From: Christoph Hellwig <hch@lst.de>
To: Andrey Albershteyn <aalbersh@kernel.org>
Cc: "Darrick J . Wong" <djwong@kernel.org>, linux-xfs@vger.kernel.org
Subject: [PATCH 23/32] xfs: calculate the log reservation for logging data checksum buffers
Date: Thu, 24 Sep 2026 12:04:13 +0200 [thread overview]
Message-ID: <20260924100512.2733748-24-hch@lst.de> (raw)
In-Reply-To: <20260924100512.2733748-1-hch@lst.de>
Source kernel commit: 78980b5ca4f00b0f664396ff5a81f44762f5a07c
The data checksum is logged in its own transaction, and only logs
transactions buffers. While the maximum size of a checksummed data write
is the same as that of a single checksum buffer, the Zone Append based
zoned write path can't guarantee alignment, so it might be spread over
up to three buffers.
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
libxfs/xfs_trans_resv.c | 21 +++++++++++++++++++++
libxfs/xfs_trans_resv.h | 4 ++++
2 files changed, 25 insertions(+)
diff --git a/libxfs/xfs_trans_resv.c b/libxfs/xfs_trans_resv.c
index 3a9de562ac2b..cc785b1476cd 100644
--- a/libxfs/xfs_trans_resv.c
+++ b/libxfs/xfs_trans_resv.c
@@ -11,6 +11,7 @@
#include "xfs_log_format.h"
#include "xfs_trans_resv.h"
#include "xfs_mount.h"
+#include "xfs_rtcsumfile.h"
#include "xfs_da_format.h"
#include "xfs_da_btree.h"
#include "xfs_inode.h"
@@ -1226,6 +1227,22 @@ xfs_calc_qm_dqalloc_reservation_minlogsize(
return xfs_calc_qm_dqalloc_reservation(mp, true);
}
+/*
+ * Log data checksums for a write.
+ *
+ * Must cover a checksum for each FSB of data written, and the checksums can
+ * span the FSB-sized checksum buffers at both ends.
+ */
+unsigned int
+xfs_calc_csum_reservation(
+ struct xfs_mount *mp,
+ unsigned int csum_len)
+{
+ return xfs_calc_buf_res(
+ howmany(csum_len, xfs_rtcsum_payload_size(mp)) + 1,
+ mp->m_rtcsum_bsize);
+}
+
/*
* Syncing the incore super block changes to disk.
* the super block to reflect the changes: sector size
@@ -1347,6 +1364,10 @@ xfs_trans_resv_calc(
xfs_calc_namespace_reservations(mp, resp);
+ resp->tr_csum.tr_logres =
+ xfs_calc_csum_reservation(mp, XFS_RTCSUM_MAX_WRITE);
+ resp->tr_csum.tr_logcount = XFS_DEFAULT_LOG_COUNT;
+
/*
* The following transactions are logged in logical format with
* a default log count.
diff --git a/libxfs/xfs_trans_resv.h b/libxfs/xfs_trans_resv.h
index 1804e821f382..127db4da31c1 100644
--- a/libxfs/xfs_trans_resv.h
+++ b/libxfs/xfs_trans_resv.h
@@ -49,6 +49,7 @@ struct xfs_trans_resv {
struct xfs_trans_res tr_sb; /* modify superblock */
struct xfs_trans_res tr_fsyncts; /* update timestamps on fsync */
struct xfs_trans_res tr_atomic_ioend; /* untorn write completion */
+ struct xfs_trans_res tr_csum; /* data checksums in metafile */
};
/* shorthand way of accessing reservation structure */
@@ -122,6 +123,9 @@ unsigned int xfs_calc_itruncate_reservation_minlogsize(struct xfs_mount *mp);
unsigned int xfs_calc_write_reservation_minlogsize(struct xfs_mount *mp);
unsigned int xfs_calc_qm_dqalloc_reservation_minlogsize(struct xfs_mount *mp);
+unsigned int xfs_calc_csum_reservation(struct xfs_mount *mp,
+ unsigned int csum_len);
+
xfs_extlen_t xfs_calc_max_atomic_write_fsblocks(struct xfs_mount *mp);
xfs_extlen_t xfs_calc_atomic_write_log_geometry(struct xfs_mount *mp,
xfs_extlen_t blockcount, unsigned int *new_logres);
--
2.53.0
next prev parent reply other threads:[~2026-09-24 10:06 UTC|newest]
Thread overview: 57+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-24 10:03 xfsprogs support for RT data checksums Christoph Hellwig
2026-09-24 10:03 ` [PATCH 01/32] man: fix alignment of the rtstart field in ioctl_xfs_fsgeometry.2 Christoph Hellwig
2026-09-24 20:30 ` Darrick J. Wong
2026-09-26 6:15 ` Christoph Hellwig
2026-09-29 10:14 ` Andrey Albershteyn
