501] RSP: 002b:00007ffdf3eab5a8 EFLAGS: 00000246 ORIG_RAX: 0000000000000003
[  323.353277] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00007f0f6f5a5f13
[  323.353947] RDX: 0000000000000001 RSI: 000055ef9f5cb360 RDI: 000000000000002a
[  323.354587] RBP: 000055ef9f5cb2e8 R08: 00000000000280e8 R09: 0000000000000143
[  323.355227] R10: 000000000000000a R11: 0000000000000246 R12: 000055ef9f5cb2e8
[  323.355872] R13: 0000000000000008 R14: 00000000ffffffff R15: 00007f0f6d3407a4
[  323.356510] Modules linked in: 9pnet_virtio isst_if_common iTCO_wdt 9pnet xhci_pci iTCO_vendor_support xhci_hcd
[  323.357477] CR2: 00007f0f1428c000
[  323.357866] ---[ end trace 081898b3c6b5ab1b ]---


I'll try your new branch next.

Greetings,

Andres Freund

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/3/20 7:04 PM, Jens Axboe wrote:
> On 6/3/20 6:59 PM, Andres Freund wrote:
>> Hi,
>>
>> I was trying to benchmark the benefits of this for the io_uring using
>> postgres I am working on. The initial results where quite promising
>> (reducing cpu usage significantly, to lower than non-uring sync io). But
>> unfortunately trying another workload triggered both panics and before
>> that seemingly returned wrong data.
>>
>> I first saw that problem with b360d424ce02, which was
>> linux-block/async-buffered.6 at the time. After hitting the issue, I
>> updated to the current linux-block/async-buffered.6, but the problem
>> persists.
>>
>> The workload that triggers the bug within a few seconds is postgres
>> doing a parallel sequential scan of a large table (and aggregating the
>> data, but that shouldn't matter). In the triggering case that boils down
>> to 9 processes sequentially reading a number of 1GB files (we chunk
>> tables internally into smaller files). Each process will read a 512kB
>> chunk of the file on its own, and then claim the next 512kB from a
>> shared memory location. Most of the IO will be READV requests, reading
>> 16 * 8kB into postgres' buffer pool (which may or may not be neighboring
>> 8kB pages).
> 
> I'll try and reproduce this, any chance you have a test case that can
> be run so I don't have to write one from scratch? The more detailed
> instructions the better.

Can you try with async-buffered.7? I've rebased it on a new mechanism,
and doing something like what you describe above I haven't been able
to trigger anything bad. I'd try your test case specifically, so do let
know if it's something I can run.

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Andres Freund]

Hi,

On 2020-06-05 08:42:28 -0600, Jens Axboe wrote:
> Can you try with async-buffered.7? I've rebased it on a new mechanism,
> and doing something like what you describe above I haven't been able
> to trigger anything bad. I'd try your test case specifically, so do let
> know if it's something I can run.

I tried my test on async-buffered.7?, and I get hangs very quickly after
starting. Unfortunately, I don't seem to get an OOPSs, not sure yet why.

Let me know if my test triggers for you.

I'll go and try to figure out why I don't see an oops...

Regards,

Andres

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/5/20 2:20 PM, Andres Freund wrote:
> Hi,
> 
> On 2020-06-05 08:42:28 -0600, Jens Axboe wrote:
>> Can you try with async-buffered.7? I've rebased it on a new mechanism,
>> and doing something like what you describe above I haven't been able
>> to trigger anything bad. I'd try your test case specifically, so do let
>> know if it's something I can run.
> 
> I tried my test on async-buffered.7?, and I get hangs very quickly after
> starting. Unfortunately, I don't seem to get an OOPSs, not sure yet why.
> 
> Let me know if my test triggers for you.
> 
> I'll go and try to figure out why I don't see an oops...

I'll try the reproducer! Thanks for testing.

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Andres Freund]

Hi,

On 2020-06-05 13:20:28 -0700, Andres Freund wrote:
> I'll go and try to figure out why I don't see an oops...

Err, that probably was a typo on my end in the serial console
config. After fixing that, I did get the below.

If helpful I can try with debugging enabled or such.

[   67.910265] tee (2577): drop_caches: 3
[   68.059674] BUG: unable to handle page fault for address: 00007f0b16a3c03c
[   68.062021] #PF: supervisor read access in kernel mode
[   68.063742] #PF: error_code(0x0000) - not-present page
[   68.065517] PGD 102e044067 P4D 102e044067 PUD 102bf7a067 PMD 0 
[   68.067519] Oops: 0000 [#1] SMP NOPTI
[   68.068800] CPU: 2 PID: 2554 Comm: postgres Not tainted 5.7.0-andres-10123-g87823242260e #44
[   68.071505] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
[   68.074139] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
[   68.075389] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
[   68.079125] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
[   68.080260] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
[   68.084115] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
[   68.085374] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
[   68.086409] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
[   68.087447] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
[   68.088697] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
[   68.089903] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   68.090776] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
[   68.091834] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[   68.092902] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[   68.093967] PKRU: 55555554
[   68.094427] Call Trace:
[   68.094845]  ? __schedule+0x2ac/0x720
[   68.095350]  ? schedule+0x55/0xc0
[   68.095797]  ? ep_read_events_proc+0xd0/0xd0
[   68.096354]  ep_scan_ready_list.constprop.0+0x16c/0x190
[   68.097016]  ep_poll+0x2a3/0x440
[   68.097449]  ? wait_woken+0x70/0x70
[   68.097904]  do_epoll_wait+0xb0/0xd0
[   68.098375]  __x64_sys_epoll_wait+0x1a/0x20
[   68.098913]  do_syscall_64+0x48/0x130
[   68.099393]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[   68.100030] RIP: 0033:0x7f0b97845606
[   68.100498] Code: 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 11 b8 e8 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 5a c3 90 48 83 ec 28 89 54 24 18 48 89 74 24
[   68.102718] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
[   68.103644] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
[   68.104533] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
[   68.105418] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
[   68.106296] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
[   68.107187] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
[   68.108104] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
[   68.109505] CR2: 00007f0b16a3c03c
[   68.109962] ---[ end trace 0ca39a5ed99162ce ]---
[   68.110547] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
[   68.111214] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
[   68.113435] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
[   68.114111] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
[   68.115016] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
[   68.115902] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
[   68.116810] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
[   68.117663] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
[   68.118520] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
[   68.119482] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   68.120181] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
[   68.121043] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[   68.121904] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[   68.122790] PKRU: 55555554
[   68.123401] general protection fault, probably for non-canonical address 0xfeeda989fef06266: 0000 [#2] SMP NOPTI
[   68.125052] CPU: 2 PID: 2554 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
[   68.126260] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
[   68.127274] RIP: 0010:__pv_queued_spin_lock_slowpath+0x1a1/0x2b0
[   68.128032] Code: c4 c1 ea 12 41 be 01 00 00 00 8d 42 ff 41 83 e4 03 4c 8d 6b 14 49 c1 e4 05 48 98 49 81 c4 00 c7 02 00 4c 03 24 c5 e0 e6 fd 82 <49> 89 1c 24 b8 00 80 00 00 eb 15 84 c0 75 0a 41 0f b6 54 24 14 84
[   68.130221] RSP: 0018:ffffc90000befce8 EFLAGS: 00010086
[   68.130867] RAX: 0000000000003ffe RBX: ffff88903f8ac700 RCX: 0000000000000001
[   68.131752] RDX: 0000000000003fff RSI: 0000000000000000 RDI: 0000000000000000
[   68.132637] RBP: ffff889037617924 R08: 0000000000000000 R09: ffffc90000befdf8
[   68.133513] R10: ffff8890334d2bf0 R11: 0000000000000018 R12: feeda989fef06266
[   68.134399] R13: ffff88903f8ac714 R14: 0000000000000001 R15: 00000000000c0000
[   68.135323] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
[   68.136307] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   68.137004] CR2: 00007f0b16a3c03c CR3: 000000000360a005 CR4: 0000000000760ee0
[   68.137866] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[   68.138748] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[   68.139606] PKRU: 55555554
[   68.139973] Call Trace:
[   68.140313]  queued_read_lock_slowpath+0x6c/0x70
[   68.140915]  _raw_read_lock_irqsave+0x26/0x30
[   68.141485]  ep_poll_callback+0x3e/0x2b0
[   68.142007]  ? set_next_entity+0xab/0x1f0
[   68.142541]  __wake_up_common+0x7a/0x140
[   68.143077]  __wake_up_common_lock+0x7c/0xc0
[   68.143651]  pipe_release+0x5b/0xd0
[   68.144150]  __fput+0xda/0x240
[   68.144574]  task_work_run+0x62/0x90
[   68.145046]  do_exit+0x35c/0xa70
[   68.145505]  ? do_epoll_wait+0xb0/0xd0
[   68.146000]  rewind_stack_do_exit+0x17/0x20
[   68.146538] RIP: 0033:0x7f0b97845606
[   68.146988] Code: Bad RIP value.
[   68.147405] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
[   68.148314] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
[   68.149157] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
[   68.150021] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
[   68.150831] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
[   68.151640] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
[   68.152459] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
[   68.153707] ---[ end trace 0ca39a5ed99162cf ]---
[   68.154282] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
[   68.154884] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
[   68.156976] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
[   68.157614] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
[   68.158436] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
[   68.159269] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
[   68.160092] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
[   68.160920] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
[   68.161746] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
[   68.162701] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   68.163386] CR2: 00007f0b978455dc CR3: 000000000360a005 CR4: 0000000000760ee0
[   68.164226] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[   68.165079] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[   68.165931] PKRU: 55555554
[   68.166298] Fixing recursive fault but reboot is needed!
[  128.173729] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
[  128.179267] rcu: 	11-...0: (1 GPs behind) idle=c56/1/0x4000000000000000 softirq=6897/6898 fqs=5976 
[  128.182232] rcu: 	19-...0: (1 GPs behind) idle=492/1/0x4000000000000000 softirq=1023/1023 fqs=5976 
[  128.185217] 	(detected by 10, t=18003 jiffies, g=15789, q=631)
[  128.186863] Sending NMI from CPU 10 to CPUs 11:
[  128.188902] NMI backtrace for cpu 11
[  128.188903] CPU: 11 PID: 2546 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
[  128.188904] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
[  128.188904] RIP: 0010:queued_read_lock_slowpath+0x5b/0x70
[  128.188906] Code: 03 00 02 00 00 8b 03 84 c0 74 08 f3 90 8b 13 84 d2 75 f8 48 89 ef e8 74 e6 ff ff 66 90 5b 5d c3 8b 07 84 c0 74 08 f3 90 8b 03 <84> c0 75 f8 5b 5d c3 89 c6 48 89 ef e8 e4 e8 ff ff 66 90 eb bf 0f
[  128.188907] RSP: 0018:ffffc90000348c08 EFLAGS: 00000086
[  128.188908] RAX: 0000000037617cc0 RBX: ffff889037617920 RCX: 00000000000000c3
[  128.188909] RDX: 0000000000000001 RSI: 0000000000000001 RDI: ffff889037617920
[  128.188910] RBP: ffff889037bb1a80 R08: 00000000000000c3 R09: ffffc90000348cd8
[  128.188910] R10: 0100000000000000 R11: 00000000de2ee17e R12: 0000000000000046
[  128.188911] R13: ffff889037617920 R14: 0000000000000001 R15: 00000000000000c3
[  128.188912] FS:  00007f0b97743740(0000) GS:ffff88903fac0000(0000) knlGS:0000000000000000
[  128.188912] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  128.188913] CR2: 00007f0b16afc00e CR3: 000000102e5c6002 CR4: 0000000000760ee0
[  128.188914] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  128.188914] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  128.188915] PKRU: 55555554
[  128.188915] Call Trace:
[  128.188916]  <IRQ>
[  128.188916]  _raw_read_lock_irqsave+0x26/0x30
[  128.188917]  ep_poll_callback+0x3e/0x2b0
[  128.188917]  ? cpumask_next_and+0x19/0x20
[  128.188918]  ? update_sd_lb_stats.constprop.0+0xfe/0x810
[  128.188918]  __wake_up_common+0x7a/0x140
[  128.188919]  __wake_up_common_lock+0x7c/0xc0
[  128.188919]  sock_def_readable+0x37/0x60
[  128.188920]  __udp_enqueue_schedule_skb+0x168/0x260
[  128.188920]  udpv6_queue_rcv_one_skb+0x284/0x3c0
[  128.188921]  udp6_unicast_rcv_skb.isra.0+0x44/0xa0
[  128.188921]  ip6_protocol_deliver_rcu+0x235/0x4b0
[  128.188922]  ip6_input_finish+0x11/0x20
[  128.188922]  ip6_input+0xa2/0xb0
[  128.188923]  ? ip6_protocol_deliver_rcu+0x4b0/0x4b0
[  128.188923]  ipv6_rcv+0xc0/0xd0
[  128.188924]  ? ip6_rcv_finish_core.isra.0+0xd0/0xd0
[  128.188924]  __netif_receive_skb_one_core+0x63/0xa0
[  128.188925]  process_backlog+0x98/0x140
[  128.188925]  net_rx_action+0x13a/0x370
[  128.188926]  __do_softirq+0xe0/0x2ca
[  128.188926]  do_softirq_own_stack+0x2a/0x40
[  128.188926]  </IRQ>
[  128.188927]  do_softirq.part.0+0x2b/0x30
[  128.188927]  __local_bh_enable_ip+0x4b/0x50
[  128.188928]  ip6_finish_output2+0x264/0x5b0
[  128.188928]  ip6_output+0x73/0x120
[  128.188929]  ? __ip6_finish_output+0x110/0x110
[  128.188929]  ip6_send_skb+0x1e/0x60
[  128.188930]  udp_v6_send_skb.isra.0+0x197/0x460
[  128.188930]  udpv6_sendmsg+0xb4f/0xdb0
[  128.188931]  ? ip_reply_glue_bits+0x40/0x40
[  128.188931]  ? update_load_avg+0x78/0x630
[  128.188932]  ? update_curr+0x73/0x1d0
[  128.188932]  ? __sys_sendto+0x108/0x190
[  128.188933]  __sys_sendto+0x108/0x190
[  128.188933]  ? __fput+0x1a5/0x240
[  128.188934]  ? _cond_resched+0x19/0x30
[  128.188934]  ? task_work_run+0x67/0x90
[  128.188935]  __x64_sys_sendto+0x25/0x30
[  128.188935]  do_syscall_64+0x48/0x130
[  128.188936]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[  128.188936] RIP: 0033:0x7f0b97a7826c
[  128.188945] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
[  128.188946] RSP: 002b:00007ffe80ffcea8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
[  128.188947] RAX: ffffffffffffffda RBX: 0000000000000058 RCX: 00007f0b97a7826c
[  128.188948] RDX: 0000000000000058 RSI: 000055fb765264c0 RDI: 0000000000000009
[  128.188949] RBP: 000055fb765264c0 R08: 0000000000000000 R09: 0000000000000000
[  128.188949] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffcfc0
[  128.188950] R13: 000000000002c600 R14: 0000000000000000 R15: 000055fb7629b751
[  128.188957] Sending NMI from CPU 10 to CPUs 19:
[  128.239939] NMI backtrace for cpu 19
[  128.239940] CPU: 19 PID: 2587 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
[  128.239940] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
[  128.239940] RIP: 0010:kvm_wait+0x30/0x50
[  128.239941] Code: 8b 05 ac cb f4 7e a9 00 00 f0 00 75 1f 9c 58 fa 0f b6 17 40 38 d6 75 12 f6 c4 02 75 10 e9 07 00 00 00 0f 00 2d 6f ce 54 01 f4 <50> 9d c3 e9 07 00 00 00 0f 00 2d 5f ce 54 01 fb f4 eb ed 66 66 2e
[  128.239942] RSP: 0018:ffffc90000cc78e8 EFLAGS: 00000046
[  128.239942] RAX: 0000000000000046 RBX: ffff88903fcec700 RCX: 0000000000000008
[  128.239943] RDX: 0000000000000003 RSI: 0000000000000003 RDI: ffff889039049d80
[  128.239943] RBP: ffff889039049d80 R08: ffff88907ffe9f80 R09: 00000000000000f8
[  128.239944] R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000000
[  128.239944] R13: 0000000000000001 R14: 0000000000000100 R15: 0000000000500000
[  128.239944] FS:  00007f0b97743740(0000) GS:ffff88903fcc0000(0000) knlGS:0000000000000000
[  128.239945] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  128.239945] CR2: 000055fb76a96030 CR3: 0000001035882004 CR4: 0000000000760ee0
[  128.239945] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  128.239946] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  128.239946] PKRU: 55555554
[  128.239946] Call Trace:
[  128.239946]  __pv_queued_spin_lock_slowpath+0x26e/0x2b0
[  128.239947]  _raw_spin_lock_irqsave+0x25/0x30
[  128.239947]  __wake_up_common_lock+0x62/0xc0
[  128.239947]  sock_def_write_space+0x49/0x90
[  128.239948]  sock_wfree+0x68/0xb0
[  128.239948]  loopback_xmit+0x49/0xe0
[  128.239948]  dev_hard_start_xmit+0x8d/0x1e0
[  128.239948]  __dev_queue_xmit+0x721/0x8e0
[  128.239949]  ip6_finish_output2+0x250/0x5b0
[  128.239949]  ip6_output+0x73/0x120
[  128.239949]  ? __ip6_finish_output+0x110/0x110
[  128.239950]  ip6_send_skb+0x1e/0x60
[  128.239950]  udp_v6_send_skb.isra.0+0x197/0x460
[  128.239950]  udpv6_sendmsg+0xb4f/0xdb0
[  128.239950]  ? release_pages+0x28f/0x2f0
[  128.239950]  ? ip_reply_glue_bits+0x40/0x40
[  128.239951]  ? _cond_resched+0x19/0x30
[  128.239951]  ? unmap_page_range+0x678/0xa60
[  128.239951]  ? __sys_sendto+0x108/0x190
[  128.239951]  __sys_sendto+0x108/0x190
[  128.239952]  ? __fput+0x1a5/0x240
[  128.239952]  ? _cond_resched+0x19/0x30
[  128.239952]  ? task_work_run+0x67/0x90
[  128.239952]  __x64_sys_sendto+0x25/0x30
[  128.239953]  do_syscall_64+0x48/0x130
[  128.239953]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[  128.239953] RIP: 0033:0x7f0b97a7826c
[  128.239954] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
[  128.239954] RSP: 002b:00007ffe80ffd7b8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
[  128.239955] RAX: ffffffffffffffda RBX: 00000000000003a8 RCX: 00007f0b97a7826c
[  128.239955] RDX: 00000000000003a8 RSI: 00007ffe80ffd800 RDI: 0000000000000009
[  128.239955] RBP: 00007ffe80ffd800 R08: 0000000000000000 R09: 0000000000000000
[  128.239956] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffd800
[  128.239956] R13: 00007ffe80ffd800 R14: 000000000000000e R15: 000055fb76b37e58

