* [PATCH 6.12.y 2/7] nfsd: Drop dprintk in blocklayout xdr functions
2025-10-20 12:52 ` [PATCH 6.12.y 1/7] nfsd: Use correct error code when decoding extents Sasha Levin
@ 2025-10-20 12:53 ` Sasha Levin
2025-10-20 12:53 ` [PATCH 6.12.y 3/7] NFSD: Rework encoding and decoding of nfsd4_deviceid Sasha Levin
` (4 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: Sasha Levin @ 2025-10-20 12:53 UTC (permalink / raw)
To: stable
Cc: Sergey Bashirov, Jeff Layton, Christoph Hellwig, Chuck Lever,
Sasha Levin
From: Sergey Bashirov <sergeybashirov@gmail.com>
[ Upstream commit e339967eecf1305557f7c697e1bc10b5cc495454 ]
Minor clean up. Instead of dprintk there are appropriate error codes.
Signed-off-by: Sergey Bashirov <sergeybashirov@gmail.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Stable-dep-of: d68886bae76a ("NFSD: Fix last write offset handling in layoutcommit")
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
fs/nfsd/blocklayoutxdr.c | 40 +++++++---------------------------------
1 file changed, 7 insertions(+), 33 deletions(-)
diff --git a/fs/nfsd/blocklayoutxdr.c b/fs/nfsd/blocklayoutxdr.c
index 669ff8e6e966e..bcf21fde91207 100644
--- a/fs/nfsd/blocklayoutxdr.c
+++ b/fs/nfsd/blocklayoutxdr.c
@@ -139,28 +139,19 @@ nfsd4_block_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
struct iomap *iomaps;
u32 nr_iomaps, i;
- if (len < sizeof(u32)) {
- dprintk("%s: extent array too small: %u\n", __func__, len);
+ if (len < sizeof(u32))
return nfserr_bad_xdr;
- }
len -= sizeof(u32);
- if (len % PNFS_BLOCK_EXTENT_SIZE) {
- dprintk("%s: extent array invalid: %u\n", __func__, len);
+ if (len % PNFS_BLOCK_EXTENT_SIZE)
return nfserr_bad_xdr;
- }
nr_iomaps = be32_to_cpup(p++);
- if (nr_iomaps != len / PNFS_BLOCK_EXTENT_SIZE) {
- dprintk("%s: extent array size mismatch: %u/%u\n",
- __func__, len, nr_iomaps);
+ if (nr_iomaps != len / PNFS_BLOCK_EXTENT_SIZE)
return nfserr_bad_xdr;
- }
iomaps = kcalloc(nr_iomaps, sizeof(*iomaps), GFP_KERNEL);
- if (!iomaps) {
- dprintk("%s: failed to allocate extent array\n", __func__);
+ if (!iomaps)
return nfserr_delay;
- }
for (i = 0; i < nr_iomaps; i++) {
struct pnfs_block_extent bex;
@@ -170,26 +161,18 @@ nfsd4_block_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
p = xdr_decode_hyper(p, &bex.foff);
if (bex.foff & (block_size - 1)) {
- dprintk("%s: unaligned offset 0x%llx\n",
- __func__, bex.foff);
goto fail;
}
p = xdr_decode_hyper(p, &bex.len);
if (bex.len & (block_size - 1)) {
- dprintk("%s: unaligned length 0x%llx\n",
- __func__, bex.foff);
goto fail;
}
p = xdr_decode_hyper(p, &bex.soff);
if (bex.soff & (block_size - 1)) {
- dprintk("%s: unaligned disk offset 0x%llx\n",
- __func__, bex.soff);
goto fail;
}
bex.es = be32_to_cpup(p++);
if (bex.es != PNFS_BLOCK_READWRITE_DATA) {
- dprintk("%s: incorrect extent state %d\n",
- __func__, bex.es);
goto fail;
}
@@ -231,38 +214,29 @@ nfsd4_scsi_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
struct iomap *iomaps;
u32 nr_iomaps, expected, i;
- if (len < sizeof(u32)) {
- dprintk("%s: extent array too small: %u\n", __func__, len);
+ if (len < sizeof(u32))
return nfserr_bad_xdr;
- }
nr_iomaps = be32_to_cpup(p++);
expected = sizeof(__be32) + nr_iomaps * PNFS_SCSI_RANGE_SIZE;
- if (len != expected) {
- dprintk("%s: extent array size mismatch: %u/%u\n",
- __func__, len, expected);
+ if (len != expected)
return nfserr_bad_xdr;
- }
iomaps = kcalloc(nr_iomaps, sizeof(*iomaps), GFP_KERNEL);
- if (!iomaps) {
- dprintk("%s: failed to allocate extent array\n", __func__);
+ if (!iomaps)
return nfserr_delay;
- }
for (i = 0; i < nr_iomaps; i++) {
u64 val;
p = xdr_decode_hyper(p, &val);
if (val & (block_size - 1)) {
- dprintk("%s: unaligned offset 0x%llx\n", __func__, val);
goto fail;
}
iomaps[i].offset = val;
p = xdr_decode_hyper(p, &val);
if (val & (block_size - 1)) {
- dprintk("%s: unaligned length 0x%llx\n", __func__, val);
goto fail;
}
iomaps[i].length = val;
--
2.51.0
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 6.12.y 3/7] NFSD: Rework encoding and decoding of nfsd4_deviceid
