Skip to contentSkip to conversation

cifs: Fix oops due to uninitialised variable

PATCH Applied David Howells 4 participants 4 messages Aug 19, 2025 – Sep 16, 2025

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 cifs: Fix oops due to uninitialised variable

Fix smb3_init_transform_rq() to initialise buffer to NULL before calling netfs_alloc_folioq_buffer() as netfs assumes it can append to the buffer it is given. Setting it to NULL means it should start a fresh buffer, but the value is currently undefined.

Fixes a2906d3316fc ("cifs: Switch crypto buffer to use a folio_queue rather than an xarray")
cc Steve French <[email protected]>
cc Paulo Alcantara <[email protected]>
---
 fs/smb/client/smb2ops.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/smb/client/smb2ops.c +1 -1
index ad8947434b71..cd0c9b5a35c3 100644
@@ -4487,7 +4487,7 @@ smb3_init_transform_rq(struct TCP_Server_Info *server, int num_rqst,
 	for (int i = 1; i < num_rqst; i++) {
 		struct smb_rqst *old = &old_rq[i - 1];
 		struct smb_rqst *new = &new_rq[i];
-		struct folio_queue *buffer;
+		struct folio_queue *buffer = NULL;
 		size_t size = iov_iter_count(&old->rq_iter);
 
 		orig_len += smb_rqst_len(server, old);
#1 lore raw
FYI: kernel test robot notices the stable kernel rule is not satisfied.

Hi,

Thanks for your patch.

FYI: kernel test robot notices the stable kernel rule is not satisfied.

The check is based on https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html#option-1

Rule: add the tag "Cc: [email protected]" in the sign-off area to have the patch automatically included in the stable tree.
Subject: [PATCH] cifs: Fix oops due to uninitialised variable

Signature · 2 lines
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
Applied
#2 lore raw
In reply to David Howells

merged into cifs-2.6.git for-next

Quoted 31 lines · David Howells

Fix smb3_init_transform_rq() to initialise buffer to NULL before calling netfs_alloc_folioq_buffer() as netfs assumes it can append to the buffer it is given. Setting it to NULL means it should start a fresh buffer, but the value is currently undefined.

Fixes a2906d3316fc ("cifs: Switch crypto buffer to use a folio_queue rather than an xarray")
cc Steve French <[email protected]>
cc Paulo Alcantara <[email protected]>
---
 fs/smb/client/smb2ops.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/smb/client/smb2ops.c +1 -1
index ad8947434b71..cd0c9b5a35c3 100644
@@ -4487,7 +4487,7 @@ smb3_init_transform_rq(struct TCP_Server_Info *server, int num_rqst,
        for (int i = 1; i < num_rqst; i++) {
                struct smb_rqst *old = &old_rq[i - 1];
                struct smb_rqst *new = &new_rq[i];
-               struct folio_queue *buffer;
+               struct folio_queue *buffer = NULL;
                size_t size = iov_iter_count(&old->rq_iter);
 
                orig_len += smb_rqst_len(server, old);
Signature · 3 lines
Thanks,

Steve
#3 lore raw
In reply to David Howells

Hi David, Steve,

Quoted 4 lines · David Howells

Fix smb3_init_transform_rq() to initialise buffer to NULL before calling netfs_alloc_folioq_buffer() as netfs assumes it can append to the buffer it is given. Setting it to NULL means it should start a fresh buffer, but the value is currently undefined.

This patch was based on David's RFC series
"netfs: [WIP] Allow the use of MSG_SPLICE_PAGES and use netmem allocator", specifically patch 15/31 "cifs: Use netfs_alloc/free_folioq_buffer()", which were never merged.

Current code in smb3_init_transform_rq() initializes buffer with cifs_alloc_folioq_buffer() and NULL-checked right after:

Quoted 12 lines

...

                struct folio_queue *buffer;
                size_t size = iov_iter_count(&old->rq_iter);

                orig_len += smb_rqst_len(server, old);
                new->rq_iov = old->rq_iov;
                new->rq_nvec = old->rq_nvec;

                if (size > 0) {
                        buffer = cifs_alloc_folioq_buffer(size);
                        if (!buffer)
                                goto err_free;
Quoted 1 line

...

Sorry not catching this earlier, but this just got my attention because there's now a CVE for this non-bug/vulnerability
https://nvd.nist.gov/vuln/detail/CVE-2025-38737

I don't know what exactly can/should be done, but I thought I'd let you know.

Cheers,

Enzo

Quoted 25 lines
Fixes a2906d3316fc ("cifs: Switch crypto buffer to use a folio_queue rather than an xarray")
cc Steve French <[email protected]>
cc Paulo Alcantara <[email protected]>

---
fs/smb/client/smb2ops.c | 2 +-

1 file changed, 1 insertion(+), 1 deletion(-)
fs/smb/client/smb2ops.c +1 -1
index ad8947434b71..cd0c9b5a35c3 100644
@@ -4487,7 +4487,7 @@ smb3_init_transform_rq(struct TCP_Server_Info *server, int num_rqst,
 	for (int i = 1; i < num_rqst; i++) {
 		struct smb_rqst *old = &old_rq[i - 1];
 		struct smb_rqst *new = &new_rq[i];
-		struct folio_queue *buffer;
+		struct folio_queue *buffer = NULL;
 		size_t size = iov_iter_count(&old->rq_iter);
 
 		orig_len += smb_rqst_len(server, old);