Greetings,

Andres Freund

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/5/20 2:36 PM, Andres Freund wrote:
> Hi,
> 
> On 2020-06-05 13:20:28 -0700, Andres Freund wrote:
>> I'll go and try to figure out why I don't see an oops...
> 
> Err, that probably was a typo on my end in the serial console
> config. After fixing that, I did get the below.
> 
> If helpful I can try with debugging enabled or such.
> 
> [   67.910265] tee (2577): drop_caches: 3
> [   68.059674] BUG: unable to handle page fault for address: 00007f0b16a3c03c
> [   68.062021] #PF: supervisor read access in kernel mode
> [   68.063742] #PF: error_code(0x0000) - not-present page
> [   68.065517] PGD 102e044067 P4D 102e044067 PUD 102bf7a067 PMD 0 
> [   68.067519] Oops: 0000 [#1] SMP NOPTI
> [   68.068800] CPU: 2 PID: 2554 Comm: postgres Not tainted 5.7.0-andres-10123-g87823242260e #44
> [   68.071505] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
> [   68.074139] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
> [   68.075389] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
> [   68.079125] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
> [   68.080260] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
> [   68.084115] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
> [   68.085374] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
> [   68.086409] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
> [   68.087447] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
> [   68.088697] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
> [   68.089903] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [   68.090776] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
> [   68.091834] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [   68.092902] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [   68.093967] PKRU: 55555554
> [   68.094427] Call Trace:
> [   68.094845]  ? __schedule+0x2ac/0x720
> [   68.095350]  ? schedule+0x55/0xc0
> [   68.095797]  ? ep_read_events_proc+0xd0/0xd0
> [   68.096354]  ep_scan_ready_list.constprop.0+0x16c/0x190
> [   68.097016]  ep_poll+0x2a3/0x440
> [   68.097449]  ? wait_woken+0x70/0x70
> [   68.097904]  do_epoll_wait+0xb0/0xd0
> [   68.098375]  __x64_sys_epoll_wait+0x1a/0x20
> [   68.098913]  do_syscall_64+0x48/0x130
> [   68.099393]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> [   68.100030] RIP: 0033:0x7f0b97845606
> [   68.100498] Code: 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 11 b8 e8 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 5a c3 90 48 83 ec 28 89 54 24 18 48 89 74 24
> [   68.102718] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
> [   68.103644] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
> [   68.104533] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
> [   68.105418] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
> [   68.106296] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
> [   68.107187] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
> [   68.108104] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
> [   68.109505] CR2: 00007f0b16a3c03c
> [   68.109962] ---[ end trace 0ca39a5ed99162ce ]---
> [   68.110547] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
> [   68.111214] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
> [   68.113435] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
> [   68.114111] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
> [   68.115016] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
> [   68.115902] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
> [   68.116810] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
> [   68.117663] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
> [   68.118520] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
> [   68.119482] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [   68.120181] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
> [   68.121043] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [   68.121904] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [   68.122790] PKRU: 55555554
> [   68.123401] general protection fault, probably for non-canonical address 0xfeeda989fef06266: 0000 [#2] SMP NOPTI
> [   68.125052] CPU: 2 PID: 2554 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
> [   68.126260] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
> [   68.127274] RIP: 0010:__pv_queued_spin_lock_slowpath+0x1a1/0x2b0
> [   68.128032] Code: c4 c1 ea 12 41 be 01 00 00 00 8d 42 ff 41 83 e4 03 4c 8d 6b 14 49 c1 e4 05 48 98 49 81 c4 00 c7 02 00 4c 03 24 c5 e0 e6 fd 82 <49> 89 1c 24 b8 00 80 00 00 eb 15 84 c0 75 0a 41 0f b6 54 24 14 84
> [   68.130221] RSP: 0018:ffffc90000befce8 EFLAGS: 00010086
> [   68.130867] RAX: 0000000000003ffe RBX: ffff88903f8ac700 RCX: 0000000000000001
> [   68.131752] RDX: 0000000000003fff RSI: 0000000000000000 RDI: 0000000000000000
> [   68.132637] RBP: ffff889037617924 R08: 0000000000000000 R09: ffffc90000befdf8
> [   68.133513] R10: ffff8890334d2bf0 R11: 0000000000000018 R12: feeda989fef06266
> [   68.134399] R13: ffff88903f8ac714 R14: 0000000000000001 R15: 00000000000c0000
> [   68.135323] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
> [   68.136307] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [   68.137004] CR2: 00007f0b16a3c03c CR3: 000000000360a005 CR4: 0000000000760ee0
> [   68.137866] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [   68.138748] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [   68.139606] PKRU: 55555554
> [   68.139973] Call Trace:
> [   68.140313]  queued_read_lock_slowpath+0x6c/0x70
> [   68.140915]  _raw_read_lock_irqsave+0x26/0x30
> [   68.141485]  ep_poll_callback+0x3e/0x2b0
> [   68.142007]  ? set_next_entity+0xab/0x1f0
> [   68.142541]  __wake_up_common+0x7a/0x140
> [   68.143077]  __wake_up_common_lock+0x7c/0xc0
> [   68.143651]  pipe_release+0x5b/0xd0
> [   68.144150]  __fput+0xda/0x240
> [   68.144574]  task_work_run+0x62/0x90
> [   68.145046]  do_exit+0x35c/0xa70
> [   68.145505]  ? do_epoll_wait+0xb0/0xd0
> [   68.146000]  rewind_stack_do_exit+0x17/0x20
> [   68.146538] RIP: 0033:0x7f0b97845606
> [   68.146988] Code: Bad RIP value.
> [   68.147405] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
> [   68.148314] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
> [   68.149157] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
> [   68.150021] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
> [   68.150831] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
> [   68.151640] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
> [   68.152459] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
> [   68.153707] ---[ end trace 0ca39a5ed99162cf ]---
> [   68.154282] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
> [   68.154884] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
> [   68.156976] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
> [   68.157614] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
> [   68.158436] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
> [   68.159269] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
> [   68.160092] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
> [   68.160920] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
> [   68.161746] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
> [   68.162701] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [   68.163386] CR2: 00007f0b978455dc CR3: 000000000360a005 CR4: 0000000000760ee0
> [   68.164226] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [   68.165079] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [   68.165931] PKRU: 55555554
> [   68.166298] Fixing recursive fault but reboot is needed!
> [  128.173729] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
> [  128.179267] rcu: 	11-...0: (1 GPs behind) idle=c56/1/0x4000000000000000 softirq=6897/6898 fqs=5976 
> [  128.182232] rcu: 	19-...0: (1 GPs behind) idle=492/1/0x4000000000000000 softirq=1023/1023 fqs=5976 
> [  128.185217] 	(detected by 10, t=18003 jiffies, g=15789, q=631)
> [  128.186863] Sending NMI from CPU 10 to CPUs 11:
> [  128.188902] NMI backtrace for cpu 11
> [  128.188903] CPU: 11 PID: 2546 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
> [  128.188904] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
> [  128.188904] RIP: 0010:queued_read_lock_slowpath+0x5b/0x70
> [  128.188906] Code: 03 00 02 00 00 8b 03 84 c0 74 08 f3 90 8b 13 84 d2 75 f8 48 89 ef e8 74 e6 ff ff 66 90 5b 5d c3 8b 07 84 c0 74 08 f3 90 8b 03 <84> c0 75 f8 5b 5d c3 89 c6 48 89 ef e8 e4 e8 ff ff 66 90 eb bf 0f
> [  128.188907] RSP: 0018:ffffc90000348c08 EFLAGS: 00000086
> [  128.188908] RAX: 0000000037617cc0 RBX: ffff889037617920 RCX: 00000000000000c3
> [  128.188909] RDX: 0000000000000001 RSI: 0000000000000001 RDI: ffff889037617920
> [  128.188910] RBP: ffff889037bb1a80 R08: 00000000000000c3 R09: ffffc90000348cd8
> [  128.188910] R10: 0100000000000000 R11: 00000000de2ee17e R12: 0000000000000046
> [  128.188911] R13: ffff889037617920 R14: 0000000000000001 R15: 00000000000000c3
> [  128.188912] FS:  00007f0b97743740(0000) GS:ffff88903fac0000(0000) knlGS:0000000000000000
> [  128.188912] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [  128.188913] CR2: 00007f0b16afc00e CR3: 000000102e5c6002 CR4: 0000000000760ee0
> [  128.188914] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [  128.188914] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [  128.188915] PKRU: 55555554
> [  128.188915] Call Trace:
> [  128.188916]  <IRQ>
> [  128.188916]  _raw_read_lock_irqsave+0x26/0x30
> [  128.188917]  ep_poll_callback+0x3e/0x2b0
> [  128.188917]  ? cpumask_next_and+0x19/0x20
> [  128.188918]  ? update_sd_lb_stats.constprop.0+0xfe/0x810
> [  128.188918]  __wake_up_common+0x7a/0x140
> [  128.188919]  __wake_up_common_lock+0x7c/0xc0
> [  128.188919]  sock_def_readable+0x37/0x60
> [  128.188920]  __udp_enqueue_schedule_skb+0x168/0x260
> [  128.188920]  udpv6_queue_rcv_one_skb+0x284/0x3c0
> [  128.188921]  udp6_unicast_rcv_skb.isra.0+0x44/0xa0
> [  128.188921]  ip6_protocol_deliver_rcu+0x235/0x4b0
> [  128.188922]  ip6_input_finish+0x11/0x20
> [  128.188922]  ip6_input+0xa2/0xb0
> [  128.188923]  ? ip6_protocol_deliver_rcu+0x4b0/0x4b0
> [  128.188923]  ipv6_rcv+0xc0/0xd0
> [  128.188924]  ? ip6_rcv_finish_core.isra.0+0xd0/0xd0
> [  128.188924]  __netif_receive_skb_one_core+0x63/0xa0
> [  128.188925]  process_backlog+0x98/0x140
> [  128.188925]  net_rx_action+0x13a/0x370
> [  128.188926]  __do_softirq+0xe0/0x2ca
> [  128.188926]  do_softirq_own_stack+0x2a/0x40
> [  128.188926]  </IRQ>
> [  128.188927]  do_softirq.part.0+0x2b/0x30
> [  128.188927]  __local_bh_enable_ip+0x4b/0x50
> [  128.188928]  ip6_finish_output2+0x264/0x5b0
> [  128.188928]  ip6_output+0x73/0x120
> [  128.188929]  ? __ip6_finish_output+0x110/0x110
> [  128.188929]  ip6_send_skb+0x1e/0x60
> [  128.188930]  udp_v6_send_skb.isra.0+0x197/0x460
> [  128.188930]  udpv6_sendmsg+0xb4f/0xdb0
> [  128.188931]  ? ip_reply_glue_bits+0x40/0x40
> [  128.188931]  ? update_load_avg+0x78/0x630
> [  128.188932]  ? update_curr+0x73/0x1d0
> [  128.188932]  ? __sys_sendto+0x108/0x190
> [  128.188933]  __sys_sendto+0x108/0x190
> [  128.188933]  ? __fput+0x1a5/0x240
> [  128.188934]  ? _cond_resched+0x19/0x30
> [  128.188934]  ? task_work_run+0x67/0x90
> [  128.188935]  __x64_sys_sendto+0x25/0x30
> [  128.188935]  do_syscall_64+0x48/0x130
> [  128.188936]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> [  128.188936] RIP: 0033:0x7f0b97a7826c
> [  128.188945] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
> [  128.188946] RSP: 002b:00007ffe80ffcea8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
> [  128.188947] RAX: ffffffffffffffda RBX: 0000000000000058 RCX: 00007f0b97a7826c
> [  128.188948] RDX: 0000000000000058 RSI: 000055fb765264c0 RDI: 0000000000000009
> [  128.188949] RBP: 000055fb765264c0 R08: 0000000000000000 R09: 0000000000000000
> [  128.188949] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffcfc0
> [  128.188950] R13: 000000000002c600 R14: 0000000000000000 R15: 000055fb7629b751
> [  128.188957] Sending NMI from CPU 10 to CPUs 19:
> [  128.239939] NMI backtrace for cpu 19
> [  128.239940] CPU: 19 PID: 2587 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
> [  128.239940] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
> [  128.239940] RIP: 0010:kvm_wait+0x30/0x50
> [  128.239941] Code: 8b 05 ac cb f4 7e a9 00 00 f0 00 75 1f 9c 58 fa 0f b6 17 40 38 d6 75 12 f6 c4 02 75 10 e9 07 00 00 00 0f 00 2d 6f ce 54 01 f4 <50> 9d c3 e9 07 00 00 00 0f 00 2d 5f ce 54 01 fb f4 eb ed 66 66 2e
> [  128.239942] RSP: 0018:ffffc90000cc78e8 EFLAGS: 00000046
> [  128.239942] RAX: 0000000000000046 RBX: ffff88903fcec700 RCX: 0000000000000008
> [  128.239943] RDX: 0000000000000003 RSI: 0000000000000003 RDI: ffff889039049d80
> [  128.239943] RBP: ffff889039049d80 R08: ffff88907ffe9f80 R09: 00000000000000f8
> [  128.239944] R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000000
> [  128.239944] R13: 0000000000000001 R14: 0000000000000100 R15: 0000000000500000
> [  128.239944] FS:  00007f0b97743740(0000) GS:ffff88903fcc0000(0000) knlGS:0000000000000000
> [  128.239945] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [  128.239945] CR2: 000055fb76a96030 CR3: 0000001035882004 CR4: 0000000000760ee0
> [  128.239945] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [  128.239946] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [  128.239946] PKRU: 55555554
> [  128.239946] Call Trace:
> [  128.239946]  __pv_queued_spin_lock_slowpath+0x26e/0x2b0
> [  128.239947]  _raw_spin_lock_irqsave+0x25/0x30
> [  128.239947]  __wake_up_common_lock+0x62/0xc0
> [  128.239947]  sock_def_write_space+0x49/0x90
> [  128.239948]  sock_wfree+0x68/0xb0
> [  128.239948]  loopback_xmit+0x49/0xe0
> [  128.239948]  dev_hard_start_xmit+0x8d/0x1e0
> [  128.239948]  __dev_queue_xmit+0x721/0x8e0
> [  128.239949]  ip6_finish_output2+0x250/0x5b0
> [  128.239949]  ip6_output+0x73/0x120
> [  128.239949]  ? __ip6_finish_output+0x110/0x110
> [  128.239950]  ip6_send_skb+0x1e/0x60
> [  128.239950]  udp_v6_send_skb.isra.0+0x197/0x460
> [  128.239950]  udpv6_sendmsg+0xb4f/0xdb0
> [  128.239950]  ? release_pages+0x28f/0x2f0
> [  128.239950]  ? ip_reply_glue_bits+0x40/0x40
> [  128.239951]  ? _cond_resched+0x19/0x30
> [  128.239951]  ? unmap_page_range+0x678/0xa60
> [  128.239951]  ? __sys_sendto+0x108/0x190
> [  128.239951]  __sys_sendto+0x108/0x190
> [  128.239952]  ? __fput+0x1a5/0x240
> [  128.239952]  ? _cond_resched+0x19/0x30
> [  128.239952]  ? task_work_run+0x67/0x90
> [  128.239952]  __x64_sys_sendto+0x25/0x30
> [  128.239953]  do_syscall_64+0x48/0x130
> [  128.239953]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> [  128.239953] RIP: 0033:0x7f0b97a7826c
> [  128.239954] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
> [  128.239954] RSP: 002b:00007ffe80ffd7b8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
> [  128.239955] RAX: ffffffffffffffda RBX: 00000000000003a8 RCX: 00007f0b97a7826c
> [  128.239955] RDX: 00000000000003a8 RSI: 00007ffe80ffd800 RDI: 0000000000000009
> [  128.239955] RBP: 00007ffe80ffd800 R08: 0000000000000000 R09: 0000000000000000
> [  128.239956] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffd800
> [  128.239956] R13: 00007ffe80ffd800 R14: 000000000000000e R15: 000055fb76b37e58

I can reproduce this, and I see what it is. I'll send out a patch soonish.