2025-10-20 12:52 ` [PATCH 6.12.y 1/7] nfsd: Use correct error code when decoding extents Sasha Levin
2025-10-20 12:53 ` [PATCH 6.12.y 2/7] nfsd: Drop dprintk in blocklayout xdr functions Sasha Levin
@ 2025-10-20 12:53 ` Sasha Levin
2025-10-20 12:53 ` [PATCH 6.12.y 4/7] NFSD: Minor cleanup in layoutcommit processing Sasha Levin
` (3 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: Sasha Levin @ 2025-10-20 12:53 UTC (permalink / raw)
To: stable; +Cc: Sergey Bashirov, Chuck Lever, Sasha Levin
From: Sergey Bashirov <sergeybashirov@gmail.com>
[ Upstream commit 832738e4b325b742940761e10487403f9aad13e8 ]
Compilers may optimize the layout of C structures, so we should not rely
on sizeof struct and memcpy to encode and decode XDR structures. The byte
order of the fields should also be taken into account.
This patch adds the correct functions to handle the deviceid4 structure
and removes the pad field, which is currently not used by NFSD, from the
runtime state. The server's byte order is preserved because the deviceid4
blob on the wire is only used as a cookie by the client.
Signed-off-by: Sergey Bashirov <sergeybashirov@gmail.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Stable-dep-of: d68886bae76a ("NFSD: Fix last write offset handling in layoutcommit")
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
fs/nfsd/blocklayoutxdr.c | 7 ++-----
fs/nfsd/flexfilelayoutxdr.c | 3 +--
fs/nfsd/nfs4layouts.c | 1 -
fs/nfsd/nfs4xdr.c | 14 +-------------
fs/nfsd/xdr4.h | 36 +++++++++++++++++++++++++++++++++++-
5 files changed, 39 insertions(+), 22 deletions(-)
diff --git a/fs/nfsd/blocklayoutxdr.c b/fs/nfsd/blocklayoutxdr.c
index bcf21fde91207..18de37ff28916 100644
--- a/fs/nfsd/blocklayoutxdr.c
+++ b/fs/nfsd/blocklayoutxdr.c
@@ -29,8 +29,7 @@ nfsd4_block_encode_layoutget(struct xdr_stream *xdr,
*p++ = cpu_to_be32(len);
*p++ = cpu_to_be32(1); /* we always return a single extent */
- p = xdr_encode_opaque_fixed(p, &b->vol_id,
- sizeof(struct nfsd4_deviceid));
+ p = svcxdr_encode_deviceid4(p, &b->vol_id);
p = xdr_encode_hyper(p, b->foff);
p = xdr_encode_hyper(p, b->len);
p = xdr_encode_hyper(p, b->soff);
@@ -156,9 +155,7 @@ nfsd4_block_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
for (i = 0; i < nr_iomaps; i++) {
struct pnfs_block_extent bex;
- memcpy(&bex.vol_id, p, sizeof(struct nfsd4_deviceid));
- p += XDR_QUADLEN(sizeof(struct nfsd4_deviceid));
-
+ p = svcxdr_decode_deviceid4(p, &bex.vol_id);
p = xdr_decode_hyper(p, &bex.foff);
if (bex.foff & (block_size - 1)) {
goto fail;
diff --git a/fs/nfsd/flexfilelayoutxdr.c b/fs/nfsd/flexfilelayoutxdr.c
index aeb71c10ff1b9..f9f7e38cba13f 100644
--- a/fs/nfsd/flexfilelayoutxdr.c
+++ b/fs/nfsd/flexfilelayoutxdr.c
@@ -54,8 +54,7 @@ nfsd4_ff_encode_layoutget(struct xdr_stream *xdr,
*p++ = cpu_to_be32(1); /* single mirror */
*p++ = cpu_to_be32(1); /* single data server */
- p = xdr_encode_opaque_fixed(p, &fl->deviceid,
- sizeof(struct nfsd4_deviceid));
+ p = svcxdr_encode_deviceid4(p, &fl->deviceid);
*p++ = cpu_to_be32(1); /* efficiency */
diff --git a/fs/nfsd/nfs4layouts.c b/fs/nfsd/nfs4layouts.c
index fbfddd3c4c943..fc5e82eddaa1a 100644
--- a/fs/nfsd/nfs4layouts.c
+++ b/fs/nfsd/nfs4layouts.c
@@ -120,7 +120,6 @@ nfsd4_set_deviceid(struct nfsd4_deviceid *id, const struct svc_fh *fhp,
id->fsid_idx = fhp->fh_export->ex_devid_map->idx;
id->generation = device_generation;
- id->pad = 0;
return 0;
}
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index 90db900b346ce..bd5c8720ea7e3 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -566,18 +566,6 @@ nfsd4_decode_state_owner4(struct nfsd4_compoundargs *argp,
}
#ifdef CONFIG_NFSD_PNFS
-static __be32
-nfsd4_decode_deviceid4(struct nfsd4_compoundargs *argp,
- struct nfsd4_deviceid *devid)
-{
- __be32 *p;
-
- p = xdr_inline_decode(argp->xdr, NFS4_DEVICEID4_SIZE);
- if (!p)
- return nfserr_bad_xdr;
- memcpy(devid, p, sizeof(*devid));
- return nfs_ok;
-}
static __be32
nfsd4_decode_layoutupdate4(struct nfsd4_compoundargs *argp,
@@ -1762,7 +1750,7 @@ nfsd4_decode_getdeviceinfo(struct nfsd4_compoundargs *argp,
__be32 status;
memset(gdev, 0, sizeof(*gdev));
- status = nfsd4_decode_deviceid4(argp, &gdev->gd_devid);
+ status = nfsd4_decode_deviceid4(argp->xdr, &gdev->gd_devid);
if (status)
return status;
if (xdr_stream_decode_u32(argp->xdr, &gdev->gd_layout_type) < 0)
diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h
index 2a21a7662e030..83263bff27dc6 100644
--- a/fs/nfsd/xdr4.h
+++ b/fs/nfsd/xdr4.h
@@ -596,9 +596,43 @@ struct nfsd4_reclaim_complete {
struct nfsd4_deviceid {
u64 fsid_idx;
u32 generation;
- u32 pad;
};
+static inline __be32 *
+svcxdr_encode_deviceid4(__be32 *p, const struct nfsd4_deviceid *devid)
+{
+ __be64 *q = (__be64 *)p;
+
+ *q = (__force __be64)devid->fsid_idx;
+ p += 2;
+ *p++ = (__force __be32)devid->generation;
+ *p++ = xdr_zero;
+ return p;
+}
+
+static inline __be32 *
+svcxdr_decode_deviceid4(__be32 *p, struct nfsd4_deviceid *devid)
+{
+ __be64 *q = (__be64 *)p;
+
+ devid->fsid_idx = (__force u64)(*q);
+ p += 2;
+ devid->generation = (__force u32)(*p++);
+ p++; /* NFSD does not use the remaining octets */
+ return p;
+}
+
+static inline __be32
+nfsd4_decode_deviceid4(struct xdr_stream *xdr, struct nfsd4_deviceid *devid)
+{
+ __be32 *p = xdr_inline_decode(xdr, NFS4_DEVICEID4_SIZE);
+
+ if (unlikely(!p))
+ return nfserr_bad_xdr;
+ svcxdr_decode_deviceid4(p, devid);
+ return nfs_ok;
+}
+
struct nfsd4_layout_seg {
u32 iomode;
u64 offset;
--
2.51.0
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 6.12.y 4/7] NFSD: Minor cleanup in layoutcommit processing
2025-10-20 12:52 ` [PATCH 6.12.y 1/7] nfsd: Use correct error code when decoding extents Sasha Levin
2025-10-20 12:53 ` [PATCH 6.12.y 2/7] nfsd: Drop dprintk in blocklayout xdr functions Sasha Levin
2025-10-20 12:53 ` [PATCH 6.12.y 3/7] NFSD: Rework encoding and decoding of nfsd4_deviceid Sasha Levin
@ 2025-10-20 12:53 ` Sasha Levin
2025-10-20 12:53 ` [PATCH 6.12.y 5/7] NFSD: Implement large extent array support in pNFS Sasha Levin
` (2 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: Sasha Levin @ 2025-10-20 12:53 UTC (permalink / raw)
To: stable; +Cc: Sergey Bashirov, Christoph Hellwig, Chuck Lever, Sasha Levin
From: Sergey Bashirov <sergeybashirov@gmail.com>
[ Upstream commit 274365a51d88658fb51cca637ba579034e90a799 ]
Remove dprintk in nfsd4_layoutcommit. These are not needed
in day to day usage, and the information is also available
in Wireshark when capturing NFS traffic.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sergey Bashirov <sergeybashirov@gmail.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Stable-dep-of: d68886bae76a ("NFSD: Fix last write offset handling in layoutcommit")
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
fs/nfsd/nfs4proc.c | 12 +++---------
1 file changed, 3 insertions(+), 9 deletions(-)
diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c
index 8f2dc7eb4fc45..c01183ddc93f2 100644
--- a/fs/nfsd/nfs4proc.c
+++ b/fs/nfsd/nfs4proc.c
@@ -2379,18 +2379,12 @@ nfsd4_layoutcommit(struct svc_rqst *rqstp,
inode = d_inode(current_fh->fh_dentry);
nfserr = nfserr_inval;
- if (new_size <= seg->offset) {
- dprintk("pnfsd: last write before layout segment\n");
+ if (new_size <= seg->offset)
goto out;
- }
- if (new_size > seg->offset + seg->length) {
- dprintk("pnfsd: last write beyond layout segment\n");
+ if (new_size > seg->offset + seg->length)
goto out;
- }
- if (!lcp->lc_newoffset && new_size > i_size_read(inode)) {
- dprintk("pnfsd: layoutcommit beyond EOF\n");
+ if (!lcp->lc_newoffset && new_size > i_size_read(inode))
goto out;
- }
nfserr = nfsd4_preprocess_layout_stateid(rqstp, cstate, &lcp->lc_sid,
false, lcp->lc_layout_type,
--
2.51.0
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 6.12.y 5/7] NFSD: Implement large extent array support in pNFS
2025-10-20 12:52 ` [PATCH 6.12.y 1/7] nfsd: Use correct error code when decoding extents Sasha Levin
` (2 preceding siblings ...)
2025-10-20 12:53 ` [PATCH 6.12.y 4/7] NFSD: Minor cleanup in layoutcommit processing Sasha Levin
@ 2025-10-20 12:53 ` Sasha Levin
2025-10-20 12:53 ` [PATCH 6.12.y 6/7] NFSD: Fix last write offset handling in layoutcommit Sasha Levin
2025-10-20 12:53 ` [PATCH 6.12.y 7/7] [SEND COMMAND] Backport of d68886bae76a4b9b3484d23e5b7df086f940fa38 to 6.12.y Sasha Levin
5 siblings, 0 replies; 8+ messages in thread
From: Sasha Levin @ 2025-10-20 12:53 UTC (permalink / raw)
To: stable
Cc: Sergey Bashirov, Konstantin Evtushenko, Jeff Layton,
Christoph Hellwig, Chuck Lever, Sasha Levin
From: Sergey Bashirov <sergeybashirov@gmail.com>
[ Upstream commit f963cf2b91a30b5614c514f3ad53ca124cb65280 ]
When pNFS client in the block or scsi layout mode sends layoutcommit
to MDS, a variable length array of modified extents is supplied within
the request. This patch allows the server to accept such extent arrays
if they do not fit within single memory page.
The issue can be reproduced when writing to a 1GB file using FIO with
O_DIRECT, 4K block and large I/O depth without preallocation of the
file. In this case, the server returns NFSERR_BADXDR to the client.
Co-developed-by: Konstantin Evtushenko <koevtushenko@yandex.com>
Signed-off-by: Konstantin Evtushenko <koevtushenko@yandex.com>
Signed-off-by: Sergey Bashirov <sergeybashirov@gmail.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Stable-dep-of: d68886bae76a ("NFSD: Fix last write offset handling in layoutcommit")
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
fs/nfsd/blocklayout.c | 20 ++++++----
fs/nfsd/blocklayoutxdr.c | 83 +++++++++++++++++++++++++++-------------
fs/nfsd/blocklayoutxdr.h | 4 +-
fs/nfsd/nfs4proc.c | 2 +-
fs/nfsd/nfs4xdr.c | 11 +++---
fs/nfsd/pnfs.h | 1 +
fs/nfsd/xdr4.h | 3 +-
7 files changed, 78 insertions(+), 46 deletions(-)
diff --git a/fs/nfsd/blocklayout.c b/fs/nfsd/blocklayout.c
index 19078a043e85c..4c936132eb440 100644
--- a/fs/nfsd/blocklayout.c
+++ b/fs/nfsd/blocklayout.c
@@ -173,16 +173,18 @@ nfsd4_block_proc_getdeviceinfo(struct super_block *sb,
}
static __be32
-nfsd4_block_proc_layoutcommit(struct inode *inode,
+nfsd4_block_proc_layoutcommit(struct inode *inode, struct svc_rqst *rqstp,
struct nfsd4_layoutcommit *lcp)
{
struct iomap *iomaps;
int nr_iomaps;
__be32 nfserr;
- nfserr = nfsd4_block_decode_layoutupdate(lcp->lc_up_layout,
- lcp->lc_up_len, &iomaps, &nr_iomaps,
- i_blocksize(inode));
+ rqstp->rq_arg = lcp->lc_up_layout;
+ svcxdr_init_decode(rqstp);
+
+ nfserr = nfsd4_block_decode_layoutupdate(&rqstp->rq_arg_stream,
+ &iomaps, &nr_iomaps, i_blocksize(inode));
if (nfserr != nfs_ok)
return nfserr;
@@ -313,16 +315,18 @@ nfsd4_scsi_proc_getdeviceinfo(struct super_block *sb,
return nfserrno(nfsd4_block_get_device_info_scsi(sb, clp, gdp));
}
static __be32
-nfsd4_scsi_proc_layoutcommit(struct inode *inode,
+nfsd4_scsi_proc_layoutcommit(struct inode *inode, struct svc_rqst *rqstp,
struct nfsd4_layoutcommit *lcp)
{
struct iomap *iomaps;
int nr_iomaps;
__be32 nfserr;
- nfserr = nfsd4_scsi_decode_layoutupdate(lcp->lc_up_layout,
- lcp->lc_up_len, &iomaps, &nr_iomaps,
- i_blocksize(inode));
+ rqstp->rq_arg = lcp->lc_up_layout;
+ svcxdr_init_decode(rqstp);
+
