On 10/03, Tom Talpey wrote:
On 10/2/2022 11:54 PM, Steve French wrote:
shash was not being initialized in one place in smb3_calc_signature
Suggested-by: Enzo Matsumiya <ematsumiya@xxxxxxx>
Signed-off-by: Steve French <stfrench@xxxxxxxxxxxxx>
I don't see the issue. The shash pointer is initialized in both
arms of the "if (allocate_crypto)" block.
True, but cifs_alloc_hash() returns 0 if *sdesc is not NULL, so
crypto_shash_setkey() got stack garbage as sdesc.
But if you do want to add this, them smb2_calc_signature has
the same code.
True. Steve will you add it to this patch please?
Tom.
Cheers,
Enzo