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/5/20 2:53 PM, Jens Axboe wrote:
> On 6/5/20 2:36 PM, Andres Freund wrote:
>> Hi,
>>
>> On 2020-06-05 13:20:28 -0700, Andres Freund wrote:
>>> I'll go and try to figure out why I don't see an oops...
>>
>> Err, that probably was a typo on my end in the serial console
>> config. After fixing that, I did get the below.
>>
>> If helpful I can try with debugging enabled or such.
>>
>> [   67.910265] tee (2577): drop_caches: 3
>> [   68.059674] BUG: unable to handle page fault for address: 00007f0b16a3c03c
>> [   68.062021] #PF: supervisor read access in kernel mode
>> [   68.063742] #PF: error_code(0x0000) - not-present page
>> [   68.065517] PGD 102e044067 P4D 102e044067 PUD 102bf7a067 PMD 0 
>> [   68.067519] Oops: 0000 [#1] SMP NOPTI
>> [   68.068800] CPU: 2 PID: 2554 Comm: postgres Not tainted 5.7.0-andres-10123-g87823242260e #44
>> [   68.071505] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>> [   68.074139] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
>> [   68.075389] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
>> [   68.079125] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
>> [   68.080260] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
>> [   68.084115] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
>> [   68.085374] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
>> [   68.086409] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
>> [   68.087447] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
>> [   68.088697] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>> [   68.089903] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [   68.090776] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
>> [   68.091834] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>> [   68.092902] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>> [   68.093967] PKRU: 55555554
>> [   68.094427] Call Trace:
>> [   68.094845]  ? __schedule+0x2ac/0x720
>> [   68.095350]  ? schedule+0x55/0xc0
>> [   68.095797]  ? ep_read_events_proc+0xd0/0xd0
>> [   68.096354]  ep_scan_ready_list.constprop.0+0x16c/0x190
>> [   68.097016]  ep_poll+0x2a3/0x440
>> [   68.097449]  ? wait_woken+0x70/0x70
>> [   68.097904]  do_epoll_wait+0xb0/0xd0
>> [   68.098375]  __x64_sys_epoll_wait+0x1a/0x20
>> [   68.098913]  do_syscall_64+0x48/0x130
>> [   68.099393]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>> [   68.100030] RIP: 0033:0x7f0b97845606
>> [   68.100498] Code: 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 11 b8 e8 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 5a c3 90 48 83 ec 28 89 54 24 18 48 89 74 24
>> [   68.102718] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
>> [   68.103644] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
>> [   68.104533] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
>> [   68.105418] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
>> [   68.106296] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
>> [   68.107187] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
>> [   68.108104] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
>> [   68.109505] CR2: 00007f0b16a3c03c
>> [   68.109962] ---[ end trace 0ca39a5ed99162ce ]---
>> [   68.110547] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
>> [   68.111214] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
>> [   68.113435] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
>> [   68.114111] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
>> [   68.115016] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
>> [   68.115902] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
>> [   68.116810] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
>> [   68.117663] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
>> [   68.118520] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>> [   68.119482] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [   68.120181] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
>> [   68.121043] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>> [   68.121904] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>> [   68.122790] PKRU: 55555554
>> [   68.123401] general protection fault, probably for non-canonical address 0xfeeda989fef06266: 0000 [#2] SMP NOPTI
>> [   68.125052] CPU: 2 PID: 2554 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
>> [   68.126260] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>> [   68.127274] RIP: 0010:__pv_queued_spin_lock_slowpath+0x1a1/0x2b0
>> [   68.128032] Code: c4 c1 ea 12 41 be 01 00 00 00 8d 42 ff 41 83 e4 03 4c 8d 6b 14 49 c1 e4 05 48 98 49 81 c4 00 c7 02 00 4c 03 24 c5 e0 e6 fd 82 <49> 89 1c 24 b8 00 80 00 00 eb 15 84 c0 75 0a 41 0f b6 54 24 14 84
>> [   68.130221] RSP: 0018:ffffc90000befce8 EFLAGS: 00010086
>> [   68.130867] RAX: 0000000000003ffe RBX: ffff88903f8ac700 RCX: 0000000000000001
>> [   68.131752] RDX: 0000000000003fff RSI: 0000000000000000 RDI: 0000000000000000
>> [   68.132637] RBP: ffff889037617924 R08: 0000000000000000 R09: ffffc90000befdf8
>> [   68.133513] R10: ffff8890334d2bf0 R11: 0000000000000018 R12: feeda989fef06266
>> [   68.134399] R13: ffff88903f8ac714 R14: 0000000000000001 R15: 00000000000c0000
>> [   68.135323] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>> [   68.136307] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [   68.137004] CR2: 00007f0b16a3c03c CR3: 000000000360a005 CR4: 0000000000760ee0
>> [   68.137866] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>> [   68.138748] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>> [   68.139606] PKRU: 55555554
>> [   68.139973] Call Trace:
>> [   68.140313]  queued_read_lock_slowpath+0x6c/0x70
>> [   68.140915]  _raw_read_lock_irqsave+0x26/0x30
>> [   68.141485]  ep_poll_callback+0x3e/0x2b0
>> [   68.142007]  ? set_next_entity+0xab/0x1f0
>> [   68.142541]  __wake_up_common+0x7a/0x140
>> [   68.143077]  __wake_up_common_lock+0x7c/0xc0
>> [   68.143651]  pipe_release+0x5b/0xd0
>> [   68.144150]  __fput+0xda/0x240
>> [   68.144574]  task_work_run+0x62/0x90
>> [   68.145046]  do_exit+0x35c/0xa70
>> [   68.145505]  ? do_epoll_wait+0xb0/0xd0
>> [   68.146000]  rewind_stack_do_exit+0x17/0x20
>> [   68.146538] RIP: 0033:0x7f0b97845606
>> [   68.146988] Code: Bad RIP value.
>> [   68.147405] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
>> [   68.148314] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
>> [   68.149157] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
>> [   68.150021] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
>> [   68.150831] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
>> [   68.151640] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
>> [   68.152459] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
>> [   68.153707] ---[ end trace 0ca39a5ed99162cf ]---
>> [   68.154282] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
>> [   68.154884] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
>> [   68.156976] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
>> [   68.157614] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
>> [   68.158436] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
>> [   68.159269] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
>> [   68.160092] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
>> [   68.160920] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
>> [   68.161746] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>> [   68.162701] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [   68.163386] CR2: 00007f0b978455dc CR3: 000000000360a005 CR4: 0000000000760ee0
>> [   68.164226] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>> [   68.165079] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>> [   68.165931] PKRU: 55555554
>> [   68.166298] Fixing recursive fault but reboot is needed!
>> [  128.173729] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
>> [  128.179267] rcu: 	11-...0: (1 GPs behind) idle=c56/1/0x4000000000000000 softirq=6897/6898 fqs=5976 
>> [  128.182232] rcu: 	19-...0: (1 GPs behind) idle=492/1/0x4000000000000000 softirq=1023/1023 fqs=5976 
>> [  128.185217] 	(detected by 10, t=18003 jiffies, g=15789, q=631)
>> [  128.186863] Sending NMI from CPU 10 to CPUs 11:
>> [  128.188902] NMI backtrace for cpu 11
>> [  128.188903] CPU: 11 PID: 2546 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
>> [  128.188904] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>> [  128.188904] RIP: 0010:queued_read_lock_slowpath+0x5b/0x70
>> [  128.188906] Code: 03 00 02 00 00 8b 03 84 c0 74 08 f3 90 8b 13 84 d2 75 f8 48 89 ef e8 74 e6 ff ff 66 90 5b 5d c3 8b 07 84 c0 74 08 f3 90 8b 03 <84> c0 75 f8 5b 5d c3 89 c6 48 89 ef e8 e4 e8 ff ff 66 90 eb bf 0f
>> [  128.188907] RSP: 0018:ffffc90000348c08 EFLAGS: 00000086
>> [  128.188908] RAX: 0000000037617cc0 RBX: ffff889037617920 RCX: 00000000000000c3
>> [  128.188909] RDX: 0000000000000001 RSI: 0000000000000001 RDI: ffff889037617920
>> [  128.188910] RBP: ffff889037bb1a80 R08: 00000000000000c3 R09: ffffc90000348cd8
>> [  128.188910] R10: 0100000000000000 R11: 00000000de2ee17e R12: 0000000000000046
>> [  128.188911] R13: ffff889037617920 R14: 0000000000000001 R15: 00000000000000c3
>> [  128.188912] FS:  00007f0b97743740(0000) GS:ffff88903fac0000(0000) knlGS:0000000000000000
>> [  128.188912] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [  128.188913] CR2: 00007f0b16afc00e CR3: 000000102e5c6002 CR4: 0000000000760ee0
>> [  128.188914] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>> [  128.188914] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>> [  128.188915] PKRU: 55555554
>> [  128.188915] Call Trace:
>> [  128.188916]  <IRQ>
>> [  128.188916]  _raw_read_lock_irqsave+0x26/0x30
>> [  128.188917]  ep_poll_callback+0x3e/0x2b0
>> [  128.188917]  ? cpumask_next_and+0x19/0x20
>> [  128.188918]  ? update_sd_lb_stats.constprop.0+0xfe/0x810
>> [  128.188918]  __wake_up_common+0x7a/0x140
>> [  128.188919]  __wake_up_common_lock+0x7c/0xc0
>> [  128.188919]  sock_def_readable+0x37/0x60
>> [  128.188920]  __udp_enqueue_schedule_skb+0x168/0x260
>> [  128.188920]  udpv6_queue_rcv_one_skb+0x284/0x3c0
>> [  128.188921]  udp6_unicast_rcv_skb.isra.0+0x44/0xa0
>> [  128.188921]  ip6_protocol_deliver_rcu+0x235/0x4b0
>> [  128.188922]  ip6_input_finish+0x11/0x20
>> [  128.188922]  ip6_input+0xa2/0xb0
>> [  128.188923]  ? ip6_protocol_deliver_rcu+0x4b0/0x4b0
>> [  128.188923]  ipv6_rcv+0xc0/0xd0
>> [  128.188924]  ? ip6_rcv_finish_core.isra.0+0xd0/0xd0
>> [  128.188924]  __netif_receive_skb_one_core+0x63/0xa0
>> [  128.188925]  process_backlog+0x98/0x140
>> [  128.188925]  net_rx_action+0x13a/0x370
>> [  128.188926]  __do_softirq+0xe0/0x2ca
>> [  128.188926]  do_softirq_own_stack+0x2a/0x40
>> [  128.188926]  </IRQ>
>> [  128.188927]  do_softirq.part.0+0x2b/0x30
>> [  128.188927]  __local_bh_enable_ip+0x4b/0x50
>> [  128.188928]  ip6_finish_output2+0x264/0x5b0
>> [  128.188928]  ip6_output+0x73/0x120
>> [  128.188929]  ? __ip6_finish_output+0x110/0x110
>> [  128.188929]  ip6_send_skb+0x1e/0x60
>> [  128.188930]  udp_v6_send_skb.isra.0+0x197/0x460
>> [  128.188930]  udpv6_sendmsg+0xb4f/0xdb0
>> [  128.188931]  ? ip_reply_glue_bits+0x40/0x40
>> [  128.188931]  ? update_load_avg+0x78/0x630
>> [  128.188932]  ? update_curr+0x73/0x1d0
>> [  128.188932]  ? __sys_sendto+0x108/0x190
>> [  128.188933]  __sys_sendto+0x108/0x190
>> [  128.188933]  ? __fput+0x1a5/0x240
>> [  128.188934]  ? _cond_resched+0x19/0x30
>> [  128.188934]  ? task_work_run+0x67/0x90
>> [  128.188935]  __x64_sys_sendto+0x25/0x30
>> [  128.188935]  do_syscall_64+0x48/0x130
>> [  128.188936]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>> [  128.188936] RIP: 0033:0x7f0b97a7826c
>> [  128.188945] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
>> [  128.188946] RSP: 002b:00007ffe80ffcea8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
>> [  128.188947] RAX: ffffffffffffffda RBX: 0000000000000058 RCX: 00007f0b97a7826c
>> [  128.188948] RDX: 0000000000000058 RSI: 000055fb765264c0 RDI: 0000000000000009
>> [  128.188949] RBP: 000055fb765264c0 R08: 0000000000000000 R09: 0000000000000000
>> [  128.188949] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffcfc0
>> [  128.188950] R13: 000000000002c600 R14: 0000000000000000 R15: 000055fb7629b751
>> [  128.188957] Sending NMI from CPU 10 to CPUs 19:
>> [  128.239939] NMI backtrace for cpu 19
>> [  128.239940] CPU: 19 PID: 2587 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
>> [  128.239940] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>> [  128.239940] RIP: 0010:kvm_wait+0x30/0x50
>> [  128.239941] Code: 8b 05 ac cb f4 7e a9 00 00 f0 00 75 1f 9c 58 fa 0f b6 17 40 38 d6 75 12 f6 c4 02 75 10 e9 07 00 00 00 0f 00 2d 6f ce 54 01 f4 <50> 9d c3 e9 07 00 00 00 0f 00 2d 5f ce 54 01 fb f4 eb ed 66 66 2e
>> [  128.239942] RSP: 0018:ffffc90000cc78e8 EFLAGS: 00000046
>> [  128.239942] RAX: 0000000000000046 RBX: ffff88903fcec700 RCX: 0000000000000008
>> [  128.239943] RDX: 0000000000000003 RSI: 0000000000000003 RDI: ffff889039049d80
>> [  128.239943] RBP: ffff889039049d80 R08: ffff88907ffe9f80 R09: 00000000000000f8
>> [  128.239944] R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000000
>> [  128.239944] R13: 0000000000000001 R14: 0000000000000100 R15: 0000000000500000
>> [  128.239944] FS:  00007f0b97743740(0000) GS:ffff88903fcc0000(0000) knlGS:0000000000000000
>> [  128.239945] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [  128.239945] CR2: 000055fb76a96030 CR3: 0000001035882004 CR4: 0000000000760ee0
>> [  128.239945] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>> [  128.239946] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>> [  128.239946] PKRU: 55555554
>> [  128.239946] Call Trace:
>> [  128.239946]  __pv_queued_spin_lock_slowpath+0x26e/0x2b0
>> [  128.239947]  _raw_spin_lock_irqsave+0x25/0x30
>> [  128.239947]  __wake_up_common_lock+0x62/0xc0
>> [  128.239947]  sock_def_write_space+0x49/0x90
>> [  128.239948]  sock_wfree+0x68/0xb0
>> [  128.239948]  loopback_xmit+0x49/0xe0
>> [  128.239948]  dev_hard_start_xmit+0x8d/0x1e0
>> [  128.239948]  __dev_queue_xmit+0x721/0x8e0
>> [  128.239949]  ip6_finish_output2+0x250/0x5b0
>> [  128.239949]  ip6_output+0x73/0x120
>> [  128.239949]  ? __ip6_finish_output+0x110/0x110
>> [  128.239950]  ip6_send_skb+0x1e/0x60
>> [  128.239950]  udp_v6_send_skb.isra.0+0x197/0x460
>> [  128.239950]  udpv6_sendmsg+0xb4f/0xdb0
>> [  128.239950]  ? release_pages+0x28f/0x2f0
>> [  128.239950]  ? ip_reply_glue_bits+0x40/0x40
>> [  128.239951]  ? _cond_resched+0x19/0x30
>> [  128.239951]  ? unmap_page_range+0x678/0xa60
>> [  128.239951]  ? __sys_sendto+0x108/0x190
>> [  128.239951]  __sys_sendto+0x108/0x190
>> [  128.239952]  ? __fput+0x1a5/0x240
>> [  128.239952]  ? _cond_resched+0x19/0x30
>> [  128.239952]  ? task_work_run+0x67/0x90
>> [  128.239952]  __x64_sys_sendto+0x25/0x30
>> [  128.239953]  do_syscall_64+0x48/0x130
>> [  128.239953]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>> [  128.239953] RIP: 0033:0x7f0b97a7826c
>> [  128.239954] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
>> [  128.239954] RSP: 002b:00007ffe80ffd7b8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
>> [  128.239955] RAX: ffffffffffffffda RBX: 00000000000003a8 RCX: 00007f0b97a7826c
>> [  128.239955] RDX: 00000000000003a8 RSI: 00007ffe80ffd800 RDI: 0000000000000009
>> [  128.239955] RBP: 00007ffe80ffd800 R08: 0000000000000000 R09: 0000000000000000
>> [  128.239956] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffd800
>> [  128.239956] R13: 00007ffe80ffd800 R14: 000000000000000e R15: 000055fb76b37e58
> 
> I can reproduce this, and I see what it is. I'll send out a patch soonish.