+ nfserr = nfsd4_scsi_decode_layoutupdate(&rqstp->rq_arg_stream,
+ &iomaps, &nr_iomaps, i_blocksize(inode));
if (nfserr != nfs_ok)
return nfserr;
diff --git a/fs/nfsd/blocklayoutxdr.c b/fs/nfsd/blocklayoutxdr.c
index 18de37ff28916..e50afe3407371 100644
--- a/fs/nfsd/blocklayoutxdr.c
+++ b/fs/nfsd/blocklayoutxdr.c
@@ -113,8 +113,7 @@ nfsd4_block_encode_getdeviceinfo(struct xdr_stream *xdr,
/**
* nfsd4_block_decode_layoutupdate - decode the block layout extent array
- * @p: pointer to the xdr data
- * @len: number of bytes to decode
+ * @xdr: subbuf set to the encoded array
* @iomapp: pointer to store the decoded extent array
* @nr_iomapsp: pointer to store the number of extents
* @block_size: alignment of extent offset and length
@@ -127,25 +126,24 @@ nfsd4_block_encode_getdeviceinfo(struct xdr_stream *xdr,
*
* Return values:
* %nfs_ok: Successful decoding, @iomapp and @nr_iomapsp are valid
- * %nfserr_bad_xdr: The encoded array in @p is invalid
+ * %nfserr_bad_xdr: The encoded array in @xdr is invalid
* %nfserr_inval: An unaligned extent found
* %nfserr_delay: Failed to allocate memory for @iomapp
*/
__be32
-nfsd4_block_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
+nfsd4_block_decode_layoutupdate(struct xdr_stream *xdr, struct iomap **iomapp,
int *nr_iomapsp, u32 block_size)
{
struct iomap *iomaps;
- u32 nr_iomaps, i;
+ u32 nr_iomaps, expected, len, i;
+ __be32 nfserr;
- if (len < sizeof(u32))
- return nfserr_bad_xdr;
- len -= sizeof(u32);
- if (len % PNFS_BLOCK_EXTENT_SIZE)
+ if (xdr_stream_decode_u32(xdr, &nr_iomaps))
return nfserr_bad_xdr;
- nr_iomaps = be32_to_cpup(p++);
- if (nr_iomaps != len / PNFS_BLOCK_EXTENT_SIZE)
+ len = sizeof(__be32) + xdr_stream_remaining(xdr);
+ expected = sizeof(__be32) + nr_iomaps * PNFS_BLOCK_EXTENT_SIZE;
+ if (len != expected)
return nfserr_bad_xdr;
iomaps = kcalloc(nr_iomaps, sizeof(*iomaps), GFP_KERNEL);
@@ -155,21 +153,44 @@ nfsd4_block_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
for (i = 0; i < nr_iomaps; i++) {
struct pnfs_block_extent bex;
- p = svcxdr_decode_deviceid4(p, &bex.vol_id);
- p = xdr_decode_hyper(p, &bex.foff);
+ if (nfsd4_decode_deviceid4(xdr, &bex.vol_id)) {
+ nfserr = nfserr_bad_xdr;
+ goto fail;
+ }
+
+ if (xdr_stream_decode_u64(xdr, &bex.foff)) {
+ nfserr = nfserr_bad_xdr;
+ goto fail;
+ }
if (bex.foff & (block_size - 1)) {
+ nfserr = nfserr_inval;
+ goto fail;
+ }
+
+ if (xdr_stream_decode_u64(xdr, &bex.len)) {
+ nfserr = nfserr_bad_xdr;
goto fail;
}
- p = xdr_decode_hyper(p, &bex.len);
if (bex.len & (block_size - 1)) {
+ nfserr = nfserr_inval;
+ goto fail;
+ }
+
+ if (xdr_stream_decode_u64(xdr, &bex.soff)) {
+ nfserr = nfserr_bad_xdr;
goto fail;
}
- p = xdr_decode_hyper(p, &bex.soff);
if (bex.soff & (block_size - 1)) {
+ nfserr = nfserr_inval;
+ goto fail;
+ }
+
+ if (xdr_stream_decode_u32(xdr, &bex.es)) {
+ nfserr = nfserr_bad_xdr;
goto fail;
}
- bex.es = be32_to_cpup(p++);
if (bex.es != PNFS_BLOCK_READWRITE_DATA) {
+ nfserr = nfserr_inval;
goto fail;
}
@@ -182,13 +203,12 @@ nfsd4_block_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
return nfs_ok;
fail:
kfree(iomaps);
- return nfserr_inval;
+ return nfserr;
}
/**
* nfsd4_scsi_decode_layoutupdate - decode the scsi layout extent array
