kernel BUG at arch/x86/include/asm/xen/hypervisor.h:86! Oops: invalid opcode: 0000 #9552
Labels
affects-4.2
This issue affects Qubes OS 4.2.
C: kernel
hardware support
P: default
Priority: default. Default priority for new issues, to be replaced given sufficient information.
R: self-closed
Voluntarily closed by the person who opened it before another resolution occurred.
T: bug
Type: bug report. A problem or defect resulting in unintended behavior in something that exists.
Comments
biergaizi
added
P: default
FWIW This is
And given this, I'm afraid you won't get much support from the i915 driver maintainers. But just an idea - maybe you'll have better luck with the new "xe" driver? See https://www.kernel.org/doc/html//next/gpu/rfc/xe.html#xe-platforms |
Both drivers used to be functional - by functional I mean at least the system boots up. But since the last (?) update, booting with |
|
I don't see much changes in Xen-related code (arch/x86/xen and drivers/xen dirs) between 6.10 and 6.11... The only remotely relevant change I see is about multicall debugging, which should be disabled by default. |
andrewdavidwong
added
C: kernel
hardware support
needs diagnosis
|
I switched the GPU to a Nvidia Quadro NVS 295, and the system still crashes multiple times an hour. This time it crashes in an interrupt handling routine, and I'm not even able to obtain the full log before the machine restarts. So my best guess is that it's either a Xen regression or defective hardware. I'll try ruling out the latter possibility by stress testing the system. |
|
Hmmmm, Perhaps my PCIe slot is broken? |
|
Got another different one, at least this trace is more complete. Looks like a PCIe write error, difficult to say if it's a hardware problem or Xen... |
|
After replacing the GPU again with an AMD Radeon RX580 from my desk drawer, the system seems to run again without problems. Thus, without better evidence, I have to conclude that the Nvidia crashes I was seeing was an unrelated and isolated hardware problem unrelated to Xen. Since the Intel DG1 is not supported by the upstream, it's unlikely that anyone from the upstream will ever attempt to debug it, so I'm closing this issue for now. |
andrewdavidwong
added
R: self-closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Qubes OS release
Brief summary
Since a week ago, I started seeing repeated Dom0 crashes once an hour. I believe it's the result of an upgrade, but I usually use the machine for months without rebooting, unfortunately it's quite hard to pinpoint the exact version... Using a serial port, I am able to capture the following log, attached at the bottom.
Steps to reproduce
i915, so perhaps the problem was GPU related (or not, can't be quite sure). This machine does have an unusual GPU, the Intel DG1, which is Intel's first modern dedicated GPU. Its support is disabled in Linux by default, and must be enabled viai915.force_probe='4908'. But this crash didn't initially occur for at least a week after the GPU has been installed, so I originally assumed it's at least functional. Whatever the reason, it's best to write it down for the record.Expected behavior
Actual behavior
The text was updated successfully, but these errors were encountered: