Discussion:
Does intel(4) support Iris Xe Graphics?
Michel von Behr
2021-04-06 07:09:07 UTC
Permalink
Hi - (not a dev, just trying to use OpenBSD snapshot) whenever I try to
launch Xorg, either via xenodm or startx, I'm getting a kernel panic,
like "pool_do_get:
drmobj : page empty" (I already sent an e-mail [1] to ***@openbsd.org with
dmesg and all).

I'm wondering if the problem could be with my video card, Intel Iris Xe?
Even though dmesg shows that is was detected and should (?) be working. But
I can't find a reason why my laptop would not run Xorg.

inteldrm0 at pci0 dev 2 function 0 "Intel Xe Graphics" rev 0x01
drm0 at inteldrm0
inteldrm0: msi, TIGERLAKE, gen 12


Any pointing to the right direction would be appreciated. (If this problem
relates to Xorg specifically and not to OpenBSD please let me know).

[1] https://marc.info/?l=openbsd-bugs&m=161754767328009&w=2

Regards,

Michel
Jonathan Gray
2021-04-06 22:33:46 UTC
Permalink
Post by Michel von Behr
Hi - (not a dev, just trying to use OpenBSD snapshot) whenever I try to
launch Xorg, either via xenodm or startx, I'm getting a kernel panic,
dmesg and all).
The pool should already be initialised via
i915_global_objects_init()
i915_globals_init()
inteldrm_attachhook()
Post by Michel von Behr
I'm wondering if the problem could be with my video card, Intel Iris Xe?
Even though dmesg shows that is was detected and should (?) be working. But
I can't find a reason why my laptop would not run Xorg.
inteldrm0 at pci0 dev 2 function 0 "Intel Xe Graphics" rev 0x01
drm0 at inteldrm0
inteldrm0: msi, TIGERLAKE, gen 12
jcs@ has/had a tiger lake machine which could run Xorg with the
linux 5.7 based drm in -current. I'm not sure what is different here.
Post by Michel von Behr
Any pointing to the right direction would be appreciated. (If this problem
relates to Xorg specifically and not to OpenBSD please let me know).
[1] https://marc.info/?l=openbsd-bugs&m=161754767328009&w=2
Regards,
Michel
Michel von Behr
2021-04-07 04:07:26 UTC
Permalink
Thank you for the reply, Jonathan - FWIW I was able to run Ubuntu on the
machine just now. I still would like to try and install OpenBSD, if anyone
can help me diagnose/fix the problem I’m willing to try.