- * @p: pointer to the xdr data
- * @len: number of bytes to decode
+ * @xdr: subbuf set to the encoded array
* @iomapp: pointer to store the decoded extent array
* @nr_iomapsp: pointer to store the number of extents
* @block_size: alignment of extent offset and length
@@ -200,21 +220,22 @@ nfsd4_block_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
*
* Return values:
* %nfs_ok: Successful decoding, @iomapp and @nr_iomapsp are valid
- * %nfserr_bad_xdr: The encoded array in @p is invalid
+ * %nfserr_bad_xdr: The encoded array in @xdr is invalid
* %nfserr_inval: An unaligned extent found
* %nfserr_delay: Failed to allocate memory for @iomapp
*/
__be32
-nfsd4_scsi_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
+nfsd4_scsi_decode_layoutupdate(struct xdr_stream *xdr, struct iomap **iomapp,
int *nr_iomapsp, u32 block_size)
{
struct iomap *iomaps;
- u32 nr_iomaps, expected, i;
+ u32 nr_iomaps, expected, len, i;
+ __be32 nfserr;
- if (len < sizeof(u32))
+ if (xdr_stream_decode_u32(xdr, &nr_iomaps))
return nfserr_bad_xdr;
- nr_iomaps = be32_to_cpup(p++);
+ len = sizeof(__be32) + xdr_stream_remaining(xdr);
expected = sizeof(__be32) + nr_iomaps * PNFS_SCSI_RANGE_SIZE;
if (len != expected)
return nfserr_bad_xdr;
@@ -226,14 +247,22 @@ nfsd4_scsi_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
for (i = 0; i < nr_iomaps; i++) {
u64 val;
- p = xdr_decode_hyper(p, &val);
+ if (xdr_stream_decode_u64(xdr, &val)) {
+ nfserr = nfserr_bad_xdr;
+ goto fail;
+ }
if (val & (block_size - 1)) {
+ nfserr = nfserr_inval;
goto fail;
}
iomaps[i].offset = val;
- p = xdr_decode_hyper(p, &val);
+ if (xdr_stream_decode_u64(xdr, &val)) {
+ nfserr = nfserr_bad_xdr;
+ goto fail;
+ }
if (val & (block_size - 1)) {
+ nfserr = nfserr_inval;
goto fail;
}
iomaps[i].length = val;
@@ -244,5 +273,5 @@ nfsd4_scsi_decode_layoutupdate(__be32 *p, u32 len, struct iomap **iomapp,
return nfs_ok;
fail:
kfree(iomaps);
- return nfserr_inval;
+ return nfserr;
}
diff --git a/fs/nfsd/blocklayoutxdr.h b/fs/nfsd/blocklayoutxdr.h
index 15b3569f3d9ad..7d25ef689671f 100644
--- a/fs/nfsd/blocklayoutxdr.h
+++ b/fs/nfsd/blocklayoutxdr.h
@@ -54,9 +54,9 @@ __be32 nfsd4_block_encode_getdeviceinfo(struct xdr_stream *xdr,
const struct nfsd4_getdeviceinfo *gdp);
__be32 nfsd4_block_encode_layoutget(struct xdr_stream *xdr,
const struct nfsd4_layoutget *lgp);
-__be32 nfsd4_block_decode_layoutupdate(__be32 *p, u32 len,
+__be32 nfsd4_block_decode_layoutupdate(struct xdr_stream *xdr,
struct iomap **iomapp, int *nr_iomapsp, u32 block_size);
-__be32 nfsd4_scsi_decode_layoutupdate(__be32 *p, u32 len,
+__be32 nfsd4_scsi_decode_layoutupdate(struct xdr_stream *xdr,
struct iomap **iomapp, int *nr_iomapsp, u32 block_size);
#endif /* _NFSD_BLOCKLAYOUTXDR_H */
diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c
index c01183ddc93f2..294fede450a01 100644
--- a/fs/nfsd/nfs4proc.c
+++ b/fs/nfsd/nfs4proc.c
@@ -2407,7 +2407,7 @@ nfsd4_layoutcommit(struct svc_rqst *rqstp,
lcp->lc_size_chg = false;
}
- nfserr = ops->proc_layoutcommit(inode, lcp);
+ nfserr = ops->proc_layoutcommit(inode, rqstp, lcp);
nfs4_put_stid(&ls->ls_stid);
out:
return nfserr;
