Bug#1112627: linux-image-6.16.3+deb14-amd64: Intel audio no longer works: DMAR: [DMA Write NO_PASID] Request device [00:1b.0] ... non-zero reserved fields in PTE
Control: tags -1 + moreinfo upstream
On Sun, Aug 31, 2025 at 02:43:30PM +0200, Francesco Poli (wintermute) wrote:
> Package: src:linux
> Version: 6.16.3-1
> Severity: important
> X-Debbugs-Cc: debian-amd64@lists.debian.org, invernomuto@paranoici.org
> User: debian-amd64@lists.debian.org
> Usertags: amd64
>
> Hello!
>
> I've just upgraded linux-image-amd64 and rebooted to
> linux-image-6.16.3+deb14-amd64, to find a very bad surprise:
> Intel integrated audio (device 00:1b.0, see the PCI list) no longer works.
>
> Everything looks normal: alsamixer shows the usual controls (nothing relevant
> was found to be accidentally muted) for the usual sound card (HDA Intel PCH);
> jackd starts as usual; audacious starts as usual and plays music as usual.
>
> But...
>
> But nothing can be heard from the speakers and /var/log/kern.log is
> filled with the following error messages (that repeat once every 5 s):
>
> kernel: dmar_fault: 19382 callbacks suppressed
> kernel: DMAR: DRHD: handling fault status reg 3
> kernel: DMAR: [DMA Write NO_PASID] Request device [00:1b.0] fault addr 0xffa02000 [fault reason 0x0c] non-zero reserved fields in PTE
> kernel: DMAR: DRHD: handling fault status reg 3
> kernel: DMAR: [DMA Write NO_PASID] Request device [00:1b.0] fault addr 0xffa02000 [fault reason 0x0c] non-zero reserved fields in PTE
> kernel: DMAR: DRHD: handling fault status reg 3
> kernel: DMAR: [DMA Write NO_PASID] Request device [00:1b.0] fault addr 0xffa02000 [fault reason 0x0c] non-zero reserved fields in PTE
> kernel: DMAR: DRHD: handling fault status reg 3
>
>
> Rebooting with linux-image-6.12.38+deb13-amd64 makes everything work
> again.
>
> The error messages look similar to the ones quoted in a comment to
> one [upstream bug] report, however that comment refers to Linux kernel
> version 6.12.23, while I only see this issue after upgrading from
> version 6.12.38 to version 6.16.3 ...
>
> [upstream bug]: <https://bugzilla.kernel.org/show_bug.cgi?id=215919#c6>
>
>
> What am I doing wrong?
> Please forward my bug report upstream and incorporate a fix as soon as
> possible.
>
> Thanks for any help you may provide!
So while this is IOMMU related, it *still* might be broken firmware
and you can try if disabling IOMMU "resolves" the issue. Still there
is indication that this might be a real regression from 6.12 to 6.16.
So additionally to the above tests I would like to ask you to do the
following:
Between your last "good" Debian revision (6.12.38-1) and your first
noticed "bad' revison (6.16.3-1) there were a couple of experimental
uploads:
6.16.3-1
6.16.1-1~exp1
6.16-1~exp1
6.16~rc7-1~exp1
6.15.6-1~exp1
6.15.5-1~exp1
6.15.4-1~exp1
6.15.3-1~exp1
6.15.2-1~exp1
6.15.1-1~exp1
6.15-1~exp1
6.15~rc7-1~exp1
6.14.6-1~exp1
6.14.5-1~exp1
6.14.3-1~exp1
6.13.11-1~exp1
6.13.10-1~exp1
6.13.9-1~exp1
6.13.8-1~exp1
6.13.7-1~exp1
6.13.6-1~exp1
6.13.5-1~exp1
6.13.4-1~exp1
6.13.3-1~exp1
6.13.2-1~exp1
6.13~rc7-1~exp1
6.13~rc6-1~exp1
6.12.43-1
6.12.41-1
6.12.38-1
Via snapshot.debian.org service, do a manual search of
linux-image-amd64 package an determine a close a possible range of
still "good" and first "bad" kernel revision. I would start here to
test first the major upstream bumps as we do not know if the potential
rgression was as well introduced backported in the back then stable
series. So you might wnant to test first 6.13~rc6-1~exp1, test,
depending on the result go up to the 6.14.y versions, test, if
behavoour changes, then search between the 6.13 and 6.14.y versions
oteherwise move the 6.15.y versions and search similarly.
Let's say you find hipotetically now you found that all up to
6.14.6-1~exp1 were good and breaks with testing 6.15~rc7-1~exp1.
(Note if the situation is not very clear and you jump between results
while moving, then it's better to directly bisect v6.12 to 6.16.3)
But let's assume we have all up to 6.14.6-1~exp1 are good and
6.15~rc7-1~exp1 is bad.
It would be great if you could bisect the problem. That would involve
compiling and testing a few kernels:
git clone ttps://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git
cd linux-stable
git checkout v6.14
cp /boot/config-$(uname -r) .config
yes '' | make localmodconfig
make savedefconfig
mv defconfig arch/x86/configs/my_defconfig
# test 6.14 to ensure this is "good"
make my_defconfig
make -j $(nproc) bindeb-pkg
... install the resulting .deb package and confirm it successfully boots / problem does not exist
# test 6.15-rc7 to ensure this is "bad"
git checkout v6.15-rc7
make my_defconfig
make -j $(nproc) bindeb-pkg
... install the resulting .deb package and confirm it fails to boot / problem exists
With that confirmed, the bisection can start:
git bisect start
git bisect good v6.14
git bisect bad v6.15-rc7
In each bisection step git checks out a state between the oldest
known-bad and the newest known-good commit. In each step test using:
make my_defconfig
make -j $(nproc) bindeb-pkg
... install, try to boot / verify if problem exists
and if the problem is hit run:
git bisect bad
and if the problem doesn't trigger run:
git bisect good
. Please pay attention to always select the just built kernel for
booting, it won't always be the default kernel picked up by grub.
Iterate until git announces to have identified the first bad commit.
Then provide the output of
git bisect log
In the course of the bisection you might have to uninstall previous
kernels again to not exhaust the disk space in /boot. Also in the end
uninstall all self-built kernels again.
Once we have a clear bad commit we might we can douple check. If the
problem seems clear we can take it from here, if not we might ask to
you to send now the report to upstream (if needed helping how to
determine the correct recipients, ideally keeping our bug in the
loop).
Let's see how this goes. Let us know if you struggle with some steps.
Again, if the first search via Debian packages is unhelpful do a git
bisect directly with upstream version, then it the bigger range of
versions you have from your initial report.
Regards,
Salvatore
Reply to: