From: Pavel Begunkov <[email protected]>
To: Song Liu <[email protected]>
Cc: "[email protected]" <[email protected]>,
Networking <[email protected]>,
"[email protected]" <[email protected]>,
"[email protected]" <[email protected]>,
Jens Axboe <[email protected]>, Alexei Starovoitov <[email protected]>,
Daniel Borkmann <[email protected]>,
Andrii Nakryiko <[email protected]>, Martin Lau <[email protected]>,
Yonghong Song <[email protected]>,
John Fastabend <[email protected]>,
KP Singh <[email protected]>,
Horst Schirmeier <[email protected]>,
"Franz-B . Tuneke" <[email protected]>,
Christian Dietrich <[email protected]>
Subject: Re: [PATCH 01/23] io_uring: shuffle rarely used ctx fields
Date: Thu, 20 May 2021 23:46:02 +0100 [thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>
On 5/20/21 10:46 PM, Song Liu wrote:
>> On May 19, 2021, at 7:13 AM, Pavel Begunkov <[email protected]> wrote:
>> There is a bunch of scattered around ctx fields that are almost never
>> used, e.g. only on ring exit, plunge them to the end, better locality,
>> better aesthetically.
>>
>> Signed-off-by: Pavel Begunkov <[email protected]>
>> ---
>> fs/io_uring.c | 36 +++++++++++++++++-------------------
>> 1 file changed, 17 insertions(+), 19 deletions(-)
>>
>> diff --git a/fs/io_uring.c b/fs/io_uring.c
>> index 9ac5e278a91e..7e3410ce100a 100644
>> --- a/fs/io_uring.c
>> +++ b/fs/io_uring.c
>> @@ -367,9 +367,6 @@ struct io_ring_ctx {
>> unsigned cached_cq_overflow;
>> unsigned long sq_check_overflow;
>>
>> - /* hashed buffered write serialization */
>> - struct io_wq_hash *hash_map;
>> -
>> struct list_head defer_list;
>> struct list_head timeout_list;
>> struct list_head cq_overflow_list;
>> @@ -386,9 +383,6 @@ struct io_ring_ctx {
>>
>> struct io_rings *rings;
>>
>> - /* Only used for accounting purposes */
>> - struct mm_struct *mm_account;
>> -
>> const struct cred *sq_creds; /* cred used for __io_sq_thread() */
>> struct io_sq_data *sq_data; /* if using sq thread polling */
>>
>> @@ -409,14 +403,6 @@ struct io_ring_ctx {
>> unsigned nr_user_bufs;
>> struct io_mapped_ubuf **user_bufs;
>>
>> - struct user_struct *user;
>> -
>> - struct completion ref_comp;
>> -
>> -#if defined(CONFIG_UNIX)
>> - struct socket *ring_sock;
>> -#endif
>> -
>> struct xarray io_buffers;
>>
>> struct xarray personalities;
>> @@ -460,12 +446,24 @@ struct io_ring_ctx {
>>
>> struct io_restriction restrictions;
>>
>> - /* exit task_work */
>> - struct callback_head *exit_task_work;
>> -
>> /* Keep this last, we don't need it for the fast path */
>> - struct work_struct exit_work;
>> - struct list_head tctx_list;
>> + struct {
>
> Why do we need an anonymous struct here? For cache line alignment?
> Do we need ____cacheline_aligned_in_smp?
Rather as a visual hint considering that most of the field historically
are in structs (____cacheline_aligned_in_smp). Also preparing to
potentially splitting it out of the ctx struct as it grows big.
First 2-3 patches are not strictly related to bpf and will go separately
earlier, just the set was based on.
