[Bug 289647] graphics/drm-61-kmod: X11 with nvidia-drm-61-kmod fails to start after commit ac934345a14f

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=289647

Bjoern A. Zeeb <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[email protected]

--- Comment #4 from Bjoern A. Zeeb <[email protected]> ---
(In reply to Tomoaki AOKI from comment #3)

No driver mapping found for PCI device 0x0000 / 0x0000
   xorg-server-21.1.18/hw/xfree86/dri2/dri2.c::dri2_probe_driver_name()

Without much investigation:
drm_public int drmGetDevice(int fd, drmDevicePtr *device)
{
    return drmGetDevice2(fd, DRM_DEVICE_GET_PCI_REVISION, device);
}
...
drmGetDeviceFromDevId ...

But that path seems to use native FreeBSD PCI ioctl and not LinuxKPI or drm.

I am not sure where it all comes together and how it is supposed to clash.

We do a sscanf from the sysctl in libdrm get_sysctl_pci_bus_info()

So a sysctl -a hw.dri from before the update and after the update would be
really good.
I fear what we all missed is a format change in printing a string (which no one
would assume would be used for anything but logging).  Specifically all the
 hw.dri.%d.busid
values.

I am seeing:
hw.dri.0.busid: pci:0000:00:06.0

which matches "pci:%04x:%02x:%02x.%u"

for:
vgapci0@pci0:0:6:0:     class=0x030000 rev=0x07 hdr=0x00 vendor=0x8086
device=0x5917 subvendor=0x103c subdevice=0x83b2

I'll do comparative builds later myself for i915kms.

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