[Bug 267028] kernel panics when booting with both (zfs,ko or vboxnetflt,ko or acpi_wmi.ko) and amdgpu.ko

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=267028

--- Comment #193 from Mark Millard <marklmi26-fbsd@yahoo.com> ---
(In reply to George Mitchell from comment #192)

Viewed at actual size or zoomed in on an iPad I was able
to read some of the text in:

https://www.m5p.com/public/george/267028/IMG_20241210_093239744_HDR.jpg

. . .
vgapci0: child drmn0 requested pci_get_powerstate
sysctl_warn_reuse: can't re-use a leaf (hw.dri.debug)!
<6>[drm] Initialized amdgpu 3.40.0 20150101 for drmn0 on minor 0

It panicked here instead of getting to the point of showing what
would normally be next: "Autoloading module: acpi_wmi"

However, the <6> is also not normal and may well be sigificant.

The panic was something like:

Fatal trap 9: general protection fault while in kernel mode
cpuid = 1: apic id = 01
instruction pointer = 0x20:0xffffffff00cf0110
. . .
current process = 25 (kldload)
trap number = 9

(I doubt that I could tell c vs. e or 0 vs. 8 .)

In the presence of the panic context I do not expect the drm_modeset
locking is able to operate correctly. In other words: I expect that
the !drm_modeset_is_locked notices are expected, given the Fatal
trap 9 and its being handling in the kernel. That in turn messes
things up even more for later --or so I guess.

I expect that the later trap 22 (in the other picture) is from the
lack of handling during the attempt to present the trap 9 information.

The other picture's backtrace suggests for the instruction
pointer reported above (general range of upper address bits):

instruction pointer = 0x2?:0xffffffff80?f?11?

where the "?"'s are about the 0 vs. 8 and c vs. e question.

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