public inbox for linux-fsdevel@vger.kernel.org
 help / color / mirror / Atom feed
From: "Darrick J. Wong" <djwong@kernel.org>
To: John Garry <john.g.garry@oracle.com>
Cc: brauner@kernel.org, cem@kernel.org, dchinner@redhat.com,
	hch@lst.de, ritesh.list@gmail.com, linux-xfs@vger.kernel.org,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	martin.petersen@oracle.com
Subject: Re: [PATCH v3 6/7] xfs: Add RT atomic write unit max to xfs_mount
Date: Tue, 7 Jan 2025 16:55:52 -0800	[thread overview]
Message-ID: <20250108005552.GC1306365@frogsfrogsfrogs> (raw)
In-Reply-To: <20250102140411.14617-7-john.g.garry@oracle.com>

On Thu, Jan 02, 2025 at 02:04:10PM +0000, John Garry wrote:
> rtvol guarantees alloc unit alignment through rt_extsize. As such, it is
> possible to atomically write multiple FS blocks in a rtvol (up to
> rt_extsize).
> 
> Add a member to xfs_mount to hold the pre-calculated atomic write unit max.
> 
> The value in rt_extsize is not necessarily a power-of-2, so find the
> largest power-of-2 evenly divisible into rt_extsize.
> 
> Signed-off-by: John Garry <john.g.garry@oracle.com>

I guess that works.
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>

--D

> ---
>  fs/xfs/libxfs/xfs_sb.c |  3 +++
>  fs/xfs/xfs_mount.h     |  1 +
>  fs/xfs/xfs_rtalloc.c   | 23 +++++++++++++++++++++++
>  fs/xfs/xfs_rtalloc.h   |  4 ++++
>  4 files changed, 31 insertions(+)
> 
> diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
> index 3b5623611eba..6381060df901 100644
> --- a/fs/xfs/libxfs/xfs_sb.c
> +++ b/fs/xfs/libxfs/xfs_sb.c
> @@ -25,6 +25,7 @@
>  #include "xfs_da_format.h"
>  #include "xfs_health.h"
>  #include "xfs_ag.h"
> +#include "xfs_rtalloc.h"
>  #include "xfs_rtbitmap.h"
>  #include "xfs_exchrange.h"
>  #include "xfs_rtgroup.h"
> @@ -1149,6 +1150,8 @@ xfs_sb_mount_rextsize(
>  		rgs->blklog = 0;
>  		rgs->blkmask = (uint64_t)-1;
>  	}
> +
> +	xfs_rt_awu_update(mp);
>  }
>  
>  /* Update incore sb rt extent size, then recompute the cached rt geometry. */
> diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
> index db9dade7d22a..f2f1d2c667cc 100644
> --- a/fs/xfs/xfs_mount.h
> +++ b/fs/xfs/xfs_mount.h
> @@ -191,6 +191,7 @@ typedef struct xfs_mount {
>  	bool			m_fail_unmount;
>  	bool			m_finobt_nores; /* no per-AG finobt resv. */
>  	bool			m_update_sb;	/* sb needs update in mount */
> +	xfs_extlen_t		m_rt_awu_max;   /* rt atomic write unit max */
>  
>  	/*
>  	 * Bitsets of per-fs metadata that have been checked and/or are sick.
> diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
> index fcfa6e0eb3ad..e3093f3c7670 100644
> --- a/fs/xfs/xfs_rtalloc.c
> +++ b/fs/xfs/xfs_rtalloc.c
> @@ -735,6 +735,28 @@ xfs_rtginode_ensure(
>  	return xfs_rtginode_create(rtg, type, true);
>  }
>  
> +void
> +xfs_rt_awu_update(
> +	struct xfs_mount	*mp)
> +{
> +	xfs_agblock_t		rsize = mp->m_sb.sb_rextsize;
> +	xfs_extlen_t		awu_max;
> +
> +	if (is_power_of_2(rsize)) {
> +		mp->m_rt_awu_max = rsize;
> +		return;
> +	}
> +
> +	/* Find highest power-of-2 evenly divisible into sb_rextsize */
> +	awu_max = 1;
> +	while (1) {
> +		if (rsize % (awu_max * 2))
> +			break;
> +		awu_max *= 2;
> +	}
> +	mp->m_rt_awu_max = awu_max;
> +}
> +
>  static struct xfs_mount *
>  xfs_growfs_rt_alloc_fake_mount(
>  	const struct xfs_mount	*mp,
> @@ -969,6 +991,7 @@ xfs_growfs_rt_bmblock(
>  	 */
>  	mp->m_rsumlevels = nmp->m_rsumlevels;
>  	mp->m_rsumblocks = nmp->m_rsumblocks;
> +	mp->m_rt_awu_max = nmp->m_rt_awu_max;
>  
>  	/*
>  	 * Recompute the growfsrt reservation from the new rsumsize.
> diff --git a/fs/xfs/xfs_rtalloc.h b/fs/xfs/xfs_rtalloc.h
> index 8e2a07b8174b..fcb7bb3df470 100644
> --- a/fs/xfs/xfs_rtalloc.h
> +++ b/fs/xfs/xfs_rtalloc.h
> @@ -42,6 +42,10 @@ xfs_growfs_rt(
>  	struct xfs_mount	*mp,	/* file system mount structure */
>  	xfs_growfs_rt_t		*in);	/* user supplied growfs struct */
>  
> +void
> +xfs_rt_awu_update(
> +	struct xfs_mount	*mp);
> +
>  int xfs_rtalloc_reinit_frextents(struct xfs_mount *mp);
>  #else
>  # define xfs_growfs_rt(mp,in)				(-ENOSYS)
> -- 
> 2.31.1
> 
> 

  reply	other threads:[~2025-01-08  0:55 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-01-02 14:04 [PATCH v3 0/7] large atomic writes for xfs John Garry
2025-01-02 14:04 ` [PATCH v3 1/7] iomap: Increase iomap_dio_zero() size limit John Garry
2025-01-02 14:04 ` [PATCH v3 2/7] iomap: Add zero unwritten mappings dio support John Garry
2025-01-02 14:04 ` [PATCH v3 3/7] iomap: Lift blocksize restriction on atomic writes John Garry
2025-01-08  0:41   ` Darrick J. Wong
2025-01-02 14:04 ` [PATCH v3 4/7] xfs: Add extent zeroing support for " John Garry
2025-01-02 14:04 ` [PATCH v3 5/7] xfs: Switch atomic write size check in xfs_file_write_iter() John Garry
2025-01-08  0:50   ` Darrick J. Wong
2025-01-02 14:04 ` [PATCH v3 6/7] xfs: Add RT atomic write unit max to xfs_mount John Garry
2025-01-08  0:55   ` Darrick J. Wong [this message]
2025-01-02 14:04 ` [PATCH v3 7/7] xfs: Update xfs_get_atomic_write_attr() for large atomic writes John Garry
2025-01-08  0:56   ` Darrick J. Wong

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=20250108005552.GC1306365@frogsfrogsfrogs \
    --to=djwong@kernel.org \
    --cc=brauner@kernel.org \
    --cc=cem@kernel.org \
    --cc=dchinner@redhat.com \
    --cc=hch@lst.de \
    --cc=john.g.garry@oracle.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-xfs@vger.kernel.org \
    --cc=martin.petersen@oracle.com \
    --cc=ritesh.list@gmail.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