lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZJOQHpKjghoGWYZ4@dread.disaster.area>
Date:   Thu, 22 Jun 2023 10:04:46 +1000
From:   Dave Chinner <david@...morbit.com>
To:     Jeremy Bongio <bongiojp@...il.com>
Cc:     Ted Tso <tytso@....edu>, "Darrick J . Wong" <djwong@...nel.org>,
        Allison Henderson <allison.henderson@...cle.com>,
        linux-ext4@...r.kernel.org, linux-fsdevel@...r.kernel.org,
        linux-xfs@...r.kernel.org
Subject: Re: [PATCH 1/1] For DIO writes with no mapped pages for inode, skip
 deferring completion.

On Wed, Jun 21, 2023 at 10:29:20AM -0700, Jeremy Bongio wrote:
> If there are no mapped pages for an DIO write then the page cache does not
> need to be updated. For very fast SSDs and direct async IO, deferring work
> completion can result in a significant performance loss.
> ---
>  fs/iomap/direct-io.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/iomap/direct-io.c b/fs/iomap/direct-io.c
> index 019cc87d0fb3..8f27d0dc4f6d 100644
> --- a/fs/iomap/direct-io.c
> +++ b/fs/iomap/direct-io.c
> @@ -168,7 +168,9 @@ void iomap_dio_bio_end_io(struct bio *bio)
>  			struct task_struct *waiter = dio->submit.waiter;
>  			WRITE_ONCE(dio->submit.waiter, NULL);
>  			blk_wake_io_task(waiter);
> -		} else if (dio->flags & IOMAP_DIO_WRITE) {
> +		} else if (dio->flags & IOMAP_DIO_WRITE &&
> +			(!dio->iocb->ki_filp->f_inode ||
> +			    dio->iocb->ki_filp->f_inode->i_mapping->nrpages))) {
>  			struct inode *inode = file_inode(dio->iocb->ki_filp);

Writes that are need O_DSYNC, unwritten extent conversion, file size
extension, etc all need to be deferred. This will break all of them,
as well as any other type of write that the filesystem itself needs
to run completion in task context.

Cheers,

Dave.

-- 
Dave Chinner
david@...morbit.com

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