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index bd5c8720ea7e3..66383eeeed15a 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -571,6 +571,8 @@ static __be32
nfsd4_decode_layoutupdate4(struct nfsd4_compoundargs *argp,
struct nfsd4_layoutcommit *lcp)
{
+ u32 len;
+
if (xdr_stream_decode_u32(argp->xdr, &lcp->lc_layout_type) < 0)
return nfserr_bad_xdr;
if (lcp->lc_layout_type < LAYOUT_NFSV4_1_FILES)
@@ -578,13 +580,10 @@ nfsd4_decode_layoutupdate4(struct nfsd4_compoundargs *argp,
if (lcp->lc_layout_type >= LAYOUT_TYPE_MAX)
return nfserr_bad_xdr;
- if (xdr_stream_decode_u32(argp->xdr, &lcp->lc_up_len) < 0)
+ if (xdr_stream_decode_u32(argp->xdr, &len) < 0)
+ return nfserr_bad_xdr;
+ if (!xdr_stream_subsegment(argp->xdr, &lcp->lc_up_layout, len))
return nfserr_bad_xdr;
- if (lcp->lc_up_len > 0) {
- lcp->lc_up_layout = xdr_inline_decode(argp->xdr, lcp->lc_up_len);
- if (!lcp->lc_up_layout)
- return nfserr_bad_xdr;
- }
return nfs_ok;
}
diff --git a/fs/nfsd/pnfs.h b/fs/nfsd/pnfs.h
index 925817f669176..dfd411d1f363f 100644
--- a/fs/nfsd/pnfs.h
+++ b/fs/nfsd/pnfs.h
@@ -35,6 +35,7 @@ struct nfsd4_layout_ops {
const struct nfsd4_layoutget *lgp);
__be32 (*proc_layoutcommit)(struct inode *inode,
+ struct svc_rqst *rqstp,
struct nfsd4_layoutcommit *lcp);
void (*fence_client)(struct nfs4_layout_stateid *ls,
diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h
index 83263bff27dc6..c75b295df206a 100644
--- a/fs/nfsd/xdr4.h
+++ b/fs/nfsd/xdr4.h
@@ -665,8 +665,7 @@ struct nfsd4_layoutcommit {
u64 lc_last_wr; /* request */
struct timespec64 lc_mtime; /* request */
u32 lc_layout_type; /* request */
- u32 lc_up_len; /* layout length */
- void *lc_up_layout; /* decoded by callback */
+ struct xdr_buf lc_up_layout; /* decoded by callback */
bool lc_size_chg; /* response */
u64 lc_newsize; /* response */
};
--
2.51.0
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 6.12.y 6/7] NFSD: Fix last write offset handling in layoutcommit
2025-10-20 12:52 ` [PATCH 6.12.y 1/7] nfsd: Use correct error code when decoding extents Sasha Levin
` (3 preceding siblings ...)
2025-10-20 12:53 ` [PATCH 6.12.y 5/7] NFSD: Implement large extent array support in pNFS Sasha Levin
@ 2025-10-20 12:53 ` Sasha Levin
2025-10-20 12:53 ` [PATCH 6.12.y 7/7] [SEND COMMAND] Backport of d68886bae76a4b9b3484d23e5b7df086f940fa38 to 6.12.y Sasha Levin
5 siblings, 0 replies; 8+ messages in thread
From: Sasha Levin @ 2025-10-20 12:53 UTC (permalink / raw)
To: stable
Cc: Sergey Bashirov, Konstantin Evtushenko, Christoph Hellwig,
Jeff Layton, Chuck Lever, Sasha Levin
From: Sergey Bashirov <sergeybashirov@gmail.com>
[ Upstream commit d68886bae76a4b9b3484d23e5b7df086f940fa38 ]
The data type of loca_last_write_offset is newoffset4 and is switched
on a boolean value, no_newoffset, that indicates if a previous write
occurred or not. If no_newoffset is FALSE, an offset is not given.
This means that client does not try to update the file size. Thus,
server should not try to calculate new file size and check if it fits
into the segment range. See RFC 8881, section 12.5.4.2.
Sometimes the current incorrect logic may cause clients to hang when
trying to sync an inode. If layoutcommit fails, the client marks the
inode as dirty again.
