regressions report as of Linux 7.1-rc6 post (was: Linux 7.1-rc6)

On 6/1/26 00:36, Linus Torvalds wrote:
> Well, I wouldn't call this "small", but it is certainly smaller than
> rc5 was. And I don't think there's anything particularly scary here,
> so maybe we're still on track for a normal release cycle. Let's see.

Hi Linus! Still experimenting a bit with a new format for the regression
report. To keep the work manageable, I chose to trim things down and
skip a few things:
* regression fixes sitting in a pending-fixes branch that most likely
will head to mainline before the next rc, as the subsystem maintainers
usually sent a weekly PR right before a new rc.
* unresolved regressions for older cycles where no ad-hoc fix is
available (e.g. neither a fix nor a revert), so there is not much we can
do about those
* fixes for regressions recently found in releases more than about one
year ago

Here we go:

#######################################################################
7.1 regressions afaics unlikely to be fixed for -rc7, unless you chose
to intervene.

---
cpufreq/amd-pstate: switching to performance profile broke
https://bugzilla.kernel.org/show_bug.cgi?id=221505

Known for 4 weeks days and fix is available, but Rafael is "unable pull
it this week":
https://lore.kernel.org/all/[email protected]/
https://lore.kernel.org/all/CAJZ5v0ghALXyU8tf0gcT+qirhde7ah8_rQ9+FryuArfbLTe01A@mail.gmail.com/

Side note: there is a related problem that developers are still looking
into:
https://bugzilla.kernel.org/show_bug.cgi?id=221473
https://lore.kernel.org/lkml/[email protected]/

#######################################################################
Regressions pre-7.0 afaics unlikely to be fixed for -rc7, unless you
chose to intervene.

---
rtl8xxxu: RTL8188FTV (0bda:f179) fails to authenticate/configure since 6.19
https://bugzilla.kernel.org/show_bug.cgi?id=221394
https://lore.kernel.org/linux-wireless/[email protected]/

Known for 6 weeks. A fix was sitting in a sub-sub-subsystem tree for a
while and recently turned up in in next as ef771eabc79d5f ("wifi:
rtl8xxxu: Detect the maximum supported channel width") [next-20260604],
but from the outside it looks a bit like it's still only on track for
merging in the next cycle:
https://lore.kernel.org/linux-wireless/[email protected]/

#######################################################################
7.1 regressions unlikely to be fixed for -rc7

---
loop: loop drive broken due to some race exposed by an unknown change,
detected by syzbot
https://lore.kernel.org/lkml/[email protected]/

Known since 7 weeks, culprit unknown. People were really busy trying to
get this resolved, but activity slowed down in the past few days without
a fix in sight (or I missed something).

---
drm: amdgpu: handling FreeSync via MCCS broke VRR
https://gitlab.freedesktop.org/drm/amd/-/issues/5286
https://lore.kernel.org/amd-gfx/[email protected]/

Known for 24 days and a fix for one of the problems is heading towards
mainline after some prodding now, but seems there are more problems that
might remain unresolved -- but there is the plan to consider a revert
next week:
https://lore.kernel.org/amd-gfx/[email protected]/
https://lore.kernel.org/all/CADnq5_McpRJzu1_UttS2NjnX5GRpTZG7H2oKpQ44yrQo3EOX%[email protected]/


#######################################################################
7.1 regressions that might or might not be fixed for -rc7

---
HID: input: Oops during boot
https://lore.kernel.org/lkml/[email protected]/
https://lore.kernel.org/lkml/[email protected]/

Fix is 426e5846eba75f ("HID: Input: Add battery list cleanup with devm
action") [next-20260602] and might make it to mainline if the HID folks
send a PR this week.


#######################################################################

Ciao, Thorsten