Unhappy Xorg upgrade

[Robert Noland]

On Mon, 2009-02-09 at 19:26 +0000, Bruce M. Simpson wrote:
> Robert,
> 
> First, thanks for all your dedicated work so far on the Xorg ports.
> 
> I realize this upgrade has been somewhat fraught with unexpected issues.
> 
> FWIW, things are not greener on the Linux side of the fence; many Ubuntu 
> and Debian users have reported issues with the newer Xorg and in 
> particular hald.
> 
> Robert Noland wrote:
> > ...
> >> I still see the USB symptoms with xorg-server port as of today -- forced 
> >> rebuild with libpciaccess also. So amd64 is still regressed -- USB is 
> >> totally unusable there after X is started. My theory was that somehow 
> >> Xorg was stomping on the USB controller registers on this machine. The 
> >> USB controller on this box is ALi, card=0x81561043.
> >>     
> >
> > Is your usb sharing interrupts with the video card?
> >   
> 
> Yes, it appears so. This is an ASUS Vintage AH-1, uniprocessor amd64 box 
> w/ioapic enabled
> 
> from devinfo -r (abbreviated):
> ohci0 17
> ohci1 18
> ohci2 19
> ehci0 23
> 
> lspci -v jibes with devinfo -r -- the primary head got IRQ 18, the 
> secondary IRQ 255.
> 
> It appears msk0 is also sharing IRQ 18, though I haven't seen any 
> problems with networking; mskc0, however, is then configured to use MSI 
> (pseudo IRQ 256, 258), it is a PCI-e device.
> 
> When the system starts, the drm module has not been loaded, so the 
> Radeon (Sapphire X550) card hasn't been allocated its IRQ by FreeBSD.
> 
> After X starts, glxinfo and glxgers work fine. kldstat reports drm.ko 
> and radeon.ko got loaded by X as I would expect. I still see no IRQ 
> allocated for the radeon, either in dmesg output or in devinfo -r, 
> however, vmstat -i does show drm0 as sharing IRQ 18.

Ok, that is odd... Once drm is loaded and X opens it, the ddx driver
should request that the irq handler be installed.  At that point, dmesg
should show something resembling the following.

vgapci0: child drm0 requested pci_enable_busmaster
info: [drm] AGP at 0xc0000000 256MB
info: [drm] Initialized i915 1.6.0 20080730
drm0: [ITHREAD]

I have code to enable msi for drm, which has been at least minimally
tested on intel and ati.  FWIW, linux does not support msi on radeon
yet, so I got the jump on them there... The outstanding issue with that
code is that I still need to implement a blacklist for certain devices
that report msi capability, but don't (intel 945gm, is the only known
chip atm).

Does the issue still occur if drm is disabled?

robert.

> At this point, I rebooted and tried manually resetting the BIOS ESCD 
> table, unfortunately the BIOS on this machine won't let me tie IRQs down 
> to particular devices.
> 
> 
> > Does the issue occur if you aren't using a usb mouse?
> >   
> 
> I see the USB problems regardless of the kind of USB devices plugged in, 
> I continue to use a PS/2 mouse on the desktop as a workaround.
> 
> I see the bump on devel/libpciaccess re typo of rombase, and forced a 
> rebuild of xorg-server against the patched libpciaccess library 
> (probably not needed, as the .so ABI didn't change).
> 
> The USB problem is still present, unfortunately.
> 
> thanks,
> BMS
> 
> _______________________________________________
> freebsd-stable at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-stable
> To unsubscribe, send any mail to "freebsd-stable-unsubscribe at freebsd.org"
-- 
Robert Noland <rnoland at FreeBSD.org>
FreeBSD
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: This is a digitally signed message part
Url : http://lists.freebsd.org/pipermail/freebsd-stable/attachments/20090209/57a51a65/attachment.pgp