From: Pavel Begunkov <[email protected]>
To: [email protected]
Cc: [email protected]
Subject: [PATCH v2 6/6] io_uring: restore back registered wait arguments
Date: Thu, 14 Nov 2024 17:38:36 +0000 [thread overview]
Message-ID: <24cce6841e4d5ebb3a33bb602a94f5ded77757c3.1731604990.git.asml.silence@gmail.com> (raw)
In-Reply-To: <[email protected]>
Now we've got a more generic region registration API, place
IORING_ENTER_EXT_ARG_REG and re-enable it.
First, the user has to register a region with the
IORING_MEM_REGION_REG_WAIT_ARG flag set. It can only be done for a
ring in a disabled state, aka IORING_SETUP_R_DISABLED, to avoid races
with already running waiters.
The other API difference is that we're now passing byte offsets instead
of indexes. The user _must_ align all offsets / pointers to the native
word size, failing to do so might but not necessarily has to lead to a
failure usually returned as -EFAULT. liburing will be hiding this
details from users.
Signed-off-by: Pavel Begunkov <[email protected]>
---
include/linux/io_uring_types.h | 3 +++
include/uapi/linux/io_uring.h | 5 +++++
io_uring/io_uring.c | 14 +++++++++++++-
io_uring/register.c | 16 +++++++++++++++-
4 files changed, 36 insertions(+), 2 deletions(-)
diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h
index e1d69123e164..aa5f5ea98076 100644
--- a/include/linux/io_uring_types.h
+++ b/include/linux/io_uring_types.h
@@ -324,6 +324,9 @@ struct io_ring_ctx {
unsigned cq_entries;
struct io_ev_fd __rcu *io_ev_fd;
unsigned cq_extra;
+
+ void *cq_wait_arg;
+ size_t cq_wait_size;
} ____cacheline_aligned_in_smp;
/*
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index 1ee35890125b..4418d0192959 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -663,6 +663,11 @@ struct io_uring_region_desc {
__u64 __resv[4];
};
+enum {
+ /* expose the region as registered wait arguments */
+ IORING_MEM_REGION_REG_WAIT_ARG = 1,
+};
+
struct io_uring_mem_region_reg {
__u64 region_uptr; /* struct io_uring_region_desc * */
__u64 flags;
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index c640b8a4ceee..c93a6a9cd47e 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -3195,7 +3195,19 @@ void __io_uring_cancel(bool cancel_all)
static struct io_uring_reg_wait *io_get_ext_arg_reg(struct io_ring_ctx *ctx,
const struct io_uring_getevents_arg __user *uarg)
{
- return ERR_PTR(-EFAULT);
+ unsigned long size = sizeof(struct io_uring_reg_wait);
+ unsigned long offset = (uintptr_t)uarg;
+ unsigned long end;
+
+ if (unlikely(offset % sizeof(long)))
+ return ERR_PTR(-EFAULT);
+
+ /* also protects from NULL ->cq_wait_arg as the size would be 0 */
+ if (unlikely(check_add_overflow(offset, size, &end) ||
+ end >= ctx->cq_wait_size))
+ return ERR_PTR(-EFAULT);
+
+ return ctx->cq_wait_arg + offset;
}
static int io_validate_ext_arg(struct io_ring_ctx *ctx, unsigned flags,
diff --git a/io_uring/register.c b/io_uring/register.c
index 2cbac3d9b288..1a60f4916649 100644
--- a/io_uring/register.c
+++ b/io_uring/register.c
@@ -588,7 +588,16 @@ static int io_register_mem_region(struct io_ring_ctx *ctx, void __user *uarg)
if (memchr_inv(®.__resv, 0, sizeof(reg.__resv)))
return -EINVAL;
- if (reg.flags)
+ if (reg.flags & ~IORING_MEM_REGION_REG_WAIT_ARG)
+ return -EINVAL;
+
+ /*
+ * This ensures there are no waiters. Waiters are unlocked and it's
+ * hard to synchronise with them, especially if we need to initialise
+ * the region.
+ */
+ if ((reg.flags & IORING_MEM_REGION_REG_WAIT_ARG) &&
+ !(ctx->flags & IORING_SETUP_R_DISABLED))
return -EINVAL;
ret = io_create_region(ctx, &ctx->param_region, &rd);
@@ -598,6 +607,11 @@ static int io_register_mem_region(struct io_ring_ctx *ctx, void __user *uarg)
io_free_region(ctx, &ctx->param_region);
return -EFAULT;
}
+
+ if (reg.flags & IORING_MEM_REGION_REG_WAIT_ARG) {
+ ctx->cq_wait_arg = io_region_get_ptr(&ctx->param_region);
+ ctx->cq_wait_size = rd.size;
+ }
return 0;
}
--
2.46.0
prev parent reply other threads:[~2024-11-14 17:38 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-11-14 17:38 [PATCH v2 0/6] regions, param pre-mapping and reg waits extension Pavel Begunkov
2024-11-14 17:38 ` [PATCH v2 1/6] io_uring: fortify io_pin_pages with a warning Pavel Begunkov
2024-11-14 17:38 ` [PATCH v2 2/6] io_uring: disable ENTER_EXT_ARG_REG for IOPOLL Pavel Begunkov
2024-11-14 17:38 ` [PATCH v2 3/6] io_uring: temporarily disable registered waits Pavel Begunkov
2024-11-14 17:38 ` [PATCH v2 4/6] io_uring: introduce concept of memory regions Pavel Begunkov
2024-11-14 17:38 ` [PATCH v2 5/6] io_uring: add memory region registration Pavel Begunkov
2024-11-14 17:38 ` Pavel Begunkov [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=24cce6841e4d5ebb3a33bb602a94f5ded77757c3.1731604990.git.asml.silence@gmail.com \
[email protected] \
[email protected] \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox