APPLIED/Cmnt: [SRU Bionic/Focal/OEM-5.14/Jammy/HWE-5.17/Kinetic 0/1] CVE-2022-40768

[Stefan Bader]

On 03.10.22 17:31, Cengiz Can wrote:
> [Impact]
> drivers/scsi/stex.c in the Linux kernel through 5.19.9 allows local
> users to obtain sensitive information from kernel memory because
> stex_queuecommand_lck lacks a memset for the PASSTHRU_CMD case.
> 
> [Fix]
> Fix is currently on linux-next.git only. There's a request to convert
> it into two separate patches which won't effect the structure of the
> fix.
> 
> [Test case]
> For each target kernel, compile and boot tested on KVM only.
> 
> [Potential regression]
> stex driver specific changes are harmless but the added const'ness to
> scsi_sg_copy_from_buffer signature might be risky.
> 
> Linus Torvalds (1):
>    scsi: stex: Properly zero out the passthrough command structure
> 
>   drivers/scsi/stex.c      | 17 +++++++++--------
>   include/scsi/scsi_cmnd.h |  2 +-
>   2 files changed, 10 insertions(+), 9 deletions(-)
> 

This is one of those examples where you should not have trusted git cherry pick. 
The submitted patch applied only to focal and bionic (-C2) but not 
jammy:linux,linux-hwe-5.17. For those the upstream version would have been 
better because the fuzz was more than can be ignore with -C. I double checked 
the result looked good there when using patch and allow all fuzz.

Applied to jammy,focal,bionic:linux/master-next and 
jammy:linux-hwe-5.17/hwe-5.17-next. Thanks.

-Stefan

-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20221005/ed00c540/attachment.sig>