Re: snapshot 20050610 [i810]

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 454-5900

Hi Alan,

> You've picked the wrong snapshot for the i915GM chip.
>=20
> You need i915-20050610 instead of i810-20050610.

thanks for the hint - I must have missed the information which driver
is for which card...:-(

Apologies!

Nevertheless, there remain problems. To begin with, I've re-installed
the common-20050610 and i915-20050610 things over a freshly installed
X.org 6.8.2, as provided by the Mandrake 2005 Limited Edition; during
boot, 915 resolution sets the 1280 x something mode to 1400 x
1050. With the plain i810 driver for the 6.8.2 server the card wasn't
recognised, and I think that's well understood because the i915 wasn't
supported at the time.

After installing the common and i915 packages, the X server does
recognise the chip and comes up in the desired resolution. Very nice!
I still need to specify i810 as driver in the xorg.conf, but I guess
that's intentional(?)

Things go wrong when starting an OpenGL program. I've attached the
xorg.conf, but with the dri module commented out (see below why). To
begin with, glxinfo says (among other things)

   direct rendering: Yes

if dri is included in the xorg.conf; but immediately after glxinfo,
I get a kernel message saying

   ... kernel: Disabling IRQ #16

From /var/log/messages:

   Jun 16 23:09:39 localhost kernel: irq 16: nobody cared!
   Jun 16 23:09:39 localhost kernel:  [dump_stack+30/32] dump_stack+0x1e/0x=
20
   Jun 16 23:09:39 localhost kernel:  [<c0103c8e>] dump_stack+0x1e/0x20
   Jun 16 23:09:39 localhost kernel:  [__report_bad_irq+43/144]
__report_bad_irq+0x2b/0x90
   Jun 16 23:09:39 localhost kernel:  [<c019117b>] __report_bad_irq+0x2b/0x=
90
   Jun 16 23:09:39 localhost kernel:  [note_interrupt+144/176]
note_interrupt+0x90/0xb0
   Jun 16 23:09:39 localhost kernel:  [<c01912a0>] note_interrupt+0x90/0xb0
   Jun 16 23:09:39 localhost kernel:  [__do_IRQ+261/272] __do_IRQ+0x105/0x1=
10
   Jun 16 23:09:39 localhost kernel:  [<c0190d55>] __do_IRQ+0x105/0x110
   Jun 16 23:09:39 localhost kernel:  [do_IRQ+28/48] do_IRQ+0x1c/0x30
   Jun 16 23:09:39 localhost kernel:  [<c010519c>] do_IRQ+0x1c/0x30
   Jun 16 23:09:39 localhost kernel:  [common_interrupt+26/32]
common_interrupt+0x1a/0x20
   Jun 16 23:09:39 localhost kernel:  [<c0103822>] common_interrupt+0x1a/0x=
20
   Jun 16 23:09:39 localhost kernel:  [cpu_idle+68/96] cpu_idle+0x44/0x60
   Jun 16 23:09:39 localhost kernel:  [<c01010f4>] cpu_idle+0x44/0x60
   Jun 16 23:09:39 localhost kernel:  [start_kernel+365/432]
start_kernel+0x16d/0x1b0
   Jun 16 23:09:39 localhost kernel:  [<c03ec89d>] start_kernel+0x16d/0x1b0
   Jun 16 23:09:39 localhost kernel:  [L6+0/2] 0xc010019f
   Jun 16 23:09:39 localhost kernel:  [<c010019f>] 0xc010019f
   Jun 16 23:09:39 localhost kernel: handlers:
   Jun 16 23:09:39 localhost kernel: [pg0+948806672/1068971008]
(usb_hcd_irq+0x0/0x80 [usbcore])
   Jun 16 23:09:39 localhost kernel: [<f8d65410>]
(usb_hcd_irq+0x0/0x80 [usbcore])
   Jun 16 23:09:39 localhost kernel: [pg0+951834816/1068971008]
(yenta_interrupt+0x0/0x40 [yenta_socket])
   Jun 16 23:09:39 localhost kernel: [<f90488c0>]
(yenta_interrupt+0x0/0x40 [yenta_socket])
   Jun 16 23:09:39 localhost kernel: Disabling IRQ #16

Attached to one USB port is a Logitech wheel mouse. The above is
reproducible; it is possible that it is the first movement of the
mouse or its wheel after the glxinfo program has finished triggers the
interupt problem. There also is a touchpad, but the problem occurs
regardless of the synaptics driver being present or not.

Running other programs using OpenGL like glxgears or tuxracer then
causes segfaults. With LIBGDL_DEBUG=3Dverbose, the output of glxgears is

   [marq@localhost ~]$ glxgears
   libGL: XF86DRIGetClientDriverName: 1.3.0 i915 (screen 0)
   libGL: OpenDriver: trying /usr/X11R6/lib/modules/dri/i915_dri.so
   drmOpenByBusid: Searching for BusID pci:0000:00:02.0
   drmOpenDevice: node name is /dev/dri/card0
   drmOpenDevice: open result is 4, (OK)
   drmOpenByBusid: drmOpenMinor returns 4
   drmOpenByBusid: drmGetBusid reports pci:0000:00:02.0
   Segmentation fault

and that of tuxracer:

   [marq@localhost ~]$ tuxracer
   Tux Racer 0.61 -- a Sunspire Studios Production
(http://www.sunspirestudios.com)
   (c) 1999-2000 Jasmin F. Patry <jf...@su...>
   "Tux Racer" is a trademark of Jasmin F. Patry
   Tux Racer comes with ABSOLUTELY NO WARRANTY. This is free software,
   and you are welcome to redistribute it under certain conditions.
   See http://www.gnu.org/copyleft/gpl.html for details.

   libGL: XF86DRIGetClientDriverName: 1.3.0 i915 (screen 0)
   libGL: OpenDriver: trying /usr/X11R6/lib/modules/dri/i915_dri.so
   drmOpenByBusid: Searching for BusID pci:0000:00:02.0
   drmOpenDevice: node name is /dev/dri/card0
   drmOpenDevice: open result is 5, (OK)
   drmOpenByBusid: drmOpenMinor returns 5
   drmOpenByBusid: drmGetBusid reports pci:0000:00:02.0
   Fatal signal: Segmentation Fault (SDL Parachute Deployed)

Eventually, I removed the dri module from the xorg.conf (that is the
version attached). Now, everything works fine, but of course without
acceleration. glxgears is fine, giving some 532 - 560 FPS, but
tuxracer is way too slow. Ok, I can live without the games...

So I assume that the problem is related to the dri driver for the
i915? I'm more than willing to run more tests if someone enlightened can
tell me what to do...

Thank you very much,

  Christian.