Thinko, can you try with this on top?


diff --git a/fs/io_uring.c b/fs/io_uring.c
index d22830a423f1..ca96ece3ac18 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -2787,9 +2787,11 @@ static int io_read(struct io_kiocb *req, bool force_nonblock)
 			/* if we can retry, do so with the callbacks armed */
 			if (io_rw_should_retry(req)) {
 				ret2 = io_iter_do_read(req, &iter);
-				if (ret2 != -EAGAIN) {
+				if (ret2 == -EIOCBQUEUED) {
+					return 0;
+				} else if (ret2 != -EAGAIN) {
 					kiocb_done(kiocb, ret2);
-					goto out_free;
+					return 0;
 				}
 			}
 			kiocb->ki_flags &= ~IOCB_WAITQ;

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/5/20 3:13 PM, Jens Axboe wrote:
> On 6/5/20 2:53 PM, Jens Axboe wrote:
>> On 6/5/20 2:36 PM, Andres Freund wrote:
>>> Hi,
>>>
>>> On 2020-06-05 13:20:28 -0700, Andres Freund wrote:
>>>> I'll go and try to figure out why I don't see an oops...
>>>
>>> Err, that probably was a typo on my end in the serial console
>>> config. After fixing that, I did get the below.
>>>
>>> If helpful I can try with debugging enabled or such.
>>>
>>> [   67.910265] tee (2577): drop_caches: 3
>>> [   68.059674] BUG: unable to handle page fault for address: 00007f0b16a3c03c
>>> [   68.062021] #PF: supervisor read access in kernel mode
>>> [   68.063742] #PF: error_code(0x0000) - not-present page
>>> [   68.065517] PGD 102e044067 P4D 102e044067 PUD 102bf7a067 PMD 0 
>>> [   68.067519] Oops: 0000 [#1] SMP NOPTI
>>> [   68.068800] CPU: 2 PID: 2554 Comm: postgres Not tainted 5.7.0-andres-10123-g87823242260e #44
>>> [   68.071505] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>>> [   68.074139] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
>>> [   68.075389] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
>>> [   68.079125] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
>>> [   68.080260] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
>>> [   68.084115] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
>>> [   68.085374] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
>>> [   68.086409] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
>>> [   68.087447] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
>>> [   68.088697] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>>> [   68.089903] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>> [   68.090776] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
>>> [   68.091834] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>> [   68.092902] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>> [   68.093967] PKRU: 55555554
>>> [   68.094427] Call Trace:
>>> [   68.094845]  ? __schedule+0x2ac/0x720
>>> [   68.095350]  ? schedule+0x55/0xc0
>>> [   68.095797]  ? ep_read_events_proc+0xd0/0xd0
>>> [   68.096354]  ep_scan_ready_list.constprop.0+0x16c/0x190
>>> [   68.097016]  ep_poll+0x2a3/0x440
>>> [   68.097449]  ? wait_woken+0x70/0x70
>>> [   68.097904]  do_epoll_wait+0xb0/0xd0
>>> [   68.098375]  __x64_sys_epoll_wait+0x1a/0x20
>>> [   68.098913]  do_syscall_64+0x48/0x130
>>> [   68.099393]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>> [   68.100030] RIP: 0033:0x7f0b97845606
>>> [   68.100498] Code: 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 11 b8 e8 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 5a c3 90 48 83 ec 28 89 54 24 18 48 89 74 24
>>> [   68.102718] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
>>> [   68.103644] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
>>> [   68.104533] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
>>> [   68.105418] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
>>> [   68.106296] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
>>> [   68.107187] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
>>> [   68.108104] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
>>> [   68.109505] CR2: 00007f0b16a3c03c
>>> [   68.109962] ---[ end trace 0ca39a5ed99162ce ]---
>>> [   68.110547] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
>>> [   68.111214] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
>>> [   68.113435] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
>>> [   68.114111] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
>>> [   68.115016] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
>>> [   68.115902] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
>>> [   68.116810] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
>>> [   68.117663] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
>>> [   68.118520] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>>> [   68.119482] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>> [   68.120181] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
>>> [   68.121043] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>> [   68.121904] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>> [   68.122790] PKRU: 55555554
>>> [   68.123401] general protection fault, probably for non-canonical address 0xfeeda989fef06266: 0000 [#2] SMP NOPTI
>>> [   68.125052] CPU: 2 PID: 2554 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
>>> [   68.126260] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>>> [   68.127274] RIP: 0010:__pv_queued_spin_lock_slowpath+0x1a1/0x2b0
>>> [   68.128032] Code: c4 c1 ea 12 41 be 01 00 00 00 8d 42 ff 41 83 e4 03 4c 8d 6b 14 49 c1 e4 05 48 98 49 81 c4 00 c7 02 00 4c 03 24 c5 e0 e6 fd 82 <49> 89 1c 24 b8 00 80 00 00 eb 15 84 c0 75 0a 41 0f b6 54 24 14 84
>>> [   68.130221] RSP: 0018:ffffc90000befce8 EFLAGS: 00010086
>>> [   68.130867] RAX: 0000000000003ffe RBX: ffff88903f8ac700 RCX: 0000000000000001
>>> [   68.131752] RDX: 0000000000003fff RSI: 0000000000000000 RDI: 0000000000000000
>>> [   68.132637] RBP: ffff889037617924 R08: 0000000000000000 R09: ffffc90000befdf8
>>> [   68.133513] R10: ffff8890334d2bf0 R11: 0000000000000018 R12: feeda989fef06266
>>> [   68.134399] R13: ffff88903f8ac714 R14: 0000000000000001 R15: 00000000000c0000
>>> [   68.135323] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>>> [   68.136307] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>> [   68.137004] CR2: 00007f0b16a3c03c CR3: 000000000360a005 CR4: 0000000000760ee0
>>> [   68.137866] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>> [   68.138748] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>> [   68.139606] PKRU: 55555554
>>> [   68.139973] Call Trace:
>>> [   68.140313]  queued_read_lock_slowpath+0x6c/0x70
>>> [   68.140915]  _raw_read_lock_irqsave+0x26/0x30
>>> [   68.141485]  ep_poll_callback+0x3e/0x2b0
>>> [   68.142007]  ? set_next_entity+0xab/0x1f0
>>> [   68.142541]  __wake_up_common+0x7a/0x140
>>> [   68.143077]  __wake_up_common_lock+0x7c/0xc0
>>> [   68.143651]  pipe_release+0x5b/0xd0
>>> [   68.144150]  __fput+0xda/0x240
>>> [   68.144574]  task_work_run+0x62/0x90
>>> [   68.145046]  do_exit+0x35c/0xa70
>>> [   68.145505]  ? do_epoll_wait+0xb0/0xd0
>>> [   68.146000]  rewind_stack_do_exit+0x17/0x20
>>> [   68.146538] RIP: 0033:0x7f0b97845606
>>> [   68.146988] Code: Bad RIP value.
>>> [   68.147405] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
>>> [   68.148314] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
>>> [   68.149157] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
>>> [   68.150021] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
>>> [   68.150831] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
>>> [   68.151640] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
>>> [   68.152459] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
>>> [   68.153707] ---[ end trace 0ca39a5ed99162cf ]---
>>> [   68.154282] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
>>> [   68.154884] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
>>> [   68.156976] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
>>> [   68.157614] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
>>> [   68.158436] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
>>> [   68.159269] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
>>> [   68.160092] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
>>> [   68.160920] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
>>> [   68.161746] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>>> [   68.162701] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>> [   68.163386] CR2: 00007f0b978455dc CR3: 000000000360a005 CR4: 0000000000760ee0
>>> [   68.164226] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>> [   68.165079] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>> [   68.165931] PKRU: 55555554
>>> [   68.166298] Fixing recursive fault but reboot is needed!
>>> [  128.173729] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
>>> [  128.179267] rcu: 	11-...0: (1 GPs behind) idle=c56/1/0x4000000000000000 softirq=6897/6898 fqs=5976 
>>> [  128.182232] rcu: 	19-...0: (1 GPs behind) idle=492/1/0x4000000000000000 softirq=1023/1023 fqs=5976 
>>> [  128.185217] 	(detected by 10, t=18003 jiffies, g=15789, q=631)
>>> [  128.186863] Sending NMI from CPU 10 to CPUs 11:
>>> [  128.188902] NMI backtrace for cpu 11
>>> [  128.188903] CPU: 11 PID: 2546 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
>>> [  128.188904] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>>> [  128.188904] RIP: 0010:queued_read_lock_slowpath+0x5b/0x70
>>> [  128.188906] Code: 03 00 02 00 00 8b 03 84 c0 74 08 f3 90 8b 13 84 d2 75 f8 48 89 ef e8 74 e6 ff ff 66 90 5b 5d c3 8b 07 84 c0 74 08 f3 90 8b 03 <84> c0 75 f8 5b 5d c3 89 c6 48 89 ef e8 e4 e8 ff ff 66 90 eb bf 0f
>>> [  128.188907] RSP: 0018:ffffc90000348c08 EFLAGS: 00000086
>>> [  128.188908] RAX: 0000000037617cc0 RBX: ffff889037617920 RCX: 00000000000000c3
>>> [  128.188909] RDX: 0000000000000001 RSI: 0000000000000001 RDI: ffff889037617920
>>> [  128.188910] RBP: ffff889037bb1a80 R08: 00000000000000c3 R09: ffffc90000348cd8
>>> [  128.188910] R10: 0100000000000000 R11: 00000000de2ee17e R12: 0000000000000046
>>> [  128.188911] R13: ffff889037617920 R14: 0000000000000001 R15: 00000000000000c3
>>> [  128.188912] FS:  00007f0b97743740(0000) GS:ffff88903fac0000(0000) knlGS:0000000000000000
>>> [  128.188912] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>> [  128.188913] CR2: 00007f0b16afc00e CR3: 000000102e5c6002 CR4: 0000000000760ee0
>>> [  128.188914] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>> [  128.188914] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>> [  128.188915] PKRU: 55555554
>>> [  128.188915] Call Trace:
>>> [  128.188916]  <IRQ>
>>> [  128.188916]  _raw_read_lock_irqsave+0x26/0x30
>>> [  128.188917]  ep_poll_callback+0x3e/0x2b0
>>> [  128.188917]  ? cpumask_next_and+0x19/0x20
>>> [  128.188918]  ? update_sd_lb_stats.constprop.0+0xfe/0x810
>>> [  128.188918]  __wake_up_common+0x7a/0x140
>>> [  128.188919]  __wake_up_common_lock+0x7c/0xc0
>>> [  128.188919]  sock_def_readable+0x37/0x60
>>> [  128.188920]  __udp_enqueue_schedule_skb+0x168/0x260
>>> [  128.188920]  udpv6_queue_rcv_one_skb+0x284/0x3c0
>>> [  128.188921]  udp6_unicast_rcv_skb.isra.0+0x44/0xa0
>>> [  128.188921]  ip6_protocol_deliver_rcu+0x235/0x4b0
>>> [  128.188922]  ip6_input_finish+0x11/0x20
>>> [  128.188922]  ip6_input+0xa2/0xb0
>>> [  128.188923]  ? ip6_protocol_deliver_rcu+0x4b0/0x4b0
>>> [  128.188923]  ipv6_rcv+0xc0/0xd0
>>> [  128.188924]  ? ip6_rcv_finish_core.isra.0+0xd0/0xd0
>>> [  128.188924]  __netif_receive_skb_one_core+0x63/0xa0
>>> [  128.188925]  process_backlog+0x98/0x140
>>> [  128.188925]  net_rx_action+0x13a/0x370
>>> [  128.188926]  __do_softirq+0xe0/0x2ca
>>> [  128.188926]  do_softirq_own_stack+0x2a/0x40
>>> [  128.188926]  </IRQ>
>>> [  128.188927]  do_softirq.part.0+0x2b/0x30
>>> [  128.188927]  __local_bh_enable_ip+0x4b/0x50
>>> [  128.188928]  ip6_finish_output2+0x264/0x5b0
>>> [  128.188928]  ip6_output+0x73/0x120
>>> [  128.188929]  ? __ip6_finish_output+0x110/0x110
>>> [  128.188929]  ip6_send_skb+0x1e/0x60
>>> [  128.188930]  udp_v6_send_skb.isra.0+0x197/0x460
>>> [  128.188930]  udpv6_sendmsg+0xb4f/0xdb0
>>> [  128.188931]  ? ip_reply_glue_bits+0x40/0x40
>>> [  128.188931]  ? update_load_avg+0x78/0x630
>>> [  128.188932]  ? update_curr+0x73/0x1d0
>>> [  128.188932]  ? __sys_sendto+0x108/0x190
>>> [  128.188933]  __sys_sendto+0x108/0x190
>>> [  128.188933]  ? __fput+0x1a5/0x240
>>> [  128.188934]  ? _cond_resched+0x19/0x30
>>> [  128.188934]  ? task_work_run+0x67/0x90
>>> [  128.188935]  __x64_sys_sendto+0x25/0x30
>>> [  128.188935]  do_syscall_64+0x48/0x130
>>> [  128.188936]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>> [  128.188936] RIP: 0033:0x7f0b97a7826c
>>> [  128.188945] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
>>> [  128.188946] RSP: 002b:00007ffe80ffcea8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
>>> [  128.188947] RAX: ffffffffffffffda RBX: 0000000000000058 RCX: 00007f0b97a7826c
>>> [  128.188948] RDX: 0000000000000058 RSI: 000055fb765264c0 RDI: 0000000000000009
>>> [  128.188949] RBP: 000055fb765264c0 R08: 0000000000000000 R09: 0000000000000000
>>> [  128.188949] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffcfc0
>>> [  128.188950] R13: 000000000002c600 R14: 0000000000000000 R15: 000055fb7629b751
>>> [  128.188957] Sending NMI from CPU 10 to CPUs 19:
>>> [  128.239939] NMI backtrace for cpu 19
>>> [  128.239940] CPU: 19 PID: 2587 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
>>> [  128.239940] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>>> [  128.239940] RIP: 0010:kvm_wait+0x30/0x50
>>> [  128.239941] Code: 8b 05 ac cb f4 7e a9 00 00 f0 00 75 1f 9c 58 fa 0f b6 17 40 38 d6 75 12 f6 c4 02 75 10 e9 07 00 00 00 0f 00 2d 6f ce 54 01 f4 <50> 9d c3 e9 07 00 00 00 0f 00 2d 5f ce 54 01 fb f4 eb ed 66 66 2e
>>> [  128.239942] RSP: 0018:ffffc90000cc78e8 EFLAGS: 00000046
>>> [  128.239942] RAX: 0000000000000046 RBX: ffff88903fcec700 RCX: 0000000000000008
>>> [  128.239943] RDX: 0000000000000003 RSI: 0000000000000003 RDI: ffff889039049d80
>>> [  128.239943] RBP: ffff889039049d80 R08: ffff88907ffe9f80 R09: 00000000000000f8
>>> [  128.239944] R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000000
>>> [  128.239944] R13: 0000000000000001 R14: 0000000000000100 R15: 0000000000500000
>>> [  128.239944] FS:  00007f0b97743740(0000) GS:ffff88903fcc0000(0000) knlGS:0000000000000000
>>> [  128.239945] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>> [  128.239945] CR2: 000055fb76a96030 CR3: 0000001035882004 CR4: 0000000000760ee0
>>> [  128.239945] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>> [  128.239946] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>> [  128.239946] PKRU: 55555554
>>> [  128.239946] Call Trace:
>>> [  128.239946]  __pv_queued_spin_lock_slowpath+0x26e/0x2b0
>>> [  128.239947]  _raw_spin_lock_irqsave+0x25/0x30
>>> [  128.239947]  __wake_up_common_lock+0x62/0xc0
>>> [  128.239947]  sock_def_write_space+0x49/0x90
>>> [  128.239948]  sock_wfree+0x68/0xb0
>>> [  128.239948]  loopback_xmit+0x49/0xe0
>>> [  128.239948]  dev_hard_start_xmit+0x8d/0x1e0
>>> [  128.239948]  __dev_queue_xmit+0x721/0x8e0
>>> [  128.239949]  ip6_finish_output2+0x250/0x5b0
>>> [  128.239949]  ip6_output+0x73/0x120
>>> [  128.239949]  ? __ip6_finish_output+0x110/0x110
>>> [  128.239950]  ip6_send_skb+0x1e/0x60
>>> [  128.239950]  udp_v6_send_skb.isra.0+0x197/0x460
>>> [  128.239950]  udpv6_sendmsg+0xb4f/0xdb0
>>> [  128.239950]  ? release_pages+0x28f/0x2f0
>>> [  128.239950]  ? ip_reply_glue_bits+0x40/0x40
>>> [  128.239951]  ? _cond_resched+0x19/0x30
>>> [  128.239951]  ? unmap_page_range+0x678/0xa60
>>> [  128.239951]  ? __sys_sendto+0x108/0x190
>>> [  128.239951]  __sys_sendto+0x108/0x190
>>> [  128.239952]  ? __fput+0x1a5/0x240
>>> [  128.239952]  ? _cond_resched+0x19/0x30
>>> [  128.239952]  ? task_work_run+0x67/0x90
>>> [  128.239952]  __x64_sys_sendto+0x25/0x30
>>> [  128.239953]  do_syscall_64+0x48/0x130
>>> [  128.239953]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>> [  128.239953] RIP: 0033:0x7f0b97a7826c
>>> [  128.239954] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
>>> [  128.239954] RSP: 002b:00007ffe80ffd7b8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
>>> [  128.239955] RAX: ffffffffffffffda RBX: 00000000000003a8 RCX: 00007f0b97a7826c
>>> [  128.239955] RDX: 00000000000003a8 RSI: 00007ffe80ffd800 RDI: 0000000000000009
>>> [  128.239955] RBP: 00007ffe80ffd800 R08: 0000000000000000 R09: 0000000000000000
>>> [  128.239956] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffd800
>>> [  128.239956] R13: 00007ffe80ffd800 R14: 000000000000000e R15: 000055fb76b37e58
>>
>> I can reproduce this, and I see what it is. I'll send out a patch soonish.
> 
> Thinko, can you try with this on top?

Sorry that was incomplete, please use this one!
 
diff --git a/fs/io_uring.c b/fs/io_uring.c
index d22830a423f1..0075107dd09a 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -2779,7 +2779,7 @@ static int io_read(struct io_kiocb *req, bool force_nonblock)
 			ret = io_setup_async_rw(req, io_size, iovec,
 						inline_vecs, &iter);
 			if (ret)
-				goto out_free;
+				goto out;
 			/* any defer here is final, must blocking retry */
 			if (!(req->flags & REQ_F_NOWAIT) &&
 			    !file_can_poll(req->file))
@@ -2787,18 +2787,18 @@ static int io_read(struct io_kiocb *req, bool force_nonblock)
 			/* if we can retry, do so with the callbacks armed */
 			if (io_rw_should_retry(req)) {
 				ret2 = io_iter_do_read(req, &iter);
-				if (ret2 != -EAGAIN) {
+				if (ret2 == -EIOCBQUEUED) {
+					goto out;
+				} else if (ret2 != -EAGAIN) {
 					kiocb_done(kiocb, ret2);
-					goto out_free;
+					goto out;
 				}
 			}
 			kiocb->ki_flags &= ~IOCB_WAITQ;
 			return -EAGAIN;
 		}
 	}
-out_free:
-	kfree(iovec);
-	req->flags &= ~REQ_F_NEED_CLEANUP;
+out:
 	return ret;
 }
 
-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Andres Freund]

Hi,

On 2020-06-05 15:21:34 -0600, Jens Axboe wrote:
> >> I can reproduce this, and I see what it is. I'll send out a patch soonish.
> > 
> > Thinko, can you try with this on top?
> 
> Sorry that was incomplete, please use this one!

That seems to fix it! Yay.


Bulk buffered reads somehow don't quite seem to be performing that well
though, looking into it. Could be on the pg side too.

