diff options
| author | Paulo Alcantara <pc@manguebit.org> | 2026-10-11 23:08:54 -0300 |
|---|---|---|
| committer | Paulo Alcantara <pc@manguebit.org> | 2026-10-13 00:52:23 -0300 |
| commit | 740619915257969048c5fd3b94c6a55b5be11032 (patch) | |
| tree | dd27a30b44ab300ef8e5190b144678639eed3600 /tools/perf/scripts/python/stackcollapse.py | |
| parent | c38f4e84c9356a9e14d602a00c4afca09c79dc7a (diff) | |
| download | linux-740619915257969048c5fd3b94c6a55b5be11032.tar.gz linux-740619915257969048c5fd3b94c6a55b5be11032.zip | |
smb: client: size smb2_compound_vars.qi_iov for the padding iov
A non-last, non-encrypted compound request gets an 8-byte alignment
padding iov appended by smb2_set_next_command() at rq_iov[rq_nvec].
qi_iov was a single struct kvec used for the query_info link of such
chains, so with rq_nvec == 1 the padding was written into the adjacent
io_iov[0] member.
That write stays within struct smb2_compound_vars and io_iov is unused on
the query_info paths, so it is harmless today, but it relies on field
ordering and would corrupt a live member if that changed. Add
SMB2_QUERY_INFO_IOV_SIZE and make qi_iov an array so the padding iov gets
its own slot.
Signed-off-by: Paulo Alcantara <pc@manguebit.org>
Reviewed-by: David Howells <dhowells@redhat.com>
Cc: Namjae Jeon <linkinjeon@kernel.org>
Cc: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Cc: Shyam Prasad N <sprasad@microsoft.com>
Cc: Tom Talpey <tom@talpey.com>
Cc: Bharath SM <bharathsm@microsoft.com>
Diffstat (limited to 'tools/perf/scripts/python/stackcollapse.py')
0 files changed, 0 insertions, 0 deletions