Regards,

Michel
Post by Jonathan Gray
Post by Michel von Behr
Hi - (not a dev, just trying to use OpenBSD snapshot) whenever I try to
launch Xorg, either via xenodm or startx, I'm getting a kernel panic,
with
Post by Michel von Behr
dmesg and all).
The pool should already be initialised via
i915_global_objects_init()
i915_globals_init()
inteldrm_attachhook()
Post by Michel von Behr
I'm wondering if the problem could be with my video card, Intel Iris Xe?
Even though dmesg shows that is was detected and should (?) be working.
But
Post by Michel von Behr
I can't find a reason why my laptop would not run Xorg.
inteldrm0 at pci0 dev 2 function 0 "Intel Xe Graphics" rev 0x01
drm0 at inteldrm0
inteldrm0: msi, TIGERLAKE, gen 12
linux 5.7 based drm in -current. I'm not sure what is different here.
Post by Michel von Behr
Any pointing to the right direction would be appreciated. (If this
problem
Post by Michel von Behr
relates to Xorg specifically and not to OpenBSD please let me know).
[1] https://marc.info/?l=openbsd-bugs&m=161754767328009&w=2
Regards,
Michel
Tom Smyth
2021-04-07 10:33:29 UTC
Permalink
Hi Michel,
if you send the dmesg from OpenBSD when it is installed and Ubuntu
it would help alot
see the hardware that your box is running (and the hardware as
detected by OpenBSD / Ubuntu
Post by Michel von Behr
Thank you for the reply, Jonathan - FWIW I was able to run Ubuntu on the
machine just now. I still would like to try and install OpenBSD, if anyone
can help me diagnose/fix the problem I’m willing to try.
Regards,
Michel
Post by Jonathan Gray
Post by Michel von Behr
Hi - (not a dev, just trying to use OpenBSD snapshot) whenever I try to
launch Xorg, either via xenodm or startx, I'm getting a kernel panic,
with
Post by Michel von Behr
dmesg and all).
The pool should already be initialised via
i915_global_objects_init()
i915_globals_init()
inteldrm_attachhook()
Post by Michel von Behr
I'm wondering if the problem could be with my video card, Intel Iris Xe?
Even though dmesg shows that is was detected and should (?) be working.
But
Post by Michel von Behr
I can't find a reason why my laptop would not run Xorg.
inteldrm0 at pci0 dev 2 function 0 "Intel Xe Graphics" rev 0x01
drm0 at inteldrm0
inteldrm0: msi, TIGERLAKE, gen 12
linux 5.7 based drm in -current. I'm not sure what is different here.
Post by Michel von Behr
Any pointing to the right direction would be appreciated. (If this
problem
Post by Michel von Behr
relates to Xorg specifically and not to OpenBSD please let me know).
[1] https://marc.info/?l=openbsd-bugs&m=161754767328009&w=2
Regards,
Michel
--
Kindest regards,
Tom Smyth.
Tom Smyth
2021-04-07 10:34:54 UTC
Permalink
Try Current and 6.8 and see if you get a different result in each..
dmesgs are key for getting help on this type of query ...
Post by Tom Smyth
Hi Michel,
if you send the dmesg from OpenBSD when it is installed and Ubuntu
it would help alot
see the hardware that your box is running (and the hardware as
detected by OpenBSD / Ubuntu
Post by Michel von Behr
Thank you for the reply, Jonathan - FWIW I was able to run Ubuntu on the
machine just now. I still would like to try and install OpenBSD, if anyone
can help me diagnose/fix the problem I’m willing to try.
Regards,
Michel
Post by Jonathan Gray
Post by Michel von Behr
Hi - (not a dev, just trying to use OpenBSD snapshot) whenever I try to
launch Xorg, either via xenodm or startx, I'm getting a kernel panic,
with
Post by Michel von Behr
dmesg and all).
The pool should already be initialised via
i915_global_objects_init()
i915_globals_init()
inteldrm_attachhook()
Post by Michel von Behr
I'm wondering if the problem could be with my video card, Intel Iris Xe?
Even though dmesg shows that is was detected and should (?) be working.
But
Post by Michel von Behr
I can't find a reason why my laptop would not run Xorg.
inteldrm0 at pci0 dev 2 function 0 "Intel Xe Graphics" rev 0x01
drm0 at inteldrm0
inteldrm0: msi, TIGERLAKE, gen 12
linux 5.7 based drm in -current. I'm not sure what is different here.
Post by Michel von Behr
Any pointing to the right direction would be appreciated. (If this
problem
Post by Michel von Behr
relates to Xorg specifically and not to OpenBSD please let me know).
[1] https://marc.info/?l=openbsd-bugs&m=161754767328009&w=2
Regards,
Michel
--
Kindest regards,
Tom Smyth.
--
Kindest regards,
Tom Smyth.
Jonathan Gray
2021-04-07 10:56:32 UTC
Permalink
Post by Tom Smyth
Try Current and 6.8 and see if you get a different result in each..
dmesgs are key for getting help on this type of query ...
There is a snapshot dmesg in the bug report. I don't see a benefit to
6.8 or linux dmesgs.
Tom Smyth
2021-04-07 11:04:51 UTC
Permalink
Hi Jonathan,
sorry missed the Bug Report... with the dmesg.
I thought the linux dnmesg where the hardware was working would
have been useful if there was an issue with the hardware detection.
I suggested trying 6.8 incase there was a bug introduced in current
and would give a baseline...
suggested current as the bug might already be fixed in current as
opposed to release ...
Post by Jonathan Gray
Post by Tom Smyth
Try Current and 6.8 and see if you get a different result in each..
dmesgs are key for getting help on this type of query ...
There is a snapshot dmesg in the bug report. I don't see a benefit to
6.8 or linux dmesgs.
--
Kindest regards,
Tom Smyth.
Eric Auge
2021-04-26 10:02:27 UTC
Permalink
Hello,

Experiencing something similar with a similar machine: x1 carbon gen 9 with
an Intel Iris Xe.
I am running -current/amd64 from April 25th on this machine.
No kernel panics here, but almost impossible to use Xorg, it hangs and I
have to restart it all (Ctrl + Alt + Backspace),
it behaves like this more or less randomly, but happens within 5 minutes of
a valid session (after login to xenodm, opening terminals, browser, etc..)
the following messages are in the kernel messages/logs:

drm:pid593:intel_cpu_fifo_underrun_irq_handler *ERROR* [drm] *ERROR* CPU
pipe A FIFO underrun
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
i915_vma_coredump_create: stub
pool_fini: stub
drm:pid23016:intel_engine_reset *NOTICE* Resetting rcs0 for stopped
heartbeat on rcs0
drm:pid23016:mark_guilty *NOTICE* Xorg[86272] context reset due to GPU hang

I've attached a dmesg, just wanted to comment on this report as it "seems"
similar.
About to properly send this to bugs@ in a separate email (with acpidump,
pcidump, blablabla sendbug(1)).

Regards,
Eric.
Post by Michel von Behr
Hi - (not a dev, just trying to use OpenBSD snapshot) whenever I try to
launch Xorg, either via xenodm or startx, I'm getting a kernel panic,
dmesg and all).
I'm wondering if the problem could be with my video card, Intel Iris Xe?
Even though dmesg shows that is was detected and should (?) be working. But
I can't find a reason why my laptop would not run Xorg.
inteldrm0 at pci0 dev 2 function 0 "Intel Xe Graphics" rev 0x01
drm0 at inteldrm0
inteldrm0: msi, TIGERLAKE, gen 12
Any pointing to the right direction would be appreciated. (If this problem
relates to Xorg specifically and not to OpenBSD please let me know).
[1] https://marc.info/?l=openbsd-bugs&m=161754767328009&w=2
Regards,
Michel
Loading...