Greetings,

Andres Freund

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Andres Freund]

Hi,

On 2020-06-05 15:30:44 -0700, Andres Freund wrote:
> On 2020-06-05 15:21:34 -0600, Jens Axboe wrote:
> > >> I can reproduce this, and I see what it is. I'll send out a patch soonish.
> > > 
> > > Thinko, can you try with this on top?
> > 
> > Sorry that was incomplete, please use this one!
> 
> That seems to fix it! Yay.
> 
> 
> Bulk buffered reads somehow don't quite seem to be performing that well
> though, looking into it. Could be on the pg side too.

While looking into that, I played with setting
/sys/<dev>/queue/read_ahead_kb to 0 and noticed that seems to result in
all/most IO done in workers. Is that to be expected?

Greetings,

Andres Freund

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/5/20 4:36 PM, Andres Freund wrote:
> Hi,
> 
> On 2020-06-05 15:30:44 -0700, Andres Freund wrote:
>> On 2020-06-05 15:21:34 -0600, Jens Axboe wrote:
>>>>> I can reproduce this, and I see what it is. I'll send out a patch soonish.
>>>>
>>>> Thinko, can you try with this on top?
>>>
>>> Sorry that was incomplete, please use this one!
>>
>> That seems to fix it! Yay.
>>
>>
>> Bulk buffered reads somehow don't quite seem to be performing that well
>> though, looking into it. Could be on the pg side too.
> 
> While looking into that, I played with setting
> /sys/<dev>/queue/read_ahead_kb to 0 and noticed that seems to result in
> all/most IO done in workers. Is that to be expected?

Yes that's expected, if we have to fallback to ->readpage(), then it'll
go to a worker. read-ahead is what drives the async nature of it, as we
issue the range (plus more, depending on RA window) as read-ahead for
the normal read, then wait for it.

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Andres Freund]

Hi,

On 2020-06-05 16:49:24 -0600, Jens Axboe wrote:
> Yes that's expected, if we have to fallback to ->readpage(), then it'll
> go to a worker. read-ahead is what drives the async nature of it, as we
> issue the range (plus more, depending on RA window) as read-ahead for
> the normal read, then wait for it.

But I assume async would still work for files with POSIX_FADV_RANDOM
set, or not? Assuming the system wide setting isn't zero, of course.

Greetings,

Andres Freund

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/5/20 4:54 PM, Andres Freund wrote:
> Hi,
> 
> On 2020-06-05 16:49:24 -0600, Jens Axboe wrote:
>> Yes that's expected, if we have to fallback to ->readpage(), then it'll
>> go to a worker. read-ahead is what drives the async nature of it, as we
>> issue the range (plus more, depending on RA window) as read-ahead for
>> the normal read, then wait for it.
> 
> But I assume async would still work for files with POSIX_FADV_RANDOM
> set, or not? Assuming the system wide setting isn't zero, of course.

Yes it'll work if FADV_RANDOM is set. But just not if read-ahead is
totally disabled. I guess we could make that work too, though not sure
that it's super important.

-- 
Jens Axboe

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Andres Freund]

Hi,

On 2020-06-05 16:56:44 -0600, Jens Axboe wrote:
> On 6/5/20 4:54 PM, Andres Freund wrote:
> > On 2020-06-05 16:49:24 -0600, Jens Axboe wrote:
> >> Yes that's expected, if we have to fallback to ->readpage(), then it'll
> >> go to a worker. read-ahead is what drives the async nature of it, as we
> >> issue the range (plus more, depending on RA window) as read-ahead for
> >> the normal read, then wait for it.
> > 
> > But I assume async would still work for files with POSIX_FADV_RANDOM
> > set, or not? Assuming the system wide setting isn't zero, of course.
> 
> Yes it'll work if FADV_RANDOM is set.

Cool.


> But just not if read-ahead is totally disabled. I guess we could make
> that work too, though not sure that it's super important.

It's not from my end. I was just trying to check if the reduced
performance I saw was related to interactions between PG prefetching and
kernel level prefetching. And changing the /sys entry seemed easier than
making postgres set POSIX_FADV_RANDOM...

Greetings,

Andres Freund

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Sedat Dilek]

On Fri, Jun 5, 2020 at 11:24 PM Jens Axboe <axboe@kernel.dk> wrote:
>
> On 6/5/20 3:13 PM, Jens Axboe wrote:
> > On 6/5/20 2:53 PM, Jens Axboe wrote:
> >> On 6/5/20 2:36 PM, Andres Freund wrote:
> >>> Hi,
> >>>
> >>> On 2020-06-05 13:20:28 -0700, Andres Freund wrote:
> >>>> I'll go and try to figure out why I don't see an oops...
> >>>
> >>> Err, that probably was a typo on my end in the serial console
> >>> config. After fixing that, I did get the below.
> >>>
> >>> If helpful I can try with debugging enabled or such.
> >>>
> >>> [   67.910265] tee (2577): drop_caches: 3
> >>> [   68.059674] BUG: unable to handle page fault for address: 00007f0b16a3c03c
> >>> [   68.062021] #PF: supervisor read access in kernel mode
> >>> [   68.063742] #PF: error_code(0x0000) - not-present page
> >>> [   68.065517] PGD 102e044067 P4D 102e044067 PUD 102bf7a067 PMD 0
> >>> [   68.067519] Oops: 0000 [#1] SMP NOPTI
> >>> [   68.068800] CPU: 2 PID: 2554 Comm: postgres Not tainted 5.7.0-andres-10123-g87823242260e #44
> >>> [   68.071505] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
> >>> [   68.074139] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
> >>> [   68.075389] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
> >>> [   68.079125] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
> >>> [   68.080260] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
> >>> [   68.084115] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
> >>> [   68.085374] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
> >>> [   68.086409] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
> >>> [   68.087447] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
> >>> [   68.088697] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
> >>> [   68.089903] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >>> [   68.090776] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
> >>> [   68.091834] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> >>> [   68.092902] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> >>> [   68.093967] PKRU: 55555554
> >>> [   68.094427] Call Trace:
> >>> [   68.094845]  ? __schedule+0x2ac/0x720
> >>> [   68.095350]  ? schedule+0x55/0xc0
> >>> [   68.095797]  ? ep_read_events_proc+0xd0/0xd0
> >>> [   68.096354]  ep_scan_ready_list.constprop.0+0x16c/0x190
> >>> [   68.097016]  ep_poll+0x2a3/0x440
> >>> [   68.097449]  ? wait_woken+0x70/0x70
> >>> [   68.097904]  do_epoll_wait+0xb0/0xd0
> >>> [   68.098375]  __x64_sys_epoll_wait+0x1a/0x20
> >>> [   68.098913]  do_syscall_64+0x48/0x130
> >>> [   68.099393]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> >>> [   68.100030] RIP: 0033:0x7f0b97845606
> >>> [   68.100498] Code: 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 11 b8 e8 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 5a c3 90 48 83 ec 28 89 54 24 18 48 89 74 24
> >>> [   68.102718] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
> >>> [   68.103644] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
> >>> [   68.104533] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
> >>> [   68.105418] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
> >>> [   68.106296] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
> >>> [   68.107187] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
> >>> [   68.108104] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
> >>> [   68.109505] CR2: 00007f0b16a3c03c
> >>> [   68.109962] ---[ end trace 0ca39a5ed99162ce ]---
> >>> [   68.110547] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
> >>> [   68.111214] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
> >>> [   68.113435] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
> >>> [   68.114111] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
> >>> [   68.115016] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
> >>> [   68.115902] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
> >>> [   68.116810] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
> >>> [   68.117663] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
> >>> [   68.118520] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
> >>> [   68.119482] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >>> [   68.120181] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
> >>> [   68.121043] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> >>> [   68.121904] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> >>> [   68.122790] PKRU: 55555554
> >>> [   68.123401] general protection fault, probably for non-canonical address 0xfeeda989fef06266: 0000 [#2] SMP NOPTI
> >>> [   68.125052] CPU: 2 PID: 2554 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
> >>> [   68.126260] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
> >>> [   68.127274] RIP: 0010:__pv_queued_spin_lock_slowpath+0x1a1/0x2b0
> >>> [   68.128032] Code: c4 c1 ea 12 41 be 01 00 00 00 8d 42 ff 41 83 e4 03 4c 8d 6b 14 49 c1 e4 05 48 98 49 81 c4 00 c7 02 00 4c 03 24 c5 e0 e6 fd 82 <49> 89 1c 24 b8 00 80 00 00 eb 15 84 c0 75 0a 41 0f b6 54 24 14 84
> >>> [   68.130221] RSP: 0018:ffffc90000befce8 EFLAGS: 00010086
> >>> [   68.130867] RAX: 0000000000003ffe RBX: ffff88903f8ac700 RCX: 0000000000000001
> >>> [   68.131752] RDX: 0000000000003fff RSI: 0000000000000000 RDI: 0000000000000000
> >>> [   68.132637] RBP: ffff889037617924 R08: 0000000000000000 R09: ffffc90000befdf8
> >>> [   68.133513] R10: ffff8890334d2bf0 R11: 0000000000000018 R12: feeda989fef06266
> >>> [   68.134399] R13: ffff88903f8ac714 R14: 0000000000000001 R15: 00000000000c0000
> >>> [   68.135323] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
> >>> [   68.136307] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >>> [   68.137004] CR2: 00007f0b16a3c03c CR3: 000000000360a005 CR4: 0000000000760ee0
> >>> [   68.137866] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> >>> [   68.138748] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> >>> [   68.139606] PKRU: 55555554
> >>> [   68.139973] Call Trace:
> >>> [   68.140313]  queued_read_lock_slowpath+0x6c/0x70
> >>> [   68.140915]  _raw_read_lock_irqsave+0x26/0x30
> >>> [   68.141485]  ep_poll_callback+0x3e/0x2b0
> >>> [   68.142007]  ? set_next_entity+0xab/0x1f0
> >>> [   68.142541]  __wake_up_common+0x7a/0x140
> >>> [   68.143077]  __wake_up_common_lock+0x7c/0xc0
> >>> [   68.143651]  pipe_release+0x5b/0xd0
> >>> [   68.144150]  __fput+0xda/0x240
> >>> [   68.144574]  task_work_run+0x62/0x90
> >>> [   68.145046]  do_exit+0x35c/0xa70
> >>> [   68.145505]  ? do_epoll_wait+0xb0/0xd0
> >>> [   68.146000]  rewind_stack_do_exit+0x17/0x20
> >>> [   68.146538] RIP: 0033:0x7f0b97845606
> >>> [   68.146988] Code: Bad RIP value.
> >>> [   68.147405] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
> >>> [   68.148314] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
> >>> [   68.149157] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
> >>> [   68.150021] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
> >>> [   68.150831] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
> >>> [   68.151640] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
> >>> [   68.152459] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
> >>> [   68.153707] ---[ end trace 0ca39a5ed99162cf ]---
> >>> [   68.154282] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
> >>> [   68.154884] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
> >>> [   68.156976] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
> >>> [   68.157614] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
> >>> [   68.158436] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
> >>> [   68.159269] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
> >>> [   68.160092] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
> >>> [   68.160920] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
> >>> [   68.161746] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
> >>> [   68.162701] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >>> [   68.163386] CR2: 00007f0b978455dc CR3: 000000000360a005 CR4: 0000000000760ee0
> >>> [   68.164226] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> >>> [   68.165079] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> >>> [   68.165931] PKRU: 55555554
> >>> [   68.166298] Fixing recursive fault but reboot is needed!
> >>> [  128.173729] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
> >>> [  128.179267] rcu:         11-...0: (1 GPs behind) idle=c56/1/0x4000000000000000 softirq=6897/6898 fqs=5976
> >>> [  128.182232] rcu:         19-...0: (1 GPs behind) idle=492/1/0x4000000000000000 softirq=1023/1023 fqs=5976
> >>> [  128.185217]      (detected by 10, t=18003 jiffies, g=15789, q=631)
> >>> [  128.186863] Sending NMI from CPU 10 to CPUs 11:
> >>> [  128.188902] NMI backtrace for cpu 11
> >>> [  128.188903] CPU: 11 PID: 2546 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
> >>> [  128.188904] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
> >>> [  128.188904] RIP: 0010:queued_read_lock_slowpath+0x5b/0x70
> >>> [  128.188906] Code: 03 00 02 00 00 8b 03 84 c0 74 08 f3 90 8b 13 84 d2 75 f8 48 89 ef e8 74 e6 ff ff 66 90 5b 5d c3 8b 07 84 c0 74 08 f3 90 8b 03 <84> c0 75 f8 5b 5d c3 89 c6 48 89 ef e8 e4 e8 ff ff 66 90 eb bf 0f
> >>> [  128.188907] RSP: 0018:ffffc90000348c08 EFLAGS: 00000086
> >>> [  128.188908] RAX: 0000000037617cc0 RBX: ffff889037617920 RCX: 00000000000000c3
> >>> [  128.188909] RDX: 0000000000000001 RSI: 0000000000000001 RDI: ffff889037617920
> >>> [  128.188910] RBP: ffff889037bb1a80 R08: 00000000000000c3 R09: ffffc90000348cd8
> >>> [  128.188910] R10: 0100000000000000 R11: 00000000de2ee17e R12: 0000000000000046
> >>> [  128.188911] R13: ffff889037617920 R14: 0000000000000001 R15: 00000000000000c3
> >>> [  128.188912] FS:  00007f0b97743740(0000) GS:ffff88903fac0000(0000) knlGS:0000000000000000
> >>> [  128.188912] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >>> [  128.188913] CR2: 00007f0b16afc00e CR3: 000000102e5c6002 CR4: 0000000000760ee0
> >>> [  128.188914] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> >>> [  128.188914] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> >>> [  128.188915] PKRU: 55555554
> >>> [  128.188915] Call Trace:
> >>> [  128.188916]  <IRQ>
> >>> [  128.188916]  _raw_read_lock_irqsave+0x26/0x30
> >>> [  128.188917]  ep_poll_callback+0x3e/0x2b0
> >>> [  128.188917]  ? cpumask_next_and+0x19/0x20
> >>> [  128.188918]  ? update_sd_lb_stats.constprop.0+0xfe/0x810
> >>> [  128.188918]  __wake_up_common+0x7a/0x140
> >>> [  128.188919]  __wake_up_common_lock+0x7c/0xc0
> >>> [  128.188919]  sock_def_readable+0x37/0x60
> >>> [  128.188920]  __udp_enqueue_schedule_skb+0x168/0x260
> >>> [  128.188920]  udpv6_queue_rcv_one_skb+0x284/0x3c0
> >>> [  128.188921]  udp6_unicast_rcv_skb.isra.0+0x44/0xa0
> >>> [  128.188921]  ip6_protocol_deliver_rcu+0x235/0x4b0
> >>> [  128.188922]  ip6_input_finish+0x11/0x20
> >>> [  128.188922]  ip6_input+0xa2/0xb0
> >>> [  128.188923]  ? ip6_protocol_deliver_rcu+0x4b0/0x4b0
> >>> [  128.188923]  ipv6_rcv+0xc0/0xd0
> >>> [  128.188924]  ? ip6_rcv_finish_core.isra.0+0xd0/0xd0
> >>> [  128.188924]  __netif_receive_skb_one_core+0x63/0xa0
> >>> [  128.188925]  process_backlog+0x98/0x140
> >>> [  128.188925]  net_rx_action+0x13a/0x370
> >>> [  128.188926]  __do_softirq+0xe0/0x2ca
> >>> [  128.188926]  do_softirq_own_stack+0x2a/0x40
> >>> [  128.188926]  </IRQ>
> >>> [  128.188927]  do_softirq.part.0+0x2b/0x30
> >>> [  128.188927]  __local_bh_enable_ip+0x4b/0x50
> >>> [  128.188928]  ip6_finish_output2+0x264/0x5b0
> >>> [  128.188928]  ip6_output+0x73/0x120
> >>> [  128.188929]  ? __ip6_finish_output+0x110/0x110
> >>> [  128.188929]  ip6_send_skb+0x1e/0x60
> >>> [  128.188930]  udp_v6_send_skb.isra.0+0x197/0x460
> >>> [  128.188930]  udpv6_sendmsg+0xb4f/0xdb0
> >>> [  128.188931]  ? ip_reply_glue_bits+0x40/0x40
> >>> [  128.188931]  ? update_load_avg+0x78/0x630
> >>> [  128.188932]  ? update_curr+0x73/0x1d0
> >>> [  128.188932]  ? __sys_sendto+0x108/0x190
> >>> [  128.188933]  __sys_sendto+0x108/0x190
> >>> [  128.188933]  ? __fput+0x1a5/0x240
> >>> [  128.188934]  ? _cond_resched+0x19/0x30
> >>> [  128.188934]  ? task_work_run+0x67/0x90
> >>> [  128.188935]  __x64_sys_sendto+0x25/0x30
> >>> [  128.188935]  do_syscall_64+0x48/0x130
> >>> [  128.188936]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> >>> [  128.188936] RIP: 0033:0x7f0b97a7826c
> >>> [  128.188945] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
> >>> [  128.188946] RSP: 002b:00007ffe80ffcea8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
> >>> [  128.188947] RAX: ffffffffffffffda RBX: 0000000000000058 RCX: 00007f0b97a7826c
> >>> [  128.188948] RDX: 0000000000000058 RSI: 000055fb765264c0 RDI: 0000000000000009
> >>> [  128.188949] RBP: 000055fb765264c0 R08: 0000000000000000 R09: 0000000000000000
> >>> [  128.188949] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffcfc0
> >>> [  128.188950] R13: 000000000002c600 R14: 0000000000000000 R15: 000055fb7629b751
> >>> [  128.188957] Sending NMI from CPU 10 to CPUs 19:
> >>> [  128.239939] NMI backtrace for cpu 19
> >>> [  128.239940] CPU: 19 PID: 2587 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
> >>> [  128.239940] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
> >>> [  128.239940] RIP: 0010:kvm_wait+0x30/0x50
> >>> [  128.239941] Code: 8b 05 ac cb f4 7e a9 00 00 f0 00 75 1f 9c 58 fa 0f b6 17 40 38 d6 75 12 f6 c4 02 75 10 e9 07 00 00 00 0f 00 2d 6f ce 54 01 f4 <50> 9d c3 e9 07 00 00 00 0f 00 2d 5f ce 54 01 fb f4 eb ed 66 66 2e
> >>> [  128.239942] RSP: 0018:ffffc90000cc78e8 EFLAGS: 00000046
> >>> [  128.239942] RAX: 0000000000000046 RBX: ffff88903fcec700 RCX: 0000000000000008
> >>> [  128.239943] RDX: 0000000000000003 RSI: 0000000000000003 RDI: ffff889039049d80
> >>> [  128.239943] RBP: ffff889039049d80 R08: ffff88907ffe9f80 R09: 00000000000000f8
> >>> [  128.239944] R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000000
> >>> [  128.239944] R13: 0000000000000001 R14: 0000000000000100 R15: 0000000000500000
> >>> [  128.239944] FS:  00007f0b97743740(0000) GS:ffff88903fcc0000(0000) knlGS:0000000000000000
> >>> [  128.239945] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >>> [  128.239945] CR2: 000055fb76a96030 CR3: 0000001035882004 CR4: 0000000000760ee0
> >>> [  128.239945] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> >>> [  128.239946] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> >>> [  128.239946] PKRU: 55555554
> >>> [  128.239946] Call Trace:
> >>> [  128.239946]  __pv_queued_spin_lock_slowpath+0x26e/0x2b0
> >>> [  128.239947]  _raw_spin_lock_irqsave+0x25/0x30
> >>> [  128.239947]  __wake_up_common_lock+0x62/0xc0
> >>> [  128.239947]  sock_def_write_space+0x49/0x90
> >>> [  128.239948]  sock_wfree+0x68/0xb0
> >>> [  128.239948]  loopback_xmit+0x49/0xe0
> >>> [  128.239948]  dev_hard_start_xmit+0x8d/0x1e0
> >>> [  128.239948]  __dev_queue_xmit+0x721/0x8e0
> >>> [  128.239949]  ip6_finish_output2+0x250/0x5b0
> >>> [  128.239949]  ip6_output+0x73/0x120
> >>> [  128.239949]  ? __ip6_finish_output+0x110/0x110
> >>> [  128.239950]  ip6_send_skb+0x1e/0x60
> >>> [  128.239950]  udp_v6_send_skb.isra.0+0x197/0x460
> >>> [  128.239950]  udpv6_sendmsg+0xb4f/0xdb0
> >>> [  128.239950]  ? release_pages+0x28f/0x2f0
> >>> [  128.239950]  ? ip_reply_glue_bits+0x40/0x40
> >>> [  128.239951]  ? _cond_resched+0x19/0x30
> >>> [  128.239951]  ? unmap_page_range+0x678/0xa60
> >>> [  128.239951]  ? __sys_sendto+0x108/0x190
> >>> [  128.239951]  __sys_sendto+0x108/0x190
> >>> [  128.239952]  ? __fput+0x1a5/0x240
> >>> [  128.239952]  ? _cond_resched+0x19/0x30
> >>> [  128.239952]  ? task_work_run+0x67/0x90
> >>> [  128.239952]  __x64_sys_sendto+0x25/0x30
> >>> [  128.239953]  do_syscall_64+0x48/0x130
> >>> [  128.239953]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> >>> [  128.239953] RIP: 0033:0x7f0b97a7826c
> >>> [  128.239954] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
> >>> [  128.239954] RSP: 002b:00007ffe80ffd7b8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
> >>> [  128.239955] RAX: ffffffffffffffda RBX: 00000000000003a8 RCX: 00007f0b97a7826c
> >>> [  128.239955] RDX: 00000000000003a8 RSI: 00007ffe80ffd800 RDI: 0000000000000009
> >>> [  128.239955] RBP: 00007ffe80ffd800 R08: 0000000000000000 R09: 0000000000000000
> >>> [  128.239956] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffd800
> >>> [  128.239956] R13: 00007ffe80ffd800 R14: 000000000000000e R15: 000055fb76b37e58
> >>
> >> I can reproduce this, and I see what it is. I'll send out a patch soonish.
> >
> > Thinko, can you try with this on top?
>
> Sorry that was incomplete, please use this one!
>
> diff --git a/fs/io_uring.c b/fs/io_uring.c
> index d22830a423f1..0075107dd09a 100644
> --- a/fs/io_uring.c
> +++ b/fs/io_uring.c
> @@ -2779,7 +2779,7 @@ static int io_read(struct io_kiocb *req, bool force_nonblock)
>                         ret = io_setup_async_rw(req, io_size, iovec,
>                                                 inline_vecs, &iter);
>                         if (ret)
> -                               goto out_free;
> +                               goto out;
>                         /* any defer here is final, must blocking retry */
>                         if (!(req->flags & REQ_F_NOWAIT) &&
>                             !file_can_poll(req->file))
> @@ -2787,18 +2787,18 @@ static int io_read(struct io_kiocb *req, bool force_nonblock)
>                         /* if we can retry, do so with the callbacks armed */
>                         if (io_rw_should_retry(req)) {
>                                 ret2 = io_iter_do_read(req, &iter);
> -                               if (ret2 != -EAGAIN) {
> +                               if (ret2 == -EIOCBQUEUED) {
> +                                       goto out;
> +                               } else if (ret2 != -EAGAIN) {
>                                         kiocb_done(kiocb, ret2);
> -                                       goto out_free;
> +                                       goto out;
>                                 }
>                         }
>                         kiocb->ki_flags &= ~IOCB_WAITQ;
>                         return -EAGAIN;
>                 }
>         }
> -out_free:
> -       kfree(iovec);
> -       req->flags &= ~REQ_F_NEED_CLEANUP;
> +out:
>         return ret;
>  }
>
> --
> Jens Axboe
>

