Linux 4.9: Reported regressions as of Sunday, 2016-12-04

Linux 4.9: Reported regressions as of Sunday, 2016-12-04

[Thorsten Leemhuis]

Hi! Here is my fifth regression report for Linux 4.9. It lists 11
regressions I'm aware of. 4 of them are new; 6 got fixed since 
the last report -- that was two weeks ago, because I yet again
didn't find any spare time to compile a report last Sunday :-/

As always: Are you aware of any other regressions? Then please let me
know (simply CC regressions@leemhuis.info). And please tell me if there
is anything in the report that shouldn't be there.

Ciao, Thorsten

== Current regressions ==

Desc: System hang up with call trace during doing S3/S4 stress
Repo: 16-12-01 https://bugzilla.kernel.org/show_bug.cgi?id=189421
Stat: n/a 
Note: Brand new

Desc: x86/unwind: Fix guess-unwinder regression // With frame pointers disabled, /proc/<pid>/stack is broken
Repo: 16-11-28 https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1281244.html
Stat: n/a 
Note: Fix heading upstream

Desc: [lkp] [mremap] 5d1904204c: will-it-scale.per_thread_ops -13.1% regression
Repo: 16-11-27 https://www.spinics.net/lists/linux-mm/msg117307.html
Stat: n/a 
Note: Aaron could not reproduce the issue on two of his machines

Desc: sched: fix find_idlest_group for fork/performance regression in hackbench
Repo: 16-11-25 https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1280514.html
Stat: 16-12-04 https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1285936.html
Note: Report contains patch to fix this, testing ongoing

Desc: GPU hang on resume from hibernation
Repo: 16-10-16 https://bugs.freedesktop.org/show_bug.cgi?id=98288#c9 https://bugzilla.kernel.org/show_bug.cgi?id=177701
Stat: n/a 
Note: confusing bug, but seems the issue is still present

Desc: GPU hang on PlaneShift
Repo: 16-10-16 https://bugs.freedesktop.org/show_bug.cgi?id=98922 https://bugzilla.kernel.org/show_bug.cgi?id=177701
Stat: n/a 
Note: confusing bug, but the issue might still be present

Desc: [i.MX6 DRM IPUv3] Regression 4.9-rc5: greenish screen with YUV420 video
Repo: 16-11-17 https://www.spinics.net/lists/kernel/msg2385550.html
Stat: 16-12-02 https://www.spinics.net/lists/kernel/msg2396720.html
Note: "patch available: 3fd8b292ae6b (""drm/imx: ipuv3-plane: merge ipu_plane_atomic_set_base into atomic_update"")"

Desc: "Failed boots/Package drops bisected to 4cd13c21b207 ""softirq: Let ksoftirqd do its job""; "
Repo: 16-11-16 https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1273344.html
Stat: 16-11-25 https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1280379.html
Note: Stalled

Desc: builddeb: fix cross-building to arm64 producing host-arch debs
Repo: 16-11-04 https://www.spinics.net/lists/linux-kbuild/msg13635.html
Stat: 16-11-11 https://www.spinics.net/lists/linux-kbuild/msg13696.html
Note: Nothing happened when Adam pinged Michael in https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1276268.html


== Stalled, waiting for feedback from reporter ==

Desc: 4.9-rc1 boot regression, ambiguous bisect result
Repo: 2016-10-19 https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1253369.html
Stat: 16-10-21 https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1255296.html
Note: Waiting for Dan or someone else to look into this

Desc: Skylake gen6 suspend/resume video regression
Repo: 16-10-16 https://bugzilla.kernel.org/show_bug.cgi?id=177731 https://bugs.freedesktop.org/show_bug.cgi?id=98517
Stat: 16-10-31 https://bugzilla.kernel.org/show_bug.cgi?id=177731#c7
Note: Stalled

== Fixed since last report ==

Desc: module loadling broken due to kbuild changes
Repo: 16-10-15 http://www.gossamer-threads.com/lists/linux/kernel/2544734 and various other threads (https://lwn.net/Articles/707520/ )
Fix:  https://git.kernel.org/torvalds/c/cd3caefb4663e3811d37cc2afad3cce642d60061 https://git.kernel.org/torvalds/c/faaae2a581435f32781a105dda3501df388fddcb (among others)

Desc: "irq 16: nobody cared (try booting with the ""irqpoll"" option) since t0b9e2988ab226 (ahci: use pci_alloc_irq_vectors)"
Repo: 16-11-19 https://bugzilla.kernel.org/show_bug.cgi?id=188181
Stat: 16-12-03 https://bugzilla.kernel.org/show_bug.cgi?id=188181#c9
Note: Fixed according to reporter (might be thx to https://git.kernel.org/torvalds/c/6929ef385e09c0065b87fda3e7b872a5070ac783 )

Desc: "build regression: make.cross ARCH=mips fails with ""No rule to make target 'alchemy/devboards/'. """
Repo: 16-10-30 https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1262410.html https://marc.info/?l=linux-kernel&m=147780880425626
Fix:  https://git.kernel.org/torvalds/c/818f38c5b7c4482abd71c64ac4d49911fbefaf9e

Desc: "oops due to 493b2ed3f760 (""crypto: algif_hash - Handle NULL hashes correctly"")"
Repo: 16-11-17 https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1273867.html
Fix:  https://git.kernel.org/torvalds/c/a8348bca2944d397a528772f5c0ccb47a8b58af4

Desc: MSI is no longer enabled for many/most Intel SATA controllers in 4.9
Repo: 16-11-16 https://bugzilla.kernel.org/show_bug.cgi?id=187821
Fix:  https://git.kernel.org/torvalds/c/6929ef385e09c0065b87fda3e7b872a5070ac783

Desc: qla2xxx: do not abort all commands in the adapter during EEH recovery
Repo: 16-11-14 https://www.mail-archive.com/linux-scsi@vger.kernel.org/msg55186.html
Fix:  https://git.kernel.org/torvalds/c/c733ab3512431436a26e0381829b45794cb13fb0

Re: Linux 4.9: Reported regressions as of Sunday, 2016-12-04

[Adam Borowski]

On Sun, Dec 04, 2016 at 01:26:05PM +0100, Thorsten Leemhuis wrote:
> Desc: builddeb: fix cross-building to arm64 producing host-arch debs
> Repo: 16-11-04 https://www.spinics.net/lists/linux-kbuild/msg13635.html
> Stat: 16-11-11 https://www.spinics.net/lists/linux-kbuild/msg13696.html
> Note: Nothing happened when Adam pinged Michael in https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1276268.html

The fix has been accepted and is sitting in kbuild/rc-fixes; I guess Michal
wanted to wait for something else to pop up to not bother Linus with
excessive one-patch pulls.  And we had the modversions brouchacha that was a
far more important regression in the kbuild land.

This cross-building fail is not that vital: arm64 has nothing in between:
* small SoCs with custom (and thus unpackaged) u-boot setups, it's less work
  to configure to boot Image rather than vmlinuz-$KVER, and they're so
  fragile their u-boot breaks if you as much as look at it funny
* fat 48-way servers that you're likely to cross-build on rather than for

I'm insisting on fixing this in 4.9 rather than 4.10 only because it's LTS
both for kernel and for Debian, thus people are going to use *deb-pkg.

So as long as the queue is flushed within this week, all is fine.


Meow!
-- 
u-boot problems can be solved with the help of your old SCSI manuals, the
parts that deal with goat termination.  You need a black-handled knife, and
an appropriate set of candles (number and color matters).  Or was it a
silver-handled knife?  Crap, need to look that up.