Received-SPF: pass (google.com: domain of linux-ext4-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 20 May 2022 13:27:39 +1000
From:   Dave Chinner <david@fromorbit.com>
To:     "Darrick J. Wong" <djwong@kernel.org>
Cc:     Eric Biggers <ebiggers@kernel.org>, linux-fsdevel@vger.kernel.org,
        linux-ext4@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net,
        linux-xfs@vger.kernel.org, linux-api@vger.kernel.org,
        linux-fscrypt@vger.kernel.org, linux-block@vger.kernel.org,
        linux-kernel@vger.kernel.org, Keith Busch <kbusch@kernel.org>
Subject: Re: [RFC PATCH v2 1/7] statx: add I/O alignment information
Message-ID: <20220520032739.GB1098723@dread.disaster.area>
References: <20220518235011.153058-1-ebiggers@kernel.org>
 <20220518235011.153058-2-ebiggers@kernel.org>
 <YobNXbYnhBiqniTH@magnolia>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YobNXbYnhBiqniTH@magnolia>
Precedence: bulk

On Thu, May 19, 2022 at 04:06:05PM -0700, Darrick J. Wong wrote:
> On Wed, May 18, 2022 at 04:50:05PM -0700, Eric Biggers wrote:
> > From: Eric Biggers <ebiggers@google.com>
> > 
> > Traditionally, the conditions for when DIO (direct I/O) is supported
> > were fairly simple: filesystems either supported DIO aligned to the
> > block device's logical block size, or didn't support DIO at all.
> > 
> > However, due to filesystem features that have been added over time (e.g,
> > data journalling, inline data, encryption, verity, compression,
> > checkpoint disabling, log-structured mode), the conditions for when DIO
> > is allowed on a file have gotten increasingly complex.  Whether a
> > particular file supports DIO, and with what alignment, can depend on
> > various file attributes and filesystem mount options, as well as which
> > block device(s) the file's data is located on.
> > 
> > XFS has an ioctl XFS_IOC_DIOINFO which exposes this information to
> > applications.  However, as discussed
> > (https://lore.kernel.org/linux-fsdevel/20220120071215.123274-1-ebiggers@kernel.org/T/#u),
> > this ioctl is rarely used and not known to be used outside of
> > XFS-specific code.  It also was never intended to indicate when a file
> > doesn't support DIO at all, and it only exposes the minimum I/O
> > alignment, not the optimal I/O alignment which has been requested too.
> > 
> > Therefore, let's expose this information via statx().  Add the
> > STATX_IOALIGN flag and three fields associated with it:
> > 
> > * stx_mem_align_dio: the alignment (in bytes) required for user memory
> >   buffers for DIO, or 0 if DIO is not supported on the file.
> > 
> > * stx_offset_align_dio: the alignment (in bytes) required for file
> >   offsets and I/O segment lengths for DIO, or 0 if DIO is not supported
> >   on the file.  This will only be nonzero if stx_mem_align_dio is
> >   nonzero, and vice versa.
> > 
> > * stx_offset_align_optimal: the alignment (in bytes) suggested for file
> >   offsets and I/O segment lengths to get optimal performance.  This
> >   applies to both DIO and buffered I/O.  It differs from stx_blocksize
> >   in that stx_offset_align_optimal will contain the real optimum I/O
> >   size, which may be a large value.  In contrast, for compatibility
> >   reasons stx_blocksize is the minimum size needed to avoid page cache
> >   read/write/modify cycles, which may be much smaller than the optimum
> >   I/O size.  For more details about the motivation for this field, see
> >   https://lore.kernel.org/r/20220210040304.GM59729@dread.disaster.area
> 
> Hmm.  So I guess this is supposed to be the filesystem's best guess at
> the IO size that will minimize RMW cycles in the entire stack?  i.e. if
> the user does not want RMW of pagecache pages, of file allocation units
> (if COW is enabled), of RAID stripes, or in the storage itself, then it
> should ensure that all IOs are aligned to this value?
> 
> I guess that means for XFS it's effectively max(pagesize, i_blocksize,
> bdev io_opt, sb_width, and (pretend XFS can reflink the realtime volume)
> the rt extent size)?  I didn't see a manpage update for statx(2) but
> that's mostly what I'm interested in. :)

Yup, xfs_stat_blksize() should give a good idea of what we should
do. It will end up being pretty much that, except without the need
to a mount option to turn on the sunit/swidth return, and always
taking into consideration extent size hints rather than just doing
that for RT inodes...

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com