Hi Jens,

I have tested with today's linux-block.git#async-buffered.7 and my
usual testcases:

[ FIO ]

root# echo 1 > /sys/kernel/debug/tracing/events/io_uring/io_uring_create/enable

root# cat /sys/kernel/debug/tracing/events/io_uring/io_uring_create/enable
1

root# cat buf-test-dileks-min
[buf-test-dileks-min]
filename=/mnt/backup/thinkpad-t530/home/sdi/Downloads/iso-images/iso_debian-stretch/firmware-stretch-live-amd64/debian-live-9.2.0-amd64-kde+nonfree.iso
buffered=1
ioengine=io_uring

root# fio ./buf-test-dileks-min
buf-test-dileks-min: (g=0): rw=read, bs=(R) 4096B-4096B, (W)
4096B-4096B, (T) 4096B-4096B, ioengine=io_uring, iodepth=1
fio-3.20
Starting 1 process
Jobs: 1 (f=1): [R(1)][100.0%][r=87.1MiB/s][r=22.3k IOPS][eta 00m:00s]
buf-test-dileks-min: (groupid=0, jobs=1): err= 0: pid=4076: Sat Jun  6
02:22:20 2020
  read: IOPS=20.8k, BW=81.2MiB/s (85.1MB/s)(2811MiB/34626msec)
    slat (usec): min=2, max=395, avg= 8.95, stdev=18.43
    clat (nsec): min=130, max=159571k, avg=36797.18, stdev=520010.15
     lat (usec): min=2, max=159779, avg=46.04, stdev=528.95
    clat percentiles (nsec):
     |  1.00th=[     225],  5.00th=[     302], 10.00th=[     330],
     | 20.00th=[     358], 30.00th=[     378], 40.00th=[     390],
     | 50.00th=[     406], 60.00th=[     418], 70.00th=[     430],
     | 80.00th=[     446], 90.00th=[     478], 95.00th=[     732],
     | 99.00th=[ 1531904], 99.50th=[ 1695744], 99.90th=[ 3555328],
     | 99.95th=[ 3686400], 99.99th=[16711680]
   bw (  KiB/s): min=34235, max=94554, per=100.00%, avg=83141.75,
stdev=11118.80, samples=69
   iops        : min= 8558, max=23638, avg=20785.14, stdev=2779.73, samples=69
  lat (nsec)   : 250=1.63%, 500=89.57%, 750=3.94%, 1000=0.85%
  lat (usec)   : 2=1.55%, 4=0.21%, 10=0.03%, 20=0.01%, 50=0.02%
  lat (usec)   : 100=0.01%, 250=0.50%, 500=0.11%, 750=0.02%, 1000=0.01%
  lat (msec)   : 2=1.18%, 4=0.34%, 10=0.01%, 20=0.01%, 50=0.01%
  lat (msec)   : 100=0.01%, 250=0.01%
  cpu          : usr=9.80%, sys=17.79%, ctx=15869, majf=0, minf=15
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=719568,0,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=1

Run status group 0 (all jobs):
   READ: bw=81.2MiB/s (85.1MB/s), 81.2MiB/s-81.2MiB/s
(85.1MB/s-85.1MB/s), io=2811MiB (2947MB), run=34626-34626msec

Disk stats (read/write):
  sdc: ios=15784/53, merge=0/89, ticks=57005/536, in_queue=58011, util=99.75%

root# fio --showcmd ./buf-test-dileks-min
fio --name=buf-test-dileks-min --buffered=1 --ioengine=io_uring
--filename=/mnt/backup/thinkpad-t530/home/sdi/Downloads/iso-images/iso_debian-stretch/firmware-stretch-live-amd64/debian-live-9.2.0-amd64-kde+nonfree.iso

root# tail /sys/kernel/debug/tracing/trace
# entries-in-buffer/entries-written: 1/1   #P:4
#
#                              _-----=> irqs-off
#                             / _----=> need-resched
#                            | / _---=> hardirq/softirq
#                            || / _--=> preempt-depth
#                            ||| /     delay
#           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
#              | |       |   ||||       |         |
             fio-4076  [001] ....   641.669846: io_uring_create: ring
00000000a88caba9, fd 5 sq size 1, cq size 2, flags 0


[ SYSTEMD-ANALYZE ]

user$ sudo systemd-analyze --no-pager time
Startup finished in 6.228s (kernel) + 41.559s (userspace) = 47.787s
graphical.target reached after 41.539s in userspace

user$ sudo systemd-analyze --no-pager blame
19.800s accounts-daemon.service
19.001s udisks2.service
13.282s NetworkManager-wait-online.service
 9.956s polkit.service
 9.590s avahi-daemon.service
 9.479s NetworkManager.service
 9.436s rtkit-daemon.service
 9.367s systemd-logind.service
 9.272s switcheroo-control.service
 9.257s wpa_supplicant.service
 9.215s dev-sdc2.device
 9.079s powertop.service
 8.875s smartmontools.service
 8.108s ModemManager.service
 7.070s zramswap.service
 5.443s networking.service
 5.181s rsyslog.service
 4.870s gpm.service
 3.846s e2scrub_reap.service
 3.718s systemd-fsck@dev-disk-by\x2duuid-cfd78a9b\x2dfadd\x2d4870\x2d88c2\x2d5277dae4e59f.service
 3.550s apparmor.service
 1.807s systemd-journal-flush.service
 1.570s systemd-tmpfiles-setup.service
 1.269s keyboard-setup.service
 1.259s atd.service
 1.207s systemd-udevd.service
 1.055s bluetooth.service
 1.052s binfmt-support.service
  935ms exim4.service
  901ms systemd-journald.service
  850ms systemd-modules-load.service
  843ms systemd-sysusers.service
  814ms upower.service
  807ms dev-hugepages.mount
  803ms pppd-dns.service
  790ms dev-mqueue.mount
  771ms sys-kernel-debug.mount
  770ms sys-kernel-tracing.mount
  697ms systemd-timesyncd.service
  677ms modprobe@drm.service
  630ms ifupdown-wait-online.service
  604ms systemd-rfkill.service
  561ms systemd-tmpfiles-setup-dev.service
  521ms systemd-remount-fs.service
  476ms systemd-sysctl.service
  397ms systemd-random-seed.service
  394ms systemd-backlight@backlight:intel_backlight.service
  376ms systemd-tmpfiles-clean.service
  289ms kmod-static-nodes.service
  269ms user@1000.service
  194ms console-setup.service
  182ms systemd-udev-trigger.service
  134ms systemd-update-utmp.service
  125ms proc-sys-fs-binfmt_misc.mount
   69ms systemd-user-sessions.service
   52ms alsa-restore.service
   21ms user-runtime-dir@1000.service
   17ms mnt-sandisk.mount
   13ms systemd-update-utmp-runlevel.service
    6ms sys-fs-fuse-connections.mount
    4ms ifupdown-pre.service

user$ sudo systemd-analyze --no-pager critical-chain
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.

graphical.target @41.539s
└─multi-user.target @41.538s
  └─exim4.service @40.601s +935ms
    └─network-online.target @40.599s
      └─NetworkManager-wait-online.service @27.314s +13.282s
        └─NetworkManager.service @17.831s +9.479s
          └─dbus.service @17.824s
            └─basic.target @17.675s
              └─sockets.target @17.672s
                └─dbus.socket @17.670s
                  └─sysinit.target @17.584s
                    └─systemd-udev-trigger.service @26.747s +182ms
                      └─systemd-udevd-kernel.socket @3.027s
                        └─system.slice @2.440s
                          └─-.slice @2.440s

Hope this helps you.

If you want some further informations and/or testing, please let me
know and give clear instructions.

Thanks.

Regards,
- Sedat -

Re: [PATCHSET v5 0/12] Add support for async buffered reads

[Jens Axboe]

