On Tue, 16 Sep 2025 14:37:57 +0000 (UTC)
"Bjoern A. Zeeb" <b...@freebsd.org> wrote:

> On Tue, 16 Sep 2025, Tomoaki AOKI wrote:
> 
> > Hi.
> >
> > After upgrading graphics/drm-61-kmod to commit
> > 09f1e5c29432d77a992f5f554f48986ef8674e9c, X11 cannot start
> > at least the combination with:
> >
> > graphics/drm-61-kmod
> > graphics/nvidia-drm-61-kmod[-devel]
> > nvidia-driver[-devel]
> >
> > Currently, nvidia drivers of master and -devel is at the same version,
> > as NFB and Beta is older than Production branch of driver.
> >
> > (Note that I'm currently testing 580.82.09, but as it seems to be
> > a minor bugfix release, I think in-tree 580.82.07 should be the same.)
> >
> > Rebuilding x11-servers/xorg-server didn't help.
> >
> > Reverting back to commit 0ec771a40199b9fc4fd671c5c64360ad6d4328fe
> > and rebuilding the above fixed the issue.
> 
> Which FreeBSD version are you exactly using?  The only difference is
> that instead of using special FreeBSD native calls for PCI info we use
> LinuxKPI as in the original (to avoid conflicts).

As noted on Bug 289647, stable/15, amd64 at commit base 1aad58b919d3.


> The one thing I could think of is that something is not set correctly
> in LinuxKPI but when I checked I hadn't noticed anything.
> 
> But if this fails I would have expected it to fail queally for Intel
> and AMD.
> 
> What is consuming the DRM_IOCTL_GET_PCIINFO ioctl anyway?  I checked
> xorg-server and mesa-dri and mesa-libs and couldn't find any use.

To look for nvidia side of codes, tried `fgrep -r -n
"DRM_IOCTL_GET_PCIINFO" 
/usr/ports/graphics/nvidia-drm-61-kmod-devel/work/NVIDIA-FreeBSD-x86_64-580.82.09/`

after both `make extract` and `make`, both showed nothing.

So would be in codes from drm-61-kmod only.


Regards.


> > The actual error found in /ver/log/Xorg.0.log is like below.
> >
> > [   138.316] (II) Initializing extension DRI2
> > [   138.316] (EE) modeset(0): Failed to create pixmap
> > [   138.316] (EE)
> > Fatal server error:
> > [   138.316] (EE) failed to create screen resources(EE)
> >
> > Logs prior to the error seems to be as usual.
> >
> > I'll file a bug on bugzilla later.
> >
> > Regards.
> >
> >
> 
> -- 
> Bjoern A. Zeeb                                                     r15:7


-- 
Tomoaki AOKI    <junch...@dec.sakura.ne.jp>

Reply via email to