ACK/Cmnt: [SRU][v3][F][J][L][M][PATCH 0/1] net: Avoid address overwrite in kernel_connect
Stefan Bader
stefan.bader at canonical.com
Wed Sep 13 07:59:43 UTC 2023
On 12.09.23 21:49, Khalid Elmously wrote:
> BugLink: https://bugs.launchpad.net/bugs/2035163
>
> An upstream fix is requested to resolve a use-case where sockaddr gets re-written which causes NFS-breakage.
>
>
> Testing:
> - GCP has confirmed that the fix works before proposing it upstream. I have verified basic network sanity with the fix applied. I will also ask them to confirm the fix with the kernel in -proposed.
>
>
> Regression potenial:
> - The fix modifies kernel_connect() which can have an effect on all kinds of network connections. The change itself is very minor and only converts a pass-by-reference to a pass-by-value - so the risk is considered minimal.
>
>
>
> v2:
> - Added nomination for J/gke as well
>
> v3:
> - nominating for all generic kernels
>
>
> Jordan Rife (1):
> net: Avoid address overwrite in kernel_connect
>
> net/socket.c | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
The READ_ONCE() which was worked around was added in v6.6 when I checked
correctly. That was done to avoid problematic compiler optimizations.
Mentioning this to set awareness that this might lead to different
results than upstream and to watch out for possible socket issues in the
cycle this gets applied to.
Also, would this not be relevant for the various OEM kernels as well?
Acked-by: Stefan Bader <stefan.bader at canonical.com>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_0xE8675DEECBEECEA3.asc
Type: application/pgp-keys
Size: 44613 bytes
Desc: OpenPGP public key
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20230913/162de914/attachment-0001.key>
-------------- 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/20230913/162de914/attachment-0001.sig>
More information about the kernel-team
mailing list