On 6/5/20 6:33 PM, Sedat Dilek wrote:
> On Fri, Jun 5, 2020 at 11:24 PM Jens Axboe <axboe@kernel.dk> wrote:
>>
>> On 6/5/20 3:13 PM, Jens Axboe wrote:
>>> On 6/5/20 2:53 PM, Jens Axboe wrote:
>>>> On 6/5/20 2:36 PM, Andres Freund wrote:
>>>>> Hi,
>>>>>
>>>>> On 2020-06-05 13:20:28 -0700, Andres Freund wrote:
>>>>>> I'll go and try to figure out why I don't see an oops...
>>>>>
>>>>> Err, that probably was a typo on my end in the serial console
>>>>> config. After fixing that, I did get the below.
>>>>>
>>>>> If helpful I can try with debugging enabled or such.
>>>>>
>>>>> [   67.910265] tee (2577): drop_caches: 3
>>>>> [   68.059674] BUG: unable to handle page fault for address: 00007f0b16a3c03c
>>>>> [   68.062021] #PF: supervisor read access in kernel mode
>>>>> [   68.063742] #PF: error_code(0x0000) - not-present page
>>>>> [   68.065517] PGD 102e044067 P4D 102e044067 PUD 102bf7a067 PMD 0
>>>>> [   68.067519] Oops: 0000 [#1] SMP NOPTI
>>>>> [   68.068800] CPU: 2 PID: 2554 Comm: postgres Not tainted 5.7.0-andres-10123-g87823242260e #44
>>>>> [   68.071505] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>>>>> [   68.074139] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
>>>>> [   68.075389] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
>>>>> [   68.079125] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
>>>>> [   68.080260] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
>>>>> [   68.084115] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
>>>>> [   68.085374] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
>>>>> [   68.086409] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
>>>>> [   68.087447] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
>>>>> [   68.088697] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>>>>> [   68.089903] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>>> [   68.090776] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
>>>>> [   68.091834] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>>>> [   68.092902] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>>>> [   68.093967] PKRU: 55555554
>>>>> [   68.094427] Call Trace:
>>>>> [   68.094845]  ? __schedule+0x2ac/0x720
>>>>> [   68.095350]  ? schedule+0x55/0xc0
>>>>> [   68.095797]  ? ep_read_events_proc+0xd0/0xd0
>>>>> [   68.096354]  ep_scan_ready_list.constprop.0+0x16c/0x190
>>>>> [   68.097016]  ep_poll+0x2a3/0x440
>>>>> [   68.097449]  ? wait_woken+0x70/0x70
>>>>> [   68.097904]  do_epoll_wait+0xb0/0xd0
>>>>> [   68.098375]  __x64_sys_epoll_wait+0x1a/0x20
>>>>> [   68.098913]  do_syscall_64+0x48/0x130
>>>>> [   68.099393]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>>>> [   68.100030] RIP: 0033:0x7f0b97845606
>>>>> [   68.100498] Code: 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 11 b8 e8 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 5a c3 90 48 83 ec 28 89 54 24 18 48 89 74 24
>>>>> [   68.102718] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
>>>>> [   68.103644] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
>>>>> [   68.104533] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
>>>>> [   68.105418] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
>>>>> [   68.106296] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
>>>>> [   68.107187] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
>>>>> [   68.108104] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
>>>>> [   68.109505] CR2: 00007f0b16a3c03c
>>>>> [   68.109962] ---[ end trace 0ca39a5ed99162ce ]---
>>>>> [   68.110547] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
>>>>> [   68.111214] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
>>>>> [   68.113435] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
>>>>> [   68.114111] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
>>>>> [   68.115016] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
>>>>> [   68.115902] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
>>>>> [   68.116810] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
>>>>> [   68.117663] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
>>>>> [   68.118520] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>>>>> [   68.119482] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>>> [   68.120181] CR2: 00007f0b16a3c03c CR3: 000000102cf7c004 CR4: 0000000000760ee0
>>>>> [   68.121043] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>>>> [   68.121904] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>>>> [   68.122790] PKRU: 55555554
>>>>> [   68.123401] general protection fault, probably for non-canonical address 0xfeeda989fef06266: 0000 [#2] SMP NOPTI
>>>>> [   68.125052] CPU: 2 PID: 2554 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
>>>>> [   68.126260] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>>>>> [   68.127274] RIP: 0010:__pv_queued_spin_lock_slowpath+0x1a1/0x2b0
>>>>> [   68.128032] Code: c4 c1 ea 12 41 be 01 00 00 00 8d 42 ff 41 83 e4 03 4c 8d 6b 14 49 c1 e4 05 48 98 49 81 c4 00 c7 02 00 4c 03 24 c5 e0 e6 fd 82 <49> 89 1c 24 b8 00 80 00 00 eb 15 84 c0 75 0a 41 0f b6 54 24 14 84
>>>>> [   68.130221] RSP: 0018:ffffc90000befce8 EFLAGS: 00010086
>>>>> [   68.130867] RAX: 0000000000003ffe RBX: ffff88903f8ac700 RCX: 0000000000000001
>>>>> [   68.131752] RDX: 0000000000003fff RSI: 0000000000000000 RDI: 0000000000000000
>>>>> [   68.132637] RBP: ffff889037617924 R08: 0000000000000000 R09: ffffc90000befdf8
>>>>> [   68.133513] R10: ffff8890334d2bf0 R11: 0000000000000018 R12: feeda989fef06266
>>>>> [   68.134399] R13: ffff88903f8ac714 R14: 0000000000000001 R15: 00000000000c0000
>>>>> [   68.135323] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>>>>> [   68.136307] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>>> [   68.137004] CR2: 00007f0b16a3c03c CR3: 000000000360a005 CR4: 0000000000760ee0
>>>>> [   68.137866] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>>>> [   68.138748] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>>>> [   68.139606] PKRU: 55555554
>>>>> [   68.139973] Call Trace:
>>>>> [   68.140313]  queued_read_lock_slowpath+0x6c/0x70
>>>>> [   68.140915]  _raw_read_lock_irqsave+0x26/0x30
>>>>> [   68.141485]  ep_poll_callback+0x3e/0x2b0
>>>>> [   68.142007]  ? set_next_entity+0xab/0x1f0
>>>>> [   68.142541]  __wake_up_common+0x7a/0x140
>>>>> [   68.143077]  __wake_up_common_lock+0x7c/0xc0
>>>>> [   68.143651]  pipe_release+0x5b/0xd0
>>>>> [   68.144150]  __fput+0xda/0x240
>>>>> [   68.144574]  task_work_run+0x62/0x90
>>>>> [   68.145046]  do_exit+0x35c/0xa70
>>>>> [   68.145505]  ? do_epoll_wait+0xb0/0xd0
>>>>> [   68.146000]  rewind_stack_do_exit+0x17/0x20
>>>>> [   68.146538] RIP: 0033:0x7f0b97845606
>>>>> [   68.146988] Code: Bad RIP value.
>>>>> [   68.147405] RSP: 002b:00007ffe80ffdba8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e8
>>>>> [   68.148314] RAX: ffffffffffffffda RBX: 000055fb76a9a998 RCX: 00007f0b97845606
>>>>> [   68.149157] RDX: 0000000000000001 RSI: 000055fb76a9aa10 RDI: 0000000000000005
>>>>> [   68.150021] RBP: 0000000005000007 R08: 0000000005000007 R09: 0000000000000003
>>>>> [   68.150831] R10: 00000000ffffffff R11: 0000000000000246 R12: 000055fb76a9a998
>>>>> [   68.151640] R13: 0000000000000001 R14: 0000000000000009 R15: 000055fb76a9a998
>>>>> [   68.152459] Modules linked in: 9pnet_virtio isst_if_common xhci_pci 9pnet iTCO_wdt intel_pmc_bxt xhci_hcd iTCO_vendor_support
>>>>> [   68.153707] ---[ end trace 0ca39a5ed99162cf ]---
>>>>> [   68.154282] RIP: 0010:__mutex_lock.isra.0+0x76/0x4e0
>>>>> [   68.154884] Code: b7 01 00 00 48 39 c6 0f 84 a4 03 00 00 65 48 8b 04 25 80 7d 01 00 48 8b 00 a8 08 75 18 49 8b 06 48 83 e0 f8 0f 84 2f 02 00 00 <8b> 50 3c 85 d2 0f 85 12 02 00 00 65 48 8b 04 25 80 7d 01 00 48 8b
>>>>> [   68.156976] RSP: 0018:ffffc90000befd50 EFLAGS: 00010206
>>>>> [   68.157614] RAX: 00007f0b16a3c000 RBX: ffff889037617910 RCX: 0000000000000000
>>>>> [   68.158436] RDX: 00007f0b16a3c000 RSI: ffff88902a4aca80 RDI: ffff8890376178c0
>>>>> [   68.159269] RBP: ffffc90000befdf0 R08: 0000000000000001 R09: 0000000000000000
>>>>> [   68.160092] R10: ffff889037617920 R11: ffff88903f8abaf0 R12: ffffc90000befe08
>>>>> [   68.160920] R13: 0000000000000000 R14: ffff8890376178c0 R15: 0000000000000000
>>>>> [   68.161746] FS:  00007f0b97743740(0000) GS:ffff88903f880000(0000) knlGS:0000000000000000
>>>>> [   68.162701] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>>> [   68.163386] CR2: 00007f0b978455dc CR3: 000000000360a005 CR4: 0000000000760ee0
>>>>> [   68.164226] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>>>> [   68.165079] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>>>> [   68.165931] PKRU: 55555554
>>>>> [   68.166298] Fixing recursive fault but reboot is needed!
>>>>> [  128.173729] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
>>>>> [  128.179267] rcu:         11-...0: (1 GPs behind) idle=c56/1/0x4000000000000000 softirq=6897/6898 fqs=5976
>>>>> [  128.182232] rcu:         19-...0: (1 GPs behind) idle=492/1/0x4000000000000000 softirq=1023/1023 fqs=5976
>>>>> [  128.185217]      (detected by 10, t=18003 jiffies, g=15789, q=631)
>>>>> [  128.186863] Sending NMI from CPU 10 to CPUs 11:
>>>>> [  128.188902] NMI backtrace for cpu 11
>>>>> [  128.188903] CPU: 11 PID: 2546 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
>>>>> [  128.188904] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>>>>> [  128.188904] RIP: 0010:queued_read_lock_slowpath+0x5b/0x70
>>>>> [  128.188906] Code: 03 00 02 00 00 8b 03 84 c0 74 08 f3 90 8b 13 84 d2 75 f8 48 89 ef e8 74 e6 ff ff 66 90 5b 5d c3 8b 07 84 c0 74 08 f3 90 8b 03 <84> c0 75 f8 5b 5d c3 89 c6 48 89 ef e8 e4 e8 ff ff 66 90 eb bf 0f
>>>>> [  128.188907] RSP: 0018:ffffc90000348c08 EFLAGS: 00000086
>>>>> [  128.188908] RAX: 0000000037617cc0 RBX: ffff889037617920 RCX: 00000000000000c3
>>>>> [  128.188909] RDX: 0000000000000001 RSI: 0000000000000001 RDI: ffff889037617920
>>>>> [  128.188910] RBP: ffff889037bb1a80 R08: 00000000000000c3 R09: ffffc90000348cd8
>>>>> [  128.188910] R10: 0100000000000000 R11: 00000000de2ee17e R12: 0000000000000046
>>>>> [  128.188911] R13: ffff889037617920 R14: 0000000000000001 R15: 00000000000000c3
>>>>> [  128.188912] FS:  00007f0b97743740(0000) GS:ffff88903fac0000(0000) knlGS:0000000000000000
>>>>> [  128.188912] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>>> [  128.188913] CR2: 00007f0b16afc00e CR3: 000000102e5c6002 CR4: 0000000000760ee0
>>>>> [  128.188914] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>>>> [  128.188914] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>>>> [  128.188915] PKRU: 55555554
>>>>> [  128.188915] Call Trace:
>>>>> [  128.188916]  <IRQ>
>>>>> [  128.188916]  _raw_read_lock_irqsave+0x26/0x30
>>>>> [  128.188917]  ep_poll_callback+0x3e/0x2b0
>>>>> [  128.188917]  ? cpumask_next_and+0x19/0x20
>>>>> [  128.188918]  ? update_sd_lb_stats.constprop.0+0xfe/0x810
>>>>> [  128.188918]  __wake_up_common+0x7a/0x140
>>>>> [  128.188919]  __wake_up_common_lock+0x7c/0xc0
>>>>> [  128.188919]  sock_def_readable+0x37/0x60
>>>>> [  128.188920]  __udp_enqueue_schedule_skb+0x168/0x260
>>>>> [  128.188920]  udpv6_queue_rcv_one_skb+0x284/0x3c0
>>>>> [  128.188921]  udp6_unicast_rcv_skb.isra.0+0x44/0xa0
>>>>> [  128.188921]  ip6_protocol_deliver_rcu+0x235/0x4b0
>>>>> [  128.188922]  ip6_input_finish+0x11/0x20
>>>>> [  128.188922]  ip6_input+0xa2/0xb0
>>>>> [  128.188923]  ? ip6_protocol_deliver_rcu+0x4b0/0x4b0
>>>>> [  128.188923]  ipv6_rcv+0xc0/0xd0
>>>>> [  128.188924]  ? ip6_rcv_finish_core.isra.0+0xd0/0xd0
>>>>> [  128.188924]  __netif_receive_skb_one_core+0x63/0xa0
>>>>> [  128.188925]  process_backlog+0x98/0x140
>>>>> [  128.188925]  net_rx_action+0x13a/0x370
>>>>> [  128.188926]  __do_softirq+0xe0/0x2ca
>>>>> [  128.188926]  do_softirq_own_stack+0x2a/0x40
>>>>> [  128.188926]  </IRQ>
>>>>> [  128.188927]  do_softirq.part.0+0x2b/0x30
>>>>> [  128.188927]  __local_bh_enable_ip+0x4b/0x50
>>>>> [  128.188928]  ip6_finish_output2+0x264/0x5b0
>>>>> [  128.188928]  ip6_output+0x73/0x120
>>>>> [  128.188929]  ? __ip6_finish_output+0x110/0x110
>>>>> [  128.188929]  ip6_send_skb+0x1e/0x60
>>>>> [  128.188930]  udp_v6_send_skb.isra.0+0x197/0x460
>>>>> [  128.188930]  udpv6_sendmsg+0xb4f/0xdb0
>>>>> [  128.188931]  ? ip_reply_glue_bits+0x40/0x40
>>>>> [  128.188931]  ? update_load_avg+0x78/0x630
>>>>> [  128.188932]  ? update_curr+0x73/0x1d0
>>>>> [  128.188932]  ? __sys_sendto+0x108/0x190
>>>>> [  128.188933]  __sys_sendto+0x108/0x190
>>>>> [  128.188933]  ? __fput+0x1a5/0x240
>>>>> [  128.188934]  ? _cond_resched+0x19/0x30
>>>>> [  128.188934]  ? task_work_run+0x67/0x90
>>>>> [  128.188935]  __x64_sys_sendto+0x25/0x30
>>>>> [  128.188935]  do_syscall_64+0x48/0x130
>>>>> [  128.188936]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>>>> [  128.188936] RIP: 0033:0x7f0b97a7826c
>>>>> [  128.188945] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
>>>>> [  128.188946] RSP: 002b:00007ffe80ffcea8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
>>>>> [  128.188947] RAX: ffffffffffffffda RBX: 0000000000000058 RCX: 00007f0b97a7826c
>>>>> [  128.188948] RDX: 0000000000000058 RSI: 000055fb765264c0 RDI: 0000000000000009
>>>>> [  128.188949] RBP: 000055fb765264c0 R08: 0000000000000000 R09: 0000000000000000
>>>>> [  128.188949] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffcfc0
>>>>> [  128.188950] R13: 000000000002c600 R14: 0000000000000000 R15: 000055fb7629b751
>>>>> [  128.188957] Sending NMI from CPU 10 to CPUs 19:
>>>>> [  128.239939] NMI backtrace for cpu 19
>>>>> [  128.239940] CPU: 19 PID: 2587 Comm: postgres Tainted: G      D           5.7.0-andres-10123-g87823242260e #44
>>>>> [  128.239940] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1 04/01/2014
>>>>> [  128.239940] RIP: 0010:kvm_wait+0x30/0x50
>>>>> [  128.239941] Code: 8b 05 ac cb f4 7e a9 00 00 f0 00 75 1f 9c 58 fa 0f b6 17 40 38 d6 75 12 f6 c4 02 75 10 e9 07 00 00 00 0f 00 2d 6f ce 54 01 f4 <50> 9d c3 e9 07 00 00 00 0f 00 2d 5f ce 54 01 fb f4 eb ed 66 66 2e
>>>>> [  128.239942] RSP: 0018:ffffc90000cc78e8 EFLAGS: 00000046
>>>>> [  128.239942] RAX: 0000000000000046 RBX: ffff88903fcec700 RCX: 0000000000000008
>>>>> [  128.239943] RDX: 0000000000000003 RSI: 0000000000000003 RDI: ffff889039049d80
>>>>> [  128.239943] RBP: ffff889039049d80 R08: ffff88907ffe9f80 R09: 00000000000000f8
>>>>> [  128.239944] R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000000
>>>>> [  128.239944] R13: 0000000000000001 R14: 0000000000000100 R15: 0000000000500000
>>>>> [  128.239944] FS:  00007f0b97743740(0000) GS:ffff88903fcc0000(0000) knlGS:0000000000000000
>>>>> [  128.239945] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>>> [  128.239945] CR2: 000055fb76a96030 CR3: 0000001035882004 CR4: 0000000000760ee0
>>>>> [  128.239945] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>>>> [  128.239946] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>>>> [  128.239946] PKRU: 55555554
>>>>> [  128.239946] Call Trace:
>>>>> [  128.239946]  __pv_queued_spin_lock_slowpath+0x26e/0x2b0
>>>>> [  128.239947]  _raw_spin_lock_irqsave+0x25/0x30
>>>>> [  128.239947]  __wake_up_common_lock+0x62/0xc0
>>>>> [  128.239947]  sock_def_write_space+0x49/0x90
>>>>> [  128.239948]  sock_wfree+0x68/0xb0
>>>>> [  128.239948]  loopback_xmit+0x49/0xe0
>>>>> [  128.239948]  dev_hard_start_xmit+0x8d/0x1e0
>>>>> [  128.239948]  __dev_queue_xmit+0x721/0x8e0
>>>>> [  128.239949]  ip6_finish_output2+0x250/0x5b0
>>>>> [  128.239949]  ip6_output+0x73/0x120
>>>>> [  128.239949]  ? __ip6_finish_output+0x110/0x110
>>>>> [  128.239950]  ip6_send_skb+0x1e/0x60
>>>>> [  128.239950]  udp_v6_send_skb.isra.0+0x197/0x460
>>>>> [  128.239950]  udpv6_sendmsg+0xb4f/0xdb0
>>>>> [  128.239950]  ? release_pages+0x28f/0x2f0
>>>>> [  128.239950]  ? ip_reply_glue_bits+0x40/0x40
>>>>> [  128.239951]  ? _cond_resched+0x19/0x30
>>>>> [  128.239951]  ? unmap_page_range+0x678/0xa60
>>>>> [  128.239951]  ? __sys_sendto+0x108/0x190
>>>>> [  128.239951]  __sys_sendto+0x108/0x190
>>>>> [  128.239952]  ? __fput+0x1a5/0x240
>>>>> [  128.239952]  ? _cond_resched+0x19/0x30
>>>>> [  128.239952]  ? task_work_run+0x67/0x90
>>>>> [  128.239952]  __x64_sys_sendto+0x25/0x30
>>>>> [  128.239953]  do_syscall_64+0x48/0x130
>>>>> [  128.239953]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>>>> [  128.239953] RIP: 0033:0x7f0b97a7826c
>>>>> [  128.239954] Code: c0 ff ff ff ff eb b9 0f 1f 80 00 00 00 00 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 19 45 31 c9 45 31 c0 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 64 c3 0f 1f 00 55 48 83 ec 20 48 89 54 24 10
>>>>> [  128.239954] RSP: 002b:00007ffe80ffd7b8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
>>>>> [  128.239955] RAX: ffffffffffffffda RBX: 00000000000003a8 RCX: 00007f0b97a7826c
>>>>> [  128.239955] RDX: 00000000000003a8 RSI: 00007ffe80ffd800 RDI: 0000000000000009
>>>>> [  128.239955] RBP: 00007ffe80ffd800 R08: 0000000000000000 R09: 0000000000000000
>>>>> [  128.239956] R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe80ffd800
>>>>> [  128.239956] R13: 00007ffe80ffd800 R14: 000000000000000e R15: 000055fb76b37e58
>>>>
>>>> I can reproduce this, and I see what it is. I'll send out a patch soonish.
>>>
>>> Thinko, can you try with this on top?
>>
>> Sorry that was incomplete, please use this one!
>>
>> diff --git a/fs/io_uring.c b/fs/io_uring.c
>> index d22830a423f1..0075107dd09a 100644
>> --- a/fs/io_uring.c
>> +++ b/fs/io_uring.c
>> @@ -2779,7 +2779,7 @@ static int io_read(struct io_kiocb *req, bool force_nonblock)
>>                         ret = io_setup_async_rw(req, io_size, iovec,
>>                                                 inline_vecs, &iter);
>>                         if (ret)
>> -                               goto out_free;
>> +                               goto out;
>>                         /* any defer here is final, must blocking retry */
>>                         if (!(req->flags & REQ_F_NOWAIT) &&
>>                             !file_can_poll(req->file))
>> @@ -2787,18 +2787,18 @@ static int io_read(struct io_kiocb *req, bool force_nonblock)
>>                         /* if we can retry, do so with the callbacks armed */
>>                         if (io_rw_should_retry(req)) {
>>                                 ret2 = io_iter_do_read(req, &iter);
>> -                               if (ret2 != -EAGAIN) {
>> +                               if (ret2 == -EIOCBQUEUED) {
>> +                                       goto out;
>> +                               } else if (ret2 != -EAGAIN) {
>>                                         kiocb_done(kiocb, ret2);
>> -                                       goto out_free;
>> +                                       goto out;
>>                                 }
>>                         }
>>                         kiocb->ki_flags &= ~IOCB_WAITQ;
>>                         return -EAGAIN;
>>                 }
>>         }
>> -out_free:
>> -       kfree(iovec);
>> -       req->flags &= ~REQ_F_NEED_CLEANUP;
>> +out:
>>         return ret;
>>  }
>>
>> --
>> Jens Axboe
>>
> 
> Hi Jens,
> 
> I have tested with today's linux-block.git#async-buffered.7 and my
> usual testcases:
> 
> [ FIO ]
> 
> root# echo 1 > /sys/kernel/debug/tracing/events/io_uring/io_uring_create/enable
> 
> root# cat /sys/kernel/debug/tracing/events/io_uring/io_uring_create/enable
> 1
> 
> root# cat buf-test-dileks-min
> [buf-test-dileks-min]
> filename=/mnt/backup/thinkpad-t530/home/sdi/Downloads/iso-images/iso_debian-stretch/firmware-stretch-live-amd64/debian-live-9.2.0-amd64-kde+nonfree.iso
> buffered=1
> ioengine=io_uring
> 
> root# fio ./buf-test-dileks-min
> buf-test-dileks-min: (g=0): rw=read, bs=(R) 4096B-4096B, (W)
> 4096B-4096B, (T) 4096B-4096B, ioengine=io_uring, iodepth=1
> fio-3.20
> Starting 1 process
> Jobs: 1 (f=1): [R(1)][100.0%][r=87.1MiB/s][r=22.3k IOPS][eta 00m:00s]
> buf-test-dileks-min: (groupid=0, jobs=1): err= 0: pid=4076: Sat Jun  6
> 02:22:20 2020
>   read: IOPS=20.8k, BW=81.2MiB/s (85.1MB/s)(2811MiB/34626msec)
>     slat (usec): min=2, max=395, avg= 8.95, stdev=18.43
>     clat (nsec): min=130, max=159571k, avg=36797.18, stdev=520010.15
>      lat (usec): min=2, max=159779, avg=46.04, stdev=528.95
>     clat percentiles (nsec):
>      |  1.00th=[     225],  5.00th=[     302], 10.00th=[     330],
>      | 20.00th=[     358], 30.00th=[     378], 40.00th=[     390],
>      | 50.00th=[     406], 60.00th=[     418], 70.00th=[     430],
>      | 80.00th=[     446], 90.00th=[     478], 95.00th=[     732],
>      | 99.00th=[ 1531904], 99.50th=[ 1695744], 99.90th=[ 3555328],
>      | 99.95th=[ 3686400], 99.99th=[16711680]
>    bw (  KiB/s): min=34235, max=94554, per=100.00%, avg=83141.75,
> stdev=11118.80, samples=69
>    iops        : min= 8558, max=23638, avg=20785.14, stdev=2779.73, samples=69
>   lat (nsec)   : 250=1.63%, 500=89.57%, 750=3.94%, 1000=0.85%
>   lat (usec)   : 2=1.55%, 4=0.21%, 10=0.03%, 20=0.01%, 50=0.02%
>   lat (usec)   : 100=0.01%, 250=0.50%, 500=0.11%, 750=0.02%, 1000=0.01%
>   lat (msec)   : 2=1.18%, 4=0.34%, 10=0.01%, 20=0.01%, 50=0.01%
>   lat (msec)   : 100=0.01%, 250=0.01%
>   cpu          : usr=9.80%, sys=17.79%, ctx=15869, majf=0, minf=15
>   IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
>      submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
>      complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
>      issued rwts: total=719568,0,0,0 short=0,0,0,0 dropped=0,0,0,0
>      latency   : target=0, window=0, percentile=100.00%, depth=1
> 
> Run status group 0 (all jobs):
>    READ: bw=81.2MiB/s (85.1MB/s), 81.2MiB/s-81.2MiB/s
> (85.1MB/s-85.1MB/s), io=2811MiB (2947MB), run=34626-34626msec
> 
> Disk stats (read/write):
>   sdc: ios=15784/53, merge=0/89, ticks=57005/536, in_queue=58011, util=99.75%
> 
> root# fio --showcmd ./buf-test-dileks-min
> fio --name=buf-test-dileks-min --buffered=1 --ioengine=io_uring
> --filename=/mnt/backup/thinkpad-t530/home/sdi/Downloads/iso-images/iso_debian-stretch/firmware-stretch-live-amd64/debian-live-9.2.0-amd64-kde+nonfree.iso
> 
> root# tail /sys/kernel/debug/tracing/trace
> # entries-in-buffer/entries-written: 1/1   #P:4
> #
> #                              _-----=> irqs-off
> #                             / _----=> need-resched
> #                            | / _---=> hardirq/softirq
> #                            || / _--=> preempt-depth
> #                            ||| /     delay
> #           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
> #              | |       |   ||||       |         |
>              fio-4076  [001] ....   641.669846: io_uring_create: ring
> 00000000a88caba9, fd 5 sq size 1, cq size 2, flags 0
> 
> 
> [ SYSTEMD-ANALYZE ]
> 
> user$ sudo systemd-analyze --no-pager time
> Startup finished in 6.228s (kernel) + 41.559s (userspace) = 47.787s
> graphical.target reached after 41.539s in userspace
> 
> user$ sudo systemd-analyze --no-pager blame
> 19.800s accounts-daemon.service
> 19.001s udisks2.service
> 13.282s NetworkManager-wait-online.service
>  9.956s polkit.service
>  9.590s avahi-daemon.service
>  9.479s NetworkManager.service
>  9.436s rtkit-daemon.service
>  9.367s systemd-logind.service
>  9.272s switcheroo-control.service
>  9.257s wpa_supplicant.service
>  9.215s dev-sdc2.device
>  9.079s powertop.service
>  8.875s smartmontools.service
>  8.108s ModemManager.service
>  7.070s zramswap.service
>  5.443s networking.service
>  5.181s rsyslog.service
>  4.870s gpm.service
>  3.846s e2scrub_reap.service
>  3.718s systemd-fsck@dev-disk-by\x2duuid-cfd78a9b\x2dfadd\x2d4870\x2d88c2\x2d5277dae4e59f.service
>  3.550s apparmor.service
>  1.807s systemd-journal-flush.service
>  1.570s systemd-tmpfiles-setup.service
>  1.269s keyboard-setup.service
>  1.259s atd.service
>  1.207s systemd-udevd.service
>  1.055s bluetooth.service
>  1.052s binfmt-support.service
>   935ms exim4.service
>   901ms systemd-journald.service
>   850ms systemd-modules-load.service
>   843ms systemd-sysusers.service
>   814ms upower.service
>   807ms dev-hugepages.mount
>   803ms pppd-dns.service
>   790ms dev-mqueue.mount
>   771ms sys-kernel-debug.mount
>   770ms sys-kernel-tracing.mount
>   697ms systemd-timesyncd.service
>   677ms modprobe@drm.service
>   630ms ifupdown-wait-online.service
>   604ms systemd-rfkill.service
>   561ms systemd-tmpfiles-setup-dev.service
>   521ms systemd-remount-fs.service
>   476ms systemd-sysctl.service
>   397ms systemd-random-seed.service
>   394ms systemd-backlight@backlight:intel_backlight.service
>   376ms systemd-tmpfiles-clean.service
>   289ms kmod-static-nodes.service
>   269ms user@1000.service
>   194ms console-setup.service
>   182ms systemd-udev-trigger.service
>   134ms systemd-update-utmp.service
>   125ms proc-sys-fs-binfmt_misc.mount
>    69ms systemd-user-sessions.service
>    52ms alsa-restore.service
>    21ms user-runtime-dir@1000.service
>    17ms mnt-sandisk.mount
>    13ms systemd-update-utmp-runlevel.service
>     6ms sys-fs-fuse-connections.mount
>     4ms ifupdown-pre.service
> 
> user$ sudo systemd-analyze --no-pager critical-chain
> The time when unit became active or started is printed after the "@" character.
> The time the unit took to start is printed after the "+" character.
> 
> graphical.target @41.539s
> └─multi-user.target @41.538s
>   └─exim4.service @40.601s +935ms
>     └─network-online.target @40.599s
>       └─NetworkManager-wait-online.service @27.314s +13.282s
>         └─NetworkManager.service @17.831s +9.479s
>           └─dbus.service @17.824s
>             └─basic.target @17.675s
>               └─sockets.target @17.672s
>                 └─dbus.socket @17.670s
>                   └─sysinit.target @17.584s
>                     └─systemd-udev-trigger.service @26.747s +182ms
>                       └─systemd-udevd-kernel.socket @3.027s
>                         └─system.slice @2.440s
>                           └─-.slice @2.440s
> 
> Hope this helps you.
> 
> If you want some further informations and/or testing, please let me
> know and give clear instructions.

