Re: [Dri-users] CVS savage driver weirdness.

[Felix <fx...@gm...>]

Am Samstag, den 12.03.2005, 19:22 -0500 schrieb Rob Buse:
> attached is my Xorg.0.log, it seems there are errors I didn't see before.
>=20
> thanks!
> rob

[from your Xorg.0.log]
(EE) SAVAGE(0): [agp] AGP not available
(WW) SAVAGE(0): [agp] AGP failed to initialize -- falling back to PCI mode.
(WW) SAVAGE(0): [agp] Make sure you have the agpgart kernel module loaded.

It is what I almost suspected. AGP initialization fails completely. This
probably means that you don't have proper AGP support in your kernel or
you have not loaded the necessary kernel modules. Quoting
http://dri.freedesktop.org/wiki/DriTroubleshooting:

        [Linux] 2.6 users need to make sure both the agpgart core module
        and the appropriate chipset-specific module are loaded. Run
        lspci | grep "AGP bridge" to find out what sort of AGP chipset
        you have. Intel needs intel-agp, VIA needs via-agp, etc.

Anyway, the driver should be able to cleanly fall back to PCI mode in
this case, which it obviously doesn't. I'll look into this.

Note that you should try to get AGP support working for the best
performance and a reasonable amount of texture memory. You won't have
much fun without this.

Regards,
  Felix

>=20
>=20
> On 3/13/2005, "Felix K=FChling" <fx...@gm...> wrote:
>=20
> >Am Samstag, den 12.03.2005, 18:26 -0500 schrieb Rob Buse:
> >> >Is this the first time you're trying the Savage driver from CVS or ha=
s
> >> >it worked for you with an older version. On what hardware are you try=
ing
> >> >this? (lspci -v)
> >>=20
> >> This is the first time I'm trying to get it working.
> >>=20
> >> $/sbin/lspci -v
> >> 01:00.0 VGA compatible controller: S3 Inc. 86C270-294 Savage/IX-MV (re=
v
> >> 13) (prog-if 00 [VGA])
> >>         Subsystem: IBM ThinkPad T20
> >>         Flags: bus master, 66Mhz, medium devsel, latency 64, IRQ 3
> >>         Memory at f0000000 (32-bit, non-prefetchable) [size=3D128M]
> >>         Expansion ROM at <unassigned> [disabled] [size=3D64K]
> >>         Capabilities: [dc] Power Management version 1
> >>         Capabilities: [80] AGP version 1.0
> >>=20
> >> >Looks like something is going wrong during driver initialization. The=
re
> >> >are some failure cases that don't print any error messages (I should =
go
> >> >fix this). It seems no one has ever hit these failure cases till now.
> >>=20
> >> The patched driver gives the following output now:
> >>=20
> >> LIBGL_DEBUG=3Dverbose glxinfo
> >> name of display: :0.0
> >> libGL: XF86DRIGetClientDriverName: 2.0.0 savage (screen 0)
> >> libGL: OpenDriver: trying /usr/X11R6/lib/modules/dri/savage_dri.so
> >> drmOpenByBusid: Searching for BusID pci:0000:01:00.0
> >> drmOpenDevice: node name is /dev/dri/card0
> >> drmOpenDevice: open result is 4, (OK)
> >> drmOpenByBusid: drmOpenMinor returns 4
> >> drmOpenByBusid: drmGetBusid reports pci:0000:01:00.0
> >> Entering savageInitDriver.
> >> Mapping agpTextures failed (handle=3D0, size=3D16777216).
> >
> >This is strange. The size indicates that you are trying to use 16MB of
> >agp textures. This is already weird, because it should have subtracted
> >2MB for vertex DMA buffers. But handle=3D0 should not happen. It's eithe=
r
> >getting lost somewhere between the 2D driver and the 3D driver
> >(unlikely) or the allocation of AGP memory that fails the the 2D driver.
> >The Xserver log would tell more.
> >
> >> libGL warning: 3D driver returned no fbconfigs.
> >> libGL error: InitDriver failed
> >> libGL error: reverting to (slow) indirect rendering
> >> display: :0  screen: 0
> >> direct rendering: No
> >> server glx vendor string: SGI
> >> server glx version string: 1.2
> >> server glx extensions:
> >>     GLX_ARB_multisample, GLX_EXT_visual_info, GLX_EXT_visual_rating,
> >>     GLX_EXT_import_context, GLX_OML_swap_method,
> >> GLX_SGI_make_current_read,
> >>     GLX_SGIS_multisample, GLX_SGIX_fbconfig
> >>=20
> >>=20
> >> CVS status doesn't show any sticky tags or dates.  All was from a fres=
h
> >> checkout.
> >>=20
> >> >Also Xorg.0.log contains some information that may be helpful (all li=
nes
> >> >containing "[junkers]"). Maybe there is something weird about the map=
s
> >> >created by the 2D driver that make 3D initialization fail. Also, are
> >> >there any suspicious messages in the kernel log?
> >>=20
> >> I'm not seeing anything unordinary in Xorg.0.log or my kernel log.
> >
> >Please attach your complete Xorg.0.log. I'd like to see what it says
> >about the agp texture map and specifically the handle. If the allocation
> >fails then the 2D driver should detect that and fail to initialize DRI
> >altogether.
> >
> >Thanks,
> >  Felix
> >
> >>=20
> >> thanks for the help!
> >> rob
> >>=20
> >
> >--=20
> >| Felix K=FChling <fx...@gm...>                     http://fxk.de.vu =
|
> >| PGP Fingerprint: 6A3C 9566 5B30 DDED 73C3  B152 151C 5CC1 D888 E595 |
> >
--=20
| Felix K=FChling <fx...@gm...>                     http://fxk.de.vu |
| PGP Fingerprint: 6A3C 9566 5B30 DDED 73C3  B152 151C 5CC1 D888 E595 |