linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Hao Xu <hao.xu@linux.dev>
To: Jens Axboe <axboe@kernel.dk>,
	"Darrick J. Wong" <djwong@kernel.org>,
	Christian Brauner <brauner@kernel.org>
Cc: Dave Chinner <david@fromorbit.com>,
	io-uring@vger.kernel.org,
	Dominique Martinet <asmadeus@codewreck.org>,
	Pavel Begunkov <asml.silence@gmail.com>,
	Alexander Viro <viro@zeniv.linux.org.uk>,
	Stefan Roesch <shr@fb.com>, Clay Harris <bugs@claycon.org>,
	linux-fsdevel@vger.kernel.org, Wanpeng Li <wanpengli@tencent.com>
Subject: Re: [PATCH 3/5] io_uring: add support for getdents
Date: Tue, 8 Aug 2023 17:33:54 +0800	[thread overview]
Message-ID: <3aacee27-6e30-ef50-ff1f-9fc1334c8924@linux.dev> (raw)
In-Reply-To: <22630618-40fc-5668-078d-6cefcb2e4962@kernel.dk>

On 8/1/23 08:28, Jens Axboe wrote:
> On 7/31/23 9:26?AM, Darrick J. Wong wrote:
>> I've watched quite a bit of NOWAIT whackamole going on over the past few
>> years (i_rwsem, the ILOCK, the IO layer, memory allocations...). IIRC
>> these filesystem ios all have to run in process context, right? If so,
>> why don't we capture the NOWAIT state in a PF flag? We already do that
>> for NOFS/NOIO memory allocations to make sure that /all/ reclaim
>> attempts cannot recurse into the fs/io stacks.
>
> I would greatly prefer passing down the context rather than capitulating
> and adding a task_struct flag for this. I think it _kind of_ makes sense
> for things like allocations, as you cannot easily track that all the way
> down, but it's a really ugly solution. It certainly creates more churn
> passing it down, but it also reveals the parts that need to check it.
> WHen new code is added, it's much more likely you'll spot the fact that
> there's passed in context. For allocation, you end up in the allocator
> anyway, which can augment the gfp mask with whatever is set in the task.
> The same is not true for locking and other bits, as they don't return a
> value to begin with. When we know they are sane, we can flag the fs as
> supporting it (like we've done for async buffered reads, for example).
>
> It's also not an absolute thing, like memory allocations are. It's
> perfectly fine to grab a mutex under NOWAIT issue. What you should not

Hi Jens,
To make sure, I'd like to ask, for memory allocation, GFP_NOIO semantics
is all we need in NOWAIT issue, GFP_NOWAIT is not necessary, do I
understand it right?

Thanks,
Hao

> do is grab a mutex that someone else can grab while waiting on IO. This
> kind of extra context is only available in the code in question, not
> generically for eg mutex locking.
>
> I'm not a huge fan of the "let's add a bool nowait". Most/all use cases
> pass down state anyway, putting it in a suitable type struct seems much
> cleaner to me than the out-of-band approach of just adding a
> current->please_nowait.
>



  parent reply	other threads:[~2023-08-08 16:14 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-18 13:21 [PATCH v4 0/5] io_uring getdents Hao Xu
2023-07-18 13:21 ` [PATCH 1/5] fs: split off vfs_getdents function of getdents64 syscall Hao Xu
2023-07-18 13:21 ` [PATCH 2/5] vfs_getdents/struct dir_context: add flags field Hao Xu
2023-07-18 13:21 ` [PATCH 3/5] io_uring: add support for getdents Hao Xu
2023-07-19  8:56   ` Hao Xu
2023-07-26 15:00   ` Christian Brauner
2023-07-27 11:51     ` Hao Xu
2023-07-27 14:27       ` Christian Brauner
2023-07-27 15:12         ` Pavel Begunkov
2023-07-27 15:52           ` Christian Brauner
2023-07-27 16:17             ` Pavel Begunkov
2023-07-27 16:28               ` Christian Brauner
2023-07-31  1:58                 ` Dave Chinner
2023-07-31  7:34                   ` Hao Xu
2023-07-31  7:50                     ` Christian Brauner
2023-07-31  7:40                   ` Christian Brauner
2023-07-30 18:02         ` Hao Xu
2023-07-31  8:18           ` Christian Brauner
2023-07-31  9:31             ` Hao Xu
2023-07-31  1:33         ` Dave Chinner
2023-07-31  8:13           ` Christian Brauner
2023-07-31 15:26             ` Darrick J. Wong
2023-07-31 22:18               ` Dave Chinner
2023-08-01  0:28               ` Jens Axboe
2023-08-01  0:47                 ` Matthew Wilcox
2023-08-01  0:49                   ` Jens Axboe
2023-08-01  1:01                     ` Matthew Wilcox
2023-08-01  7:00                       ` Christian Brauner
2023-08-01  6:59                     ` Christian Brauner
2023-08-01  7:17                 ` Christian Brauner
2023-08-08  4:34                 ` Hao Xu
2023-08-08  5:18                   ` Hao Xu
2023-08-08  9:33                 ` Hao Xu [this message]
2023-08-08 22:55                   ` Jens Axboe
2023-08-01 18:39             ` Hao Xu
2023-07-18 13:21 ` [PATCH 4/5] xfs: add NOWAIT semantics for readdir Hao Xu
2023-07-19  2:35   ` kernel test robot
2023-07-18 13:21 ` [PATCH RFC 5/5] disable fixed file for io_uring getdents for now Hao Xu
2023-07-26 14:23   ` Christian Brauner
2023-07-27 12:09     ` Hao Xu
2023-07-19  6:04 ` [PATCH v4 0/5] io_uring getdents Christian Brauner

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=3aacee27-6e30-ef50-ff1f-9fc1334c8924@linux.dev \
    --to=hao.xu@linux.dev \
    --cc=asmadeus@codewreck.org \
    --cc=asml.silence@gmail.com \
    --cc=axboe@kernel.dk \
    --cc=brauner@kernel.org \
    --cc=bugs@claycon.org \
    --cc=david@fromorbit.com \
    --cc=djwong@kernel.org \
    --cc=io-uring@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=shr@fb.com \
    --cc=viro@zeniv.linux.org.uk \
    --cc=wanpengli@tencent.com \
    /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;
as well as URLs for NNTP newsgroup(s).