[Bug 98386] [NVE7] bus: MMIO write of FAULT at [ IBUS ], Pointer to {TDMS, flat panel) table invalid

[bugzilla-daemon-CC+yJ3UmIYqDUpFQwHEjaQ@public.gmane.org]

[-- Attachment #1.1: Type: text/plain, Size: 1506 bytes --]

https://bugs.freedesktop.org/show_bug.cgi?id=98386

--- Comment #1 from Ilia Mirkin <imirkin-FrUbXkNCsVf2fBVCVOL8/A@public.gmane.org> ---
(In reply to bruno.pagani from comment #0)
> [  454.919271] nouveau 0000:02:00.0: bus: MMIO write of 00000002 FAULT at
> 4188ac [ IBUS ]

That's bad.

4188ac = PGRAPH.GPC_BROADCAST.FFB.PART_CONFIG

I believe we do this here:

drm/nouveau/nvkm/engine/gr/gk104.c:     nvkm_wr32(device, GPC_BCAST(0x08ac),
nvkm_rd32(device, 0x100800));

> [  454.989954] nouveau 0000:02:00.0: DRM: Pointer to TMDS table invalid
> [  454.989957] nouveau 0000:02:00.0: DRM: Pointer to flat panel table invalid

These are (semi-)expected. Should probably just make them no longer errors.

> [  470.245180] nouveau 0000:02:00.0: bus: MMIO write of ffffff1f FAULT at
> 6013d4 [ IBUS ]

That's bad too. This is related to the legacy 0x3d4 VGA IO port (which still
has to be used for various display-related things). Unless you've skimped on
the dmesg, you don't have any DCB entries and it may be reasonable to believe
you don't have a display block at all. (I assume there are no outputs wired up
to the NVIDIA GPU?) In that case, we should probably track down where we're
trying to do this from and skip it. Based on the location in the logs, probably
something VBIOS- or other early-init-related.

Unfortunately I don't have any actual debugging steps for you. Perhaps Ben
will.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[-- Attachment #1.2: Type: text/html, Size: 2477 bytes --]

[-- Attachment #2: Type: text/plain, Size: 154 bytes --]

_______________________________________________
Nouveau mailing list
Nouveau@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/nouveau