2026-09-24 10:03 ` [PATCH 02/32] add cpu_to_le64 and le64_to_cpu_helpers Christoph Hellwig
2026-09-24 20:30 ` Darrick J. Wong
2026-09-24 10:03 ` [PATCH 03/32] libfrog: add a crc64_nvme implementation Christoph Hellwig
2026-09-24 10:03 ` [PATCH 04/32] libxfs: add DIV_ROUND_UP_ULL Christoph Hellwig
2026-09-24 20:44 ` Darrick J. Wong
2026-09-25 6:13 ` Christoph Hellwig
2026-09-24 10:03 ` [PATCH 05/32] libxfs: remove a spurious xfs_rtbitmap.h include in xfs_sb.c Christoph Hellwig
2026-09-25 23:27 ` Darrick J. Wong
2026-09-24 10:03 ` [PATCH 06/32] libxfs: add SZ_* constants Christoph Hellwig
2026-09-25 23:27 ` Darrick J. Wong
2026-09-24 10:03 ` [PATCH 07/32] xfs: remove spurious XBF_DONE clearing on readahead validation failure Christoph Hellwig
2026-09-24 10:03 ` [PATCH 08/32] xfs: hide b_flags manipulation from code outside of xfs_buf.c Christoph Hellwig
2026-09-24 10:03 ` [PATCH 09/32] FIXUP Christoph Hellwig
2026-09-25 23:29 ` Darrick J. Wong
2026-09-26 6:17 ` Christoph Hellwig
2026-09-24 10:04 ` [PATCH 10/32] xfs: add error injection for lazy bounce buffering Christoph Hellwig
2026-09-24 10:04 ` [PATCH 11/32] xfs: add xfs_daddr_to_rgno and xfs_daddr_to_rgbno helpers Christoph Hellwig
2026-09-24 10:04 ` [PATCH 12/32] FIXUP Christoph Hellwig
2026-09-24 10:04 ` [PATCH 13/32] xfs: introduce XFS_BLI_PREALLOC Christoph Hellwig
2026-09-24 10:04 ` [PATCH 14/32] xfs: centralize setting of buf_ops/buf_type/magic for rtblocks Christoph Hellwig
2026-09-24 10:04 ` [PATCH 15/32] xfs: factor out a xfs_rtfile_initialize_buf helper Christoph Hellwig
2026-09-24 10:04 ` [PATCH 16/32] xfs: add a xfs_rtblock_payload helper Christoph Hellwig
2026-09-24 10:04 ` [PATCH 17/32] xfs: prepare xfs_rtfile_initialize_blocks for larger than FSB blocks Christoph Hellwig
2026-09-24 10:04 ` [PATCH 18/32] FIXUP Christoph Hellwig
2026-09-24 10:04 ` [PATCH 19/32] xfs: define the RT data checksum on-disk format Christoph Hellwig
2026-09-24 10:04 ` [PATCH 20/32] FIXUP Christoph Hellwig
2026-09-24 10:04 ` [PATCH 21/32] xfs: add support for per-RTG csum files Christoph Hellwig
2026-09-24 10:04 ` [PATCH 22/32] FIXUP Christoph Hellwig
2026-09-24 10:04 ` Christoph Hellwig [this message]
2026-09-24 10:04 ` [PATCH 24/32] xfs: report RT data checksum information via XFS_FSOP_GEOM Christoph Hellwig
2026-09-24 10:04 ` [PATCH 25/32] xfs: enable RT data checksums Christoph Hellwig
2026-09-24 10:04 ` [PATCH 26/32] man: document the rtcsum geom fields Christoph Hellwig
2026-09-25 23:31 ` Darrick J. Wong
2026-09-26 6:17 ` Christoph Hellwig
2026-09-24 10:04 ` [PATCH 27/32] libfrog: print csum geometry information Christoph Hellwig
2026-09-25 23:32 ` Darrick J. Wong
2026-09-24 10:04 ` [PATCH 28/32] xfs_io: report checksum information from fs geometry in statfs Christoph Hellwig
2026-09-25 23:33 ` Darrick J. Wong
2026-09-26 6:18 ` Christoph Hellwig
2026-09-24 10:04 ` [PATCH 29/32] mkfs: support RT data checksums Christoph Hellwig
2026-09-25 23:39 ` Darrick J. Wong
2026-09-26 6:19 ` Christoph Hellwig
2026-09-24 10:04 ` [PATCH 30/32] xfs_db: support RT data checksum Christoph Hellwig
2026-09-25 23:43 ` Darrick J. Wong
2026-09-26 6:21 ` Christoph Hellwig
2026-09-26 18:20 ` Darrick J. Wong
2026-09-24 10:04 ` [PATCH 31/32] repair: support RT data checksums Christoph Hellwig
2026-09-25 23:54 ` Darrick J. Wong
2026-09-26 6:23 ` Christoph Hellwig
2026-09-27 0:01 ` Darrick J. Wong
2026-09-24 10:04 ` [PATCH 32/32] xfs_scrub: don't merge over unused space when data checksums are enabled Christoph Hellwig
2026-09-25 23:51 ` 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=20260924100512.2733748-24-hch@lst.de \
--to=hch@lst.de \
--cc=aalbersh@kernel.org \
--cc=djwong@kernel.org \
--cc=linux-xfs@vger.kernel.org \
/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