Fixes: 9cf514ccfacb ("nfsd: implement pNFS operations")
Cc: stable@vger.kernel.org
Co-developed-by: Konstantin Evtushenko <koevtushenko@yandex.com>
Signed-off-by: Konstantin Evtushenko <koevtushenko@yandex.com>
Signed-off-by: Sergey Bashirov <sergeybashirov@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
fs/nfsd/blocklayout.c | 5 ++---
fs/nfsd/nfs4proc.c | 30 +++++++++++++++---------------
2 files changed, 17 insertions(+), 18 deletions(-)
diff --git a/fs/nfsd/blocklayout.c b/fs/nfsd/blocklayout.c
index 4c936132eb440..0822d8a119c6f 100644
--- a/fs/nfsd/blocklayout.c
+++ b/fs/nfsd/blocklayout.c
@@ -118,7 +118,6 @@ nfsd4_block_commit_blocks(struct inode *inode, struct nfsd4_layoutcommit *lcp,
struct iomap *iomaps, int nr_iomaps)
{
struct timespec64 mtime = inode_get_mtime(inode);
- loff_t new_size = lcp->lc_last_wr + 1;
struct iattr iattr = { .ia_valid = 0 };
int error;
@@ -128,9 +127,9 @@ nfsd4_block_commit_blocks(struct inode *inode, struct nfsd4_layoutcommit *lcp,
iattr.ia_valid |= ATTR_ATIME | ATTR_CTIME | ATTR_MTIME;
iattr.ia_atime = iattr.ia_ctime = iattr.ia_mtime = lcp->lc_mtime;
- if (new_size > i_size_read(inode)) {
+ if (lcp->lc_size_chg) {
iattr.ia_valid |= ATTR_SIZE;
- iattr.ia_size = new_size;
+ iattr.ia_size = lcp->lc_newsize;
}
error = inode->i_sb->s_export_op->commit_blocks(inode, iomaps,
diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c
index 294fede450a01..6040de0923f85 100644
--- a/fs/nfsd/nfs4proc.c
+++ b/fs/nfsd/nfs4proc.c
@@ -2362,7 +2362,6 @@ nfsd4_layoutcommit(struct svc_rqst *rqstp,
const struct nfsd4_layout_seg *seg = &lcp->lc_seg;
struct svc_fh *current_fh = &cstate->current_fh;
const struct nfsd4_layout_ops *ops;
- loff_t new_size = lcp->lc_last_wr + 1;
struct inode *inode;
struct nfs4_layout_stateid *ls;
__be32 nfserr;
@@ -2378,13 +2377,21 @@ nfsd4_layoutcommit(struct svc_rqst *rqstp,
goto out;
inode = d_inode(current_fh->fh_dentry);
- nfserr = nfserr_inval;
- if (new_size <= seg->offset)
- goto out;
- if (new_size > seg->offset + seg->length)
- goto out;
- if (!lcp->lc_newoffset && new_size > i_size_read(inode))
- goto out;
+ lcp->lc_size_chg = false;
+ if (lcp->lc_newoffset) {
+ loff_t new_size = lcp->lc_last_wr + 1;
+
+ nfserr = nfserr_inval;
+ if (new_size <= seg->offset)
+ goto out;
+ if (new_size > seg->offset + seg->length)
+ goto out;
+
+ if (new_size > i_size_read(inode)) {
+ lcp->lc_size_chg = true;
+ lcp->lc_newsize = new_size;
+ }
+ }
nfserr = nfsd4_preprocess_layout_stateid(rqstp, cstate, &lcp->lc_sid,
false, lcp->lc_layout_type,
@@ -2400,13 +2407,6 @@ nfsd4_layoutcommit(struct svc_rqst *rqstp,
/* LAYOUTCOMMIT does not require any serialization */
mutex_unlock(&ls->ls_mutex);
- if (new_size > i_size_read(inode)) {
- lcp->lc_size_chg = true;
- lcp->lc_newsize = new_size;
- } else {
- lcp->lc_size_chg = false;
- }
-
nfserr = ops->proc_layoutcommit(inode, rqstp, lcp);
nfs4_put_stid(&ls->ls_stid);
out:
--
2.51.0
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 6.12.y 7/7] [SEND COMMAND] Backport of d68886bae76a4b9b3484d23e5b7df086f940fa38 to 6.12.y
2025-10-20 12:52 ` [PATCH 6.12.y 1/7] nfsd: Use correct error code when decoding extents Sasha Levin
` (4 preceding siblings ...)
2025-10-20 12:53 ` [PATCH 6.12.y 6/7] NFSD: Fix last write offset handling in layoutcommit Sasha Levin
@ 2025-10-20 12:53 ` Sasha Levin
5 siblings, 0 replies; 8+ messages in thread
From: Sasha Levin @ 2025-10-20 12:53 UTC (permalink / raw)
To: stable; +Cc: Sasha Levin
Send command:
git send-email --to stable@vger.kernel.org --in-reply-to '2025101657-preseason-garnish-c1e0@gregkh' --subject-prefix 'PATCH 6.12.y' stable/linux-6.12.y..HEAD
--
2.51.0
^ permalink raw reply [flat|nested] 8+ messages in thread