Home / Browse / Direct Rendering Infrastructure / Mail / Archive

Direct Rendering Infrastructure

Email Archive: dri-devel (read-only)

1999:	_Jan	_Feb	_Mar	_Apr	_May	_Jun	_Jul	_Aug	_Sep	_Oct	_Nov	_Dec (115)
2000:	_Jan (143)	_Feb (177)	_Mar (390)	_Apr (285)	_May (316)	_Jun (241)	_Jul (303)	_Aug (504)	_Sep (322)	_Oct (368)	_Nov (398)	_Dec (474)
2001:	_Jan (734)	_Feb (712)	_Mar (736)	_Apr (358)	_May (403)	_Jun (317)	_Jul (286)	_Aug (299)	_Sep (304)	_Oct (398)	_Nov (169)	_Dec (313)
2002:	_Jan (406)	_Feb (506)	_Mar (520)	_Apr (629)	_May (714)	_Jun (711)	_Jul (761)	_Aug (665)	_Sep (542)	_Oct (713)	_Nov (641)	_Dec (639)
2003:	_Jan (468)	_Feb (748)	_Mar (781)	_Apr (812)	_May (789)	_Jun (731)	_Jul (567)	_Aug (579)	_Sep (624)	_Oct (647)	_Nov (387)	_Dec (422)
2004:	_Jan (592)	_Feb (630)	_Mar (514)	_Apr (457)	_May (647)	_Jun (388)	_Jul (276)	_Aug (528)	_Sep (840)	_Oct (831)	_Nov (350)	_Dec (458)
2005:	_Jan (584)	_Feb (654)	_Mar (706)	_Apr (229)	_May (411)	_Jun (594)	_Jul (341)	_Aug (435)	_Sep (251)	_Oct (297)	_Nov (196)	_Dec (286)
2006:	_Jan (295)	_Feb (378)	_Mar (300)	_Apr (204)	_May (241)	_Jun (316)	_Jul (256)	_Aug (346)	_Sep (338)	_Oct (352)	_Nov (288)	_Dec (272)
2007:	_Jan (194)	_Feb (242)	_Mar (329)	_Apr (357)	_May (254)	_Jun (309)	_Jul (291)	_Aug (370)	_Sep (279)	_Oct (336)	_Nov (357)	_Dec (465)
2008:	_Jan (396)	_Feb (370)	_Mar (407)	_Apr (350)	_May (337)	_Jun (339)	_Jul (352)	_Aug (314)	_Sep (338)	_Oct (299)	_Nov (279)	_Dec (365)
2009:	_Jan (596)	_Feb (601)	_Mar (588)	_Apr (542)	_May (731)	_Jun (701)	_Jul (673)	_Aug (1050)	_Sep (740)	_Oct (750)	_Nov (774)	_Dec (1044)
2010:	_Jan (835)	_Feb (1215)	_Mar (1249)	_Apr (485)	_May (138)	_Jun (164)	_Jul (143)	_Aug (148)	_Sep (102)	_Oct (121)	_Nov (74)	_Dec (83)
2011:	_Jan (131)	_Feb (200)	_Mar (122)	_Apr (111)	_May (125)	_Jun (6)	_Jul	_Aug	_Sep (1)	_Oct (6)	_Nov (1)	_Dec (4)
2012:	_Jan	_Feb (4)	_Mar (2)	_Apr (3)	_May	_Jun	_Jul	_Aug (6)	_Sep (2)	_Oct (2)	_Nov	_Dec
2013:	_Jan (4)	_Feb (1)	_Mar	_Apr	_May	_Jun	_Jul	_Aug	_Sep	_Oct	_Nov	_Dec

[Dri-devel] Multiple PCI Radeon AIW

From: James Fung <fungja@ey...> - 2002-10-29 03:19

Hi,

Thanks for the help my PCI radeon.  

I'm now trying to use multiple PCI radeons on a computer, with DRI
hardware acceleration on each of them. 

So far, when I plug two radeons in, they show up on the PCI scanbus at
0:9:0 and 0:13:0.  I can use the 0:9:0 card, but when I change my
XF86Config to use the 0:13:0 card, X has problems - I get a bunch of
stripes and X locks up. I have made sure the cards themselves are
physically ok by running them each individually.

Similarly, when I try an XF86Config configured for multiple head support I
get blank screens.

Is there anything I could do to get multiple Radeons working?

Thanks.
-James

Re: [Dri-devel] Multiple PCI Radeon AIW

From: Michel <michel@da...> - 2002-10-29 17:49

On Die, 2002-10-29 at 04:22, James Fung wrote:
>=20
> Thanks for the help my PCI radeon. =20
>=20
> I'm now trying to use multiple PCI radeons on a computer, with DRI
> hardware acceleration on each of them.=20

I'm not sure that's supposed to work at this time.

> So far, when I plug two radeons in, they show up on the PCI scanbus at
> 0:9:0 and 0:13:0.  I can use the 0:9:0 card, but when I change my
> XF86Config to use the 0:13:0 card, X has problems - I get a bunch of
> stripes and X locks up. I have made sure the cards themselves are
> physically ok by running them each individually.

I suppose both cards work with DRI disabled? Please provide config
files, logs and kernel output, preferrably somewhere on the web.


--=20
Earthling Michel D=E4nzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
XFree86 and DRI project member   /  CS student, Free Software enthusiast

Re: [Dri-devel] Multiple PCI Radeon AIW

From: James Fung <fungja@ey...> - 2002-10-29 23:45

> I suppose both cards work with DRI disabled? Please provide config
> files, logs and kernel output, preferrably somewhere on the web.

Oops - my XF86Config-4 file had some whitespace problems which I fixed.  
I've got two PCI radeons working now.

However, only one has dri enabled.   Is there a way I can get the multiple
cards each to use hardware acceleration?

I put some log files at http://www.eyetap.org/~fungja/radeon/ 


II) RADEON(0): [drm] installed DRM signal handler
(II) RADEON(0): [DRI] installation complete
(II) RADEON(0): [drm] Added 32 65536 byte vertex/indirect buffers
(II) RADEON(0): [drm] Mapped 32 vertex/indirect buffers
(II) RADEON(0): Direct rendering enabled
drmOpenDevice: minor is 0
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 5, (OK)
drmOpenDevice: minor is 0
drmOpenDevice: node name is /dev/dri/card0

...[snip]...

drmOpenDevice: node name is /dev/dri/card14
drmOpenDevice: open result is -1, (No such device)
drmOpenDevice: open result is -1, (No such device)
drmOpenDevice: Open failed
drmOpenByBusid: busid is PCI:0:9:0
drmOpenDevice: minor is 0
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 5, (OK)
drmOpenByBusid: drmOpenMinor returns 5
drmOpenByBusid: drmGetBusid reports PCI:0:9:0
(II) RADEON(1): [drm] drmSetBusid failed (5, PCI:0:13:0), Device or
resource busy
(EE) RADEON(1): [dri] DRIScreenInit failed.  Disabling DRI.

...


Thanks.

-James

Re: [Dri-devel] Multiple PCI Radeon AIW

From: Michel <michel@da...> - 2002-11-01 15:36

Attachments: radeon_drv.diff

On Mit, 2002-10-30 at 00:48, James Fung wrote:=20
> > I suppose both cards work with DRI disabled? Please provide config
> > files, logs and kernel output, preferrably somewhere on the web.
>=20
> Oops - my XF86Config-4 file had some whitespace problems which I fixed. =20
> I've got two PCI radeons working now.
>=20
> However, only one has dri enabled.   Is there a way I can get the multipl=
e
> cards each to use hardware acceleration?
>=20
> I put some log files at http://www.eyetap.org/~fungja/radeon/=20
>=20
>=20
> II) RADEON(0): [drm] installed DRM signal handler
> (II) RADEON(0): [DRI] installation complete
> (II) RADEON(0): [drm] Added 32 65536 byte vertex/indirect buffers
> (II) RADEON(0): [drm] Mapped 32 vertex/indirect buffers
> (II) RADEON(0): Direct rendering enabled
> drmOpenDevice: minor is 0
> drmOpenDevice: node name is /dev/dri/card0
> drmOpenDevice: open result is 5, (OK)
> drmOpenDevice: minor is 0
> drmOpenDevice: node name is /dev/dri/card0
>=20
> ...[snip]...
>=20
> drmOpenDevice: node name is /dev/dri/card14
> drmOpenDevice: open result is -1, (No such device)
> drmOpenDevice: open result is -1, (No such device)
> drmOpenDevice: Open failed
> drmOpenByBusid: busid is PCI:0:9:0
> drmOpenDevice: minor is 0
> drmOpenDevice: node name is /dev/dri/card0
> drmOpenDevice: open result is 5, (OK)
> drmOpenByBusid: drmOpenMinor returns 5
> drmOpenByBusid: drmGetBusid reports PCI:0:9:0
> (II) RADEON(1): [drm] drmSetBusid failed (5, PCI:0:13:0), Device or
> resource busy
> (EE) RADEON(1): [dri] DRIScreenInit failed.  Disabling DRI.

After looking at the source, support for multiple cards is only enabled
in the tdfx DRM yet. The attached hack might serve as a quick'n'dirty
way to find out the next problem. :)


--=20
Earthling Michel D=E4nzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
XFree86 and DRI project member   /  CS student, Free Software enthusiast

Re: [Dri-devel] Multiple PCI Radeon AIW

From: James Fung <fungja@ey...> - 2002-11-01 20:45

> After looking at the source, support for multiple cards is only enabled
> in the tdfx DRM yet. The attached hack might serve as a quick'n'dirty
> way to find out the next problem. :)

I added in the #define from the patch.  

When I insmod, it now sees both cards.  I get something like:
[drm] Initialized radeon 1.6.0 20020828 on minor 0
[drm] Initialized radeon 1.6.0 20020828 on minor 1

( Here's "the next problem" :) )
X, however, doesnt work.  I get 1 blank screen and one with crazy lines
zipping around. 

both cards are seen by X now: 
(II) RADEON(0): [drm] created "radeon" driver at busid "PCI:0:9:0"
...
(II) RADEON(1): [drm] created "radeon" driver at busid "PCI:0:13:0"
...
but then I get 

(II) RADEON(1): X context handle = 0x00000001
(II) RADEON(1): [drm] installed DRM signal handler
(II) RADEON(1): [DRI] installation complete
(II) RADEON(1): [drm] Added 32 65536 byte vertex/indirect buffers
(II) RADEON(1): [drm] Mapped 32 vertex/indirect buffers
(EE) RADEON(1): RADEONDRICPInit: CP start -1007
(EE) RADEON(1): RADEONCPGetBuffer: CP GetBuffer -1007
(EE) RADEON(1): GetBuffer timed out, resetting engine...
(EE) RADEON(1): RADEONCPGetBuffer: CP reset -1007
(EE) RADEON(1): RADEONCPGetBuffer: CP start -1007
(EE) RADEON(1): RADEONCPGetBuffer: CP GetBuffer -1007
(EE) RADEON(1): GetBuffer timed out, resetting engine...
(EE) RADEON(1): RADEONCPGetBuffer: CP reset -1007
(EE) RADEON(1): RADEONCPGetBuffer: CP start -1007
(EE) RADEON(1): RADEONCPGetBuffer: CP GetBuffer -1007
(EE) RADEON(1): GetBuffer timed out, resetting engine...

and so on.

The full log is at 
http://www.eyetap.org/~fungja/radeon/XFree86.0.log.dualradeon

-James

Re: [Dri-devel] Multiple PCI Radeon AIW

From: Michel <michel@da...> - 2002-11-02 14:02

On Fre, 2002-11-01 at 21:48, James Fung wrote:=20
> > After looking at the source, support for multiple cards is only enabled
> > in the tdfx DRM yet. The attached hack might serve as a quick'n'dirty
> > way to find out the next problem. :)
>=20
> I added in the #define from the patch. =20
>=20
> When I insmod, it now sees both cards.  I get something like:
> [drm] Initialized radeon 1.6.0 20020828 on minor 0
> [drm] Initialized radeon 1.6.0 20020828 on minor 1
>=20
> ( Here's "the next problem" :) )
> X, however, doesnt work.  I get 1 blank screen and one with crazy lines
> zipping around.=20
>=20
> both cards are seen by X now:=20
> (II) RADEON(0): [drm] created "radeon" driver at busid "PCI:0:9:0"
> ...
> (II) RADEON(1): [drm] created "radeon" driver at busid "PCI:0:13:0"
> ...
> but then I get=20
>=20
> (II) RADEON(1): X context handle =3D 0x00000001
> (II) RADEON(1): [drm] installed DRM signal handler
> (II) RADEON(1): [DRI] installation complete
> (II) RADEON(1): [drm] Added 32 65536 byte vertex/indirect buffers
> (II) RADEON(1): [drm] Mapped 32 vertex/indirect buffers
> (EE) RADEON(1): RADEONDRICPInit: CP start -1007
> (EE) RADEON(1): RADEONCPGetBuffer: CP GetBuffer -1007
> (EE) RADEON(1): GetBuffer timed out, resetting engine...
> (EE) RADEON(1): RADEONCPGetBuffer: CP reset -1007
> (EE) RADEON(1): RADEONCPGetBuffer: CP start -1007
> (EE) RADEON(1): RADEONCPGetBuffer: CP GetBuffer -1007
> (EE) RADEON(1): GetBuffer timed out, resetting engine...
> (EE) RADEON(1): RADEONCPGetBuffer: CP reset -1007
> (EE) RADEON(1): RADEONCPGetBuffer: CP start -1007
> (EE) RADEON(1): RADEONCPGetBuffer: CP GetBuffer -1007
> (EE) RADEON(1): GetBuffer timed out, resetting engine...
>=20
> and so on.

That's what I meant by 'not sure if this is supposed to work at this
time'. :/ Looks like it might take some serious hacking to get this
working...

Is there anything in the kernel output though? (if you insmod the radeon
kernel module with drm_opts=3Ddebug ?)


PS: Any specific reason why you're trying 4.2.0 instead of the DRI CVS
trunk?

PPS: What's 'subtraction clamp blending'? :)

--=20
Earthling Michel D=E4nzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
XFree86 and DRI project member   /  CS student, Free Software enthusiast