>> + #if defined(CONFIG_UNIX)
>> + struct socket *ring_sock;
>> + #endif
>> + /* hashed buffered write serialization */
>> + struct io_wq_hash *hash_map;
>> +
>> + /* Only used for accounting purposes */
>> + struct user_struct *user;
>> + struct mm_struct *mm_account;
>> +
>> + /* ctx exit and cancelation */
>> + struct callback_head *exit_task_work;
>> + struct work_struct exit_work;
>> + struct list_head tctx_list;
>> + struct completion ref_comp;
>> + };
>> };
>>
>> struct io_uring_task {
--
Pavel Begunkov
next prev parent reply other threads:[~2021-05-20 22:46 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-19 14:13 [RFC v2 00/23] io_uring BPF requests Pavel Begunkov
2021-05-19 14:13 ` [PATCH 01/23] io_uring: shuffle rarely used ctx fields Pavel Begunkov
2021-05-20 21:46 ` Song Liu
2021-05-20 22:46 ` Pavel Begunkov [this message]
2021-05-19 14:13 ` [PATCH 02/23] io_uring: localise fixed resources fields Pavel Begunkov
2021-05-19 14:13 ` [PATCH 03/23] io_uring: remove dependency on ring->sq/cq_entries Pavel Begunkov
2021-05-19 14:13 ` [PATCH 04/23] io_uring: deduce cq_mask from cq_entries Pavel Begunkov
2021-05-19 14:13 ` [PATCH 05/23] io_uring: kill cached_cq_overflow Pavel Begunkov
2021-05-19 14:13 ` [PATCH 06/23] io_uring: rename io_get_cqring Pavel Begunkov
2021-05-19 14:13 ` [PATCH 07/23] io_uring: extract struct for CQ Pavel Begunkov
2021-05-19 14:13 ` [PATCH 08/23] io_uring: internally pass CQ indexes Pavel Begunkov
2021-05-19 14:13 ` [PATCH 09/23] io_uring: extract cq size helper Pavel Begunkov
2021-05-19 14:13 ` [PATCH 10/23] io_uring: add support for multiple CQs Pavel Begunkov
2021-05-19 14:13 ` [PATCH 11/23] io_uring: enable mmap'ing additional CQs Pavel Begunkov
2021-05-19 14:13 ` [PATCH 12/23] bpf: add IOURING program type Pavel Begunkov
2021-05-20 23:34 ` Song Liu
2021-05-21 0:56 ` Pavel Begunkov
2021-05-19 14:13 ` [PATCH 13/23] io_uring: implement bpf prog registration Pavel Begunkov
2021-05-20 23:45 ` Song Liu
2021-05-21 0:43 ` Pavel Begunkov
2021-05-19 14:13 ` [PATCH 14/23] io_uring: add support for bpf requests Pavel Begunkov
2021-05-21 0:42 ` Pavel Begunkov
2021-05-19 14:13 ` [PATCH 15/23] io_uring: enable BPF to submit SQEs Pavel Begunkov
2021-05-21 0:06 ` Song Liu
2021-05-21 1:07 ` Alexei Starovoitov
2021-05-21 9:33 ` Pavel Begunkov
2021-05-19 14:13 ` [PATCH 16/23] io_uring: enable bpf to submit CQEs Pavel Begunkov
2021-05-19 14:13 ` [PATCH 17/23] io_uring: enable bpf to reap CQEs Pavel Begunkov
2021-05-19 14:13 ` [PATCH 18/23] libbpf: support io_uring Pavel Begunkov
2021-05-19 17:38 ` Andrii Nakryiko
2021-05-20 9:58 ` Pavel Begunkov
2021-05-20 17:23 ` Andrii Nakryiko
2021-05-19 14:13 ` [PATCH 19/23] io_uring: pass user_data to bpf executor Pavel Begunkov
2021-05-19 14:13 ` [PATCH 20/23] bpf: Add bpf_copy_to_user() helper Pavel Begunkov
2021-05-19 14:13 ` [PATCH 21/23] io_uring: wire bpf copy to user Pavel Begunkov
2021-05-19 14:13 ` [PATCH 22/23] io_uring: don't wait on CQ exclusively Pavel Begunkov
2021-05-19 14:13 ` [PATCH 23/23] io_uring: enable bpf reqs to wait for CQs Pavel Begunkov
2021-05-21 0:35 ` [RFC v2 00/23] io_uring BPF requests Song Liu
2021-05-21 0:58 ` Pavel Begunkov
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 \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[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