Thanks for testing, appreciate it!

-- 
Jens Axboe

[PATCHSET v4 0/12] Add support for async buffered reads

[Jens Axboe]

We technically support this already through io_uring, but it's
implemented with a thread backend to support cases where we would
block. This isn't ideal.

After a few prep patches, the core of this patchset is adding support
for async callbacks on page unlock. With this primitive, we can simply
retry the IO operation. With io_uring, this works a lot like poll based
retry for files that support it. If a page is currently locked and
needed, -EIOCBQUEUED is returned with a callback armed. The callers
callback is responsible for restarting the operation.

With this callback primitive, we can add support for
generic_file_buffered_read(), which is what most file systems end up
using for buffered reads. XFS/ext4/btrfs/bdev is wired up, but probably
trivial to add more.

The file flags support for this by setting FMODE_BUF_RASYNC, similar
to what we do for FMODE_NOWAIT. Open to suggestions here if this is
the preferred method or not.

In terms of results, I wrote a small test app that randomly reads 4G
of data in 4K chunks from a file hosted by ext4. The app uses a queue
depth of 32. If you want to test yourself, you can just use buffered=1
with ioengine=io_uring with fio. No application changes are needed to
use the more optimized buffered async read.

preadv for comparison:
	real    1m13.821s
	user    0m0.558s
	sys     0m11.125s
	CPU	~13%

Mainline:
	real    0m12.054s
	user    0m0.111s
	sys     0m5.659s
	CPU	~32% + ~50% == ~82%

This patchset:
	real    0m9.283s
	user    0m0.147s
	sys     0m4.619s
	CPU	~52%

The CPU numbers are just a rough estimate. For the mainline io_uring
run, this includes the app itself and all the threads doing IO on its
behalf (32% for the app, ~1.6% per worker and 32 of them). Context
switch rate is much smaller with the patchset, since we only have the
one task performing IO.

Also ran a simple fio based test case, varying the queue depth from 1
to 16, doubling every time:

[buf-test]
filename=/data/file
direct=0
ioengine=io_uring
norandommap
rw=randread
bs=4k
iodepth=${QD}
randseed=89
runtime=10s

QD/Test		Patchset IOPS		Mainline IOPS
1		9046			8294
2		19.8k			18.9k
4		39.2k			28.5k
8		64.4k			31.4k
16		65.7k			37.8k

Outside of my usual environment, so this is just running on a virtualized
NVMe device in qemu, using ext4 as the file system. NVMe isn't very
efficient virtualized, so we run out of steam at ~65K which is why we
flatline on the patched side (nvme_submit_cmd() eats ~75% of the test app
CPU). Before that happens, it's a linear increase. Not shown is context
switch rate, which is massively lower with the new code. The old thread
offload adds a blocking thread per pending IO, so context rate quickly
goes through the roof.

The goal here is efficiency. Async thread offload adds latency, and
it also adds noticable overhead on items such as adding pages to the
page cache. By allowing proper async buffered read support, we don't
have X threads hammering on the same inode page cache, we have just
the single app actually doing IO.

Been beating on this and it's solid for me, and I'm now pretty happy
with how it all turned out. Not aware of any missing bits/pieces or
code cleanups that need doing.

Series can also be found here:

https://git.kernel.dk/cgit/linux-block/log/?h=async-buffered.4

or pull from:

git://git.kernel.dk/linux-block async-buffered.4

 fs/block_dev.c            |   2 +-
 fs/btrfs/file.c           |   2 +-
 fs/ext4/file.c            |   2 +-
 fs/io_uring.c             | 114 ++++++++++++++++++++++++++++++++++++++
 fs/xfs/xfs_file.c         |   2 +-
 include/linux/blk_types.h |   3 +-
 include/linux/fs.h        |  10 +++-
 include/linux/pagemap.h   |  67 ++++++++++++++++++++++
 mm/filemap.c              | 111 ++++++++++++++++++++++++-------------
 9 files changed, 267 insertions(+), 46 deletions(-)

Changes since v3:
- io_uring: don't retry if REQ_F_NOWAIT is set
- io_uring: alloc req->io if the request type didn't already
- Add iocb->ki_waitq instead of (ab)using iocb->private
Changes since v2:
- Get rid of unnecessary wait_page_async struct, just use wait_page_async
- Add another prep handler, adding wake_page_match()
- Use wake_page_match() in both callers
Changes since v1:
- Fix an issue with inline page locking
- Fix a potential race with __wait_on_page_locked_async()
- Fix a hang related to not setting page_match, thus missing a wakeup

-- 
Jens Axboe

[PATCH 02/12] mm: allow read-ahead with IOCB_NOWAIT set

[Jens Axboe]

The read-ahead shouldn't block, so allow it to be done even if
IOCB_NOWAIT is set in the kiocb.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 mm/filemap.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/mm/filemap.c b/mm/filemap.c
index 23a051a7ef0f..80747f1377d5 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -2031,8 +2031,6 @@ static ssize_t generic_file_buffered_read(struct kiocb *iocb,
 
 		page = find_get_page(mapping, index);
 		if (!page) {
-			if (iocb->ki_flags & IOCB_NOWAIT)
-				goto would_block;
 			page_cache_sync_readahead(mapping,
 					ra, filp,
 					index, last_index - index);
-- 
2.26.2

[PATCHSET v2 0/12] Add support for async buffered reads

[Jens Axboe]

We technically support this already through io_uring, but it's
implemented with a thread backend to support cases where we would
block. This isn't ideal.

After a few prep patches, the core of this patchset is adding support
for async callbacks on page unlock. With this primitive, we can simply
retry the IO operation. With io_uring, this works a lot like poll based
retry for files that support it. If a page is currently locked and
needed, -EIOCBQUEUED is returned with a callback armed. The callers
callback is responsible for restarting the operation.

With this callback primitive, we can add support for
generic_file_buffered_read(), which is what most file systems end up
using for buffered reads. XFS/ext4/btrfs/bdev is wired up, but probably
trivial to add more.

The file flags support for this by setting FMODE_BUF_RASYNC, similar
to what we do for FMODE_NOWAIT. Open to suggestions here if this is
the preferred method or not.

In terms of results, I wrote a small test app that randomly reads 4G
of data in 4K chunks from a file hosted by ext4. The app uses a queue
depth of 32. If you want to test yourself, you can just use buffered=1
with ioengine=io_uring with fio. No application changes are needed to
use the more optimized buffered async read.

preadv for comparison:
	real    1m13.821s
	user    0m0.558s
	sys     0m11.125s
	CPU	~13%

Mainline:
	real    0m12.054s
	user    0m0.111s
	sys     0m5.659s
	CPU	~32% + ~50% == ~82%

This patchset:
	real    0m9.283s
	user    0m0.147s
	sys     0m4.619s
	CPU	~52%

The CPU numbers are just a rough estimate. For the mainline io_uring
run, this includes the app itself and all the threads doing IO on its
behalf (32% for the app, ~1.6% per worker and 32 of them). Context
switch rate is much smaller with the patchset, since we only have the
one task performing IO.

The goal here is efficiency. Async thread offload adds latency, and
it also adds noticable overhead on items such as adding pages to the
page cache. By allowing proper async buffered read support, we don't
have X threads hammering on the same inode page cache, we have just
the single app actually doing IO.

Been beating on this and it's solid for me, and I'm now pretty happy
with how it all turned out. Not aware of any missing bits/pieces or
code cleanups that need doing.

Series can also be found here:

https://git.kernel.dk/cgit/linux-block/log/?h=async-buffered.3

or pull from:

git://git.kernel.dk/linux-block async-buffered.3

 fs/block_dev.c            |   2 +-
 fs/btrfs/file.c           |   2 +-
 fs/ext4/file.c            |   2 +-
 fs/io_uring.c             |  99 ++++++++++++++++++++++++++++++++++
 fs/xfs/xfs_file.c         |   2 +-
 include/linux/blk_types.h |   3 +-
 include/linux/fs.h        |   5 ++
 include/linux/pagemap.h   |  64 ++++++++++++++++++++++
 mm/filemap.c              | 111 ++++++++++++++++++++++++--------------
 9 files changed, 245 insertions(+), 45 deletions(-)

Changes since v2:
- Get rid of unnecessary wait_page_async struct, just use wait_page_async
- Add another prep handler, adding wake_page_match()
- Use wake_page_match() in both callers
Changes since v1:
- Fix an issue with inline page locking
- Fix a potential race with __wait_on_page_locked_async()
- Fix a hang related to not setting page_match, thus missing a wakeup

-- 
Jens Axboe

[PATCH 02/12] mm: allow read-ahead with IOCB_NOWAIT set

[Jens Axboe]

The read-ahead shouldn't block, so allow it to be done even if
IOCB_NOWAIT is set in the kiocb.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 mm/filemap.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/mm/filemap.c b/mm/filemap.c
index 23a051a7ef0f..80747f1377d5 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -2031,8 +2031,6 @@ static ssize_t generic_file_buffered_read(struct kiocb *iocb,
 
 		page = find_get_page(mapping, index);
 		if (!page) {
-			if (iocb->ki_flags & IOCB_NOWAIT)
-				goto would_block;
 			page_cache_sync_readahead(mapping,
 					ra, filp,
 					index, last_index - index);
-- 
2.26.2