Bug#991475: unblock: linux/5.10.46-3 (pre-approval)
Hi,
On Sun, Jul 25, 2021 at 04:23:56PM +0200, Salvatore Bonaccorso wrote:
> Hi,
>
> On Sun, Jul 25, 2021 at 03:30:00PM +0200, Salvatore Bonaccorso wrote:
> > On Sun, Jul 25, 2021 at 11:29:23AM +0200, Salvatore Bonaccorso wrote:
> > > Package: release.debian.org
> > > Severity: normal
> > > User: release.debian.org@packages.debian.org
> > > Usertags: unblock
> > > X-Debbugs-Cc: carnil@debian.org,kibi@debian.org
> > > Control: tags -1 + d-i
> > >
> > > Hi Release Team, hi Cyril
> > >
> > > I'm seeking for approval tor another update for linux (hopefullye the
> > > last one) before d-i RC3. It will contain the following changes:
> > >
> > > + * [armhf] Add mdio-aspeed to nic-modules.
> > > + Thanks to Joel Stanley <joel@jms.id.au> (Closes: #991262)
> > >
> > > Support for installing on an ASPEED AST2600 EVB. Installation report
> > > done by Joel Stanley. Cyril supported the inclusion of this change in
> > > https://salsa.debian.org/kernel-team/linux/-/merge_requests/377#note_248751
> > > .
> > >
> > > + * Revert "drm/amdgpu/gfx9: fix the doorbell missing when in CGPG issue."
> > > + (Closes: #990312)
> > > + * Revert "drm/amdgpu/gfx10: enlarge CP_MEC_DOORBELL_RANGE_UPPER to cover full
> > > + doorbell." (Closes: #990312)
> > >
> > > Those two commits were in upstream 5.10.47 and fixes regressions
> > > reported as #990312 (and #991453, #991467). It regressed from the
> > > previous version in testing, which did not exhibit the issue. I have
> > > raised the severities here to RC and think this fixes need to go in
> > > before the bullseye release (they would be pending for the next 5.10.y
> > > based upload, but these seem severe enought to have fixed).
> > >
> > > + * Input: joydev - prevent use of not validated data in JSIOCSBTNMAP ioctl
> > > + (CVE-2021-3612)
> > > + * sctp: validate from_addr_param return (CVE-2021-3655)
> > > + * sctp: add size validation when walking chunks (CVE-2021-3655)
> > >
> > > Pending CVE fixes already in 5.10.y versions upstream.
> > >
> > > + * [s390x] sclp_vt220: fix console name to match device (Closes: #961056)
> > >
> > > Fixes an issue related in #961056, where installer fails (s390x
> > > specific, it as recently addressed by a patch form Valentin Vidić
> > > (vvidic) upstream, merged into 5.10.52, 4.19.198 and 4.9.276.
> > >
> > > Please unblock package linux
> > >
> > > Attached (filtered) debdiff, expluding the ABI reference and generated
> > > rules file. Changelog entry is still set to UNRELEASED.
> >
> > Additionally would like to include another change, the bugfix for
> > https://lists.debian.org/debian-kernel/2021/07/msg00134.html .
>
> And attached the new filtered debdiff.
The upload has happened, including the above bugfix as well and
another followup for the sctp changes[1].
[1]: https://lore.kernel.org/linux-sctp/599e6c1fdcc50f16597380118c9b3b6790241d50.1627439903.git.marcelo.leitner@gmail.com/
All builds arrived in meanwhile and Ansgar poked the signing service
to get in linux-signed-{amd64,arm64,i386} as well.
Discussed with Cyril on IRC, I propose to wait now 24h for little more
exposure in unstable, before letting it migrate to testing, and so
clear the way for d-i RC3.
Regards,
Salvatore
Reply to: