Re: r200 oops
Dave Airlie <[EMAIL PROTECTED]> writes: > can you insert the module with drm_opts=debug option and see what it > says.. > > can you also send on lspci -v and the contents of /proc/iomem unfortunately my disk just died so i won't be able to do any testing until ibm sends a new one. sorry. --alex-- -- | I believe the moment is at hand when, by a paranoiac and active | | advance of the mind, it will be possible (simultaneously with | | automatism and other passive states) to systematize confusion | | and thus to help to discredit completely the world of reality. | --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Mesa 6.1 released
I've wrapped up Mesa 6.1. The CVS tree has been tagged with 'mesa_6_1' and the tarballs can be downloaded from the SourceForge site. I'll leave it up to the DRI/X.org developers to pull this into the upcoming X server release. There have been only minor changes over the past week, I believe. Mesa developers: if you're anxious to start checking in new development / new feature code into the Mesa CVS tree, let me know and I'll create a 'mesa_6_2_branch' branch. Then the trunk will be open for new development while the mesa_6_2_branch will be for bug fixes for the 6.2 release (the usual system). But until I create that branch, the CVS trunk should be for bug fixes only. Thanks. -Brian --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: drm round 2...
Alan Cox wrote: On Mer, 2004-08-18 at 12:32, Keith Whitwell wrote: Once again, I predict the gamma driver which reportedly doesn't work and doesn't have any users will prove to be the stumbling block... I would suggest the gamma driver is retired. And I think I say that as about the only Linux user other than Alan Hourihane who has one 8) I've got a card, but I don't have an AGP 2x system to put it in. I also have an older 500TX based card. My vote would be to chop the existing driver. When the time comes, we can write a new driver that will support the whole "family" (i.e., 500TX, DMX, GMX, Permedia2, etc.) of cards. I've been wanting to do this for over a year now, but I'm too swamped with other stuff. :( --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: drm round 2...
On Wed, Aug 18, 2004 at 07:57:56PM +0100, Keith Whitwell wrote: > Alan Cox wrote: > >On Mer, 2004-08-18 at 12:32, Keith Whitwell wrote: > > > >>Once again, I predict the gamma driver which reportedly doesn't work and > >>doesn't have any users will prove to be the stumbling block... > > > > > >I would suggest the gamma driver is retired. And I think I say that as > >about the only Linux user other than Alan Hourihane who has one 8) > > > > That would be fine with me... Dave, AlanH, has the moment arrived? Chop it! Alan. --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: drm round 2...
Alan Cox wrote: On Mer, 2004-08-18 at 12:32, Keith Whitwell wrote: Once again, I predict the gamma driver which reportedly doesn't work and doesn't have any users will prove to be the stumbling block... I would suggest the gamma driver is retired. And I think I say that as about the only Linux user other than Alan Hourihane who has one 8) That would be fine with me... Dave, AlanH, has the moment arrived? Keith --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: drm round 2...
On Mer, 2004-08-18 at 12:32, Keith Whitwell wrote: > Once again, I predict the gamma driver which reportedly doesn't work and > doesn't have any users will prove to be the stumbling block... I would suggest the gamma driver is retired. And I think I say that as about the only Linux user other than Alan Hourihane who has one 8) --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
vertex program support in r200 driver
I've just added vertex program support to the r200 driver. It uses the software emulation Mesa provies, though the hardware could do accelerated vertex programs. I'll do some more testing of GL_ARB_vertex_program, though I haven't seen any problems so far. GL_NV_vertex_program is completely untested and only enabled since Mesa requires it when using GL_ARB_vertex_program. Philipp Klaus Krause --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: drm round 2...
Dave Airlie wrote: Okay take a look at http://www.skynet.ie/~airlied/patches/dri/mtrr_removal.diff This is how I intend dumping the __HAVE_ set of macros, I've just patched the radeon in this patch.. any objections to this approach any neater ways to do it? That looks good. Having the bits explicitly tracked in the driver at run-time may help debugging in some cases. My only comment is a minor code style issue. I believe that GCC gives a warning for 'if ( foo & BAR )', so I usually use 'if ( (foo & BAR) != 0 )'. Using in-line functions (as suggested by Sam) may be even better. In any case, thank you *very* much for tackling this with so much gusto! :) I'm sure this clean-up hasn't been the most exciting way to spend your coding time. --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: drm round 2...
On Wed, Aug 18, 2004 at 01:10:56PM +0100, Dave Airlie wrote:
>
> >
> > A few points (based only on the diff).
> > 1) Rename CONFIG_MTRR to CONFIG_DRM_MTRR to provide better context
>
> well we are using the kernels CONFIG_MTRR so if the kernel has MTRR
> built-in we want to use it
Back home with access to src.
Yes - I realise MTRR is nothing DRM specific so for sure use it as it is.
> .. so I've noticed it already supplies stubs
> for mtrr_add/mtrr_del if it isn't built in so I'll look into removing the
> MTRR detects completely..
>
> > 3) The following code
> > +#if __REALLY_HAVE_AGP && defined(CONFIG_MTRR)
> > + if ( (dev->driver_features & DRIVER_USE_MTRR) && dev->agp && ...
> >
> > Could be replaced by a stub function:
> > #ifdef CONFIG_DRM_MTRR
> > static inline int driver_has_mtrr(...)
> > {
> >return (feature & DRIVER_USE_MTRR) != 0;
> > }
> > #else
> > static inline int driver_has_mtrr(...)
> > {
> >return 0;
> > }
> > #endif
> > This check seems to be used more than one place.
> >
> I might do something like that alright.. looks nicer...
>
> > 4) Use #ifdef, not #if defined()
> > 5) Why not post the diff, makes it _much_ easier to reply wtih code comments.
>
> I'll probably start doing that.. it's just a change in my workflow :-),
> the machine I email from isn't the same one I work on so it means moving
> things around differently..
Good, a bit more trouble for you, but makes it much easier for many..
Sam
---
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
--
___
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: Weekly IRC meeting reminder
On 16 Aug 2004 18:30:00 -, Ian Romanick <[EMAIL PROTECTED]> wrote: > > This is just a friendly reminder that the weekly dri-devel IRC meeting will > be starting in the #dri-devel channel on irc.freenode.net at 2100 UTC (or > 5:00PM EDT or 2:00PM PDT, if you prefer). > > Time zone conversion available at: > > http://www.timezoneconverter.com/cgi-bin/tzc.tzc > > Logs of previous IRC meetings are available at: > > http://dri.sourceforge.net/IRC-logs/ Does anyone have a log of this week's meeting? I was only able to join for a few minutes. Thanks, Alex --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: drm round 2...
>
> A few points (based only on the diff).
> 1) Rename CONFIG_MTRR to CONFIG_DRM_MTRR to provide better context
well we are using the kernels CONFIG_MTRR so if the kernel has MTRR
built-in we want to use it .. so I've noticed it already supplies stubs
for mtrr_add/mtrr_del if it isn't built in so I'll look into removing the
MTRR detects completely..
> 3) The following code
> +#if __REALLY_HAVE_AGP && defined(CONFIG_MTRR)
> + if ( (dev->driver_features & DRIVER_USE_MTRR) && dev->agp && ...
>
> Could be replaced by a stub function:
> #ifdef CONFIG_DRM_MTRR
> static inline int driver_has_mtrr(...)
> {
>return (feature & DRIVER_USE_MTRR) != 0;
> }
> #else
> static inline int driver_has_mtrr(...)
> {
>return 0;
> }
> #endif
> This check seems to be used more than one place.
>
I might do something like that alright.. looks nicer...
> 4) Use #ifdef, not #if defined()
> 5) Why not post the diff, makes it _much_ easier to reply wtih code comments.
I'll probably start doing that.. it's just a change in my workflow :-),
the machine I email from isn't the same one I work on so it means moving
things around differently..
Dave.
--
David Airlie, Software Engineer
http://www.skynet.ie/~airlied / airlied at skynet.ie
pam_smb / Linux DECstation / Linux VAX / ILUG person
---
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
--
___
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: drm round 2...
>
> Okay take a look at
>
> http://www.skynet.ie/~airlied/patches/dri/mtrr_removal.diff
>
> This is how I intend dumping the __HAVE_ set of macros, I've just patched
> the radeon in this patch.. any objections to this approach any neater ways
> to do it?
A few points (based only on the diff).
1) Rename CONFIG_MTRR to CONFIG_DRM_MTRR to provide better context
2) When CONFIR_DRM_MTRR is not defined provide dummy stubs for:
mtrr_add and mtrr_del
See for example include/linux/procfs.h for good examples
Implementing the dummy functions would significanly improve readability of
code.
3) The following code
+#if __REALLY_HAVE_AGP && defined(CONFIG_MTRR)
+ if ( (dev->driver_features & DRIVER_USE_MTRR) && dev->agp && ...
Could be replaced by a stub function:
#ifdef CONFIG_DRM_MTRR
static inline int driver_has_mtrr(...)
{
return (feature & DRIVER_USE_MTRR) != 0;
}
#else
static inline int driver_has_mtrr(...)
{
return 0;
}
#endif
This check seems to be used more than one place.
4) Use #ifdef, not #if defined()
5) Why not post the diff, makes it _much_ easier to reply wtih code comments.
Sam
---
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
--
___
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: drm round 2...
> This does make it a lot clearer what's going on with those REALLY_ macros. I'm > wondering if there are any of those capability bits that can be dropped > because all or no drivers use them. Like CTX_BITMAP, for instance. Yes CTX_BITMAP is what I'm looking at now.. it is the gamma that causes the problem with it, I'm trying to move as many of the ugly hacks into the gamma ... > Once again, I predict the gamma driver which reportedly doesn't work and > doesn't have any users will prove to be the stumbling block... The gamma does keep me honest, every shortcut I attempt it messes it up :-) I've replaced the CONFIG_AGP and CONFIG_MTRR with OS_HAS_AGP and OS_HAS_MTRR macros in my current tree as BSD needs to define them as well for some drivers... I've started another branch drmfntbl-0-0-2-branch and I've checked in the first AGP/MTRR on it .. (I don't think it builds at the moment.. but I'm really keeping the branches for my own sanity in case I make a big mistake..) Dave. -- David Airlie, Software Engineer http://www.skynet.ie/~airlied / airlied at skynet.ie pam_smb / Linux DECstation / Linux VAX / ILUG person --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: drm round 2...
Dave Airlie wrote: Okay take a look at http://www.skynet.ie/~airlied/patches/dri/mtrr_removal.diff This is how I intend dumping the __HAVE_ set of macros, I've just patched the radeon in this patch.. any objections to this approach any neater ways to do it? Regards, Dave. This does make it a lot clearer what's going on with those REALLY_ macros. I'm wondering if there are any of those capability bits that can be dropped because all or no drivers use them. Like CTX_BITMAP, for instance. Once again, I predict the gamma driver which reportedly doesn't work and doesn't have any users will prove to be the stumbling block... Keith --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
drm round 2...
Okay take a look at http://www.skynet.ie/~airlied/patches/dri/mtrr_removal.diff This is how I intend dumping the __HAVE_ set of macros, I've just patched the radeon in this patch.. any objections to this approach any neater ways to do it? Regards, Dave. -- David Airlie, Software Engineer http://www.skynet.ie/~airlied / airlied at skynet.ie pam_smb / Linux DECstation / Linux VAX / ILUG person --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: r200 oops
can you insert the module with drm_opts=debug option and see what it says.. can you also send on lspci -v and the contents of /proc/iomem Dave. > using the latest cvs (as of this morning) i get the following oops > when loading the radeon module: > > kernel: mtrr: 0xe000,0x200 overlaps existing 0xe000,0x40 > kernel: ACPI: PCI interrupt :01:00.0[A] -> GSI 4 (level, low) -> IRQ 4 > kernel: [drm:radeon_register_regions] *ERROR* cannot reserve FB region > kernel: Unable to handle kernel NULL pointer dereference at virtual address > kernel: printing eip: > kernel: e1195520 > kernel: *pde = > kernel: Oops: 0002 [#1] > kernel: PREEMPT > kernel: Modules linked in: radeon sd_mod scsi_mod autofs4 microcode aes_i586 airo > ehci_hcd uhci_hcd usbcore parport_pc parport irtty_sir sir_dev irda crc_ccitt joydev > evdev nls_iso8859_1 ntfs yenta_socket > kernel: CPU:0 > kernel: EIP:0060:[__crc_snd_ctl_free_one+3747574/4594735]Not tainted > kernel: EFLAGS: 00013246 (2.6.8.1) > kernel: EIP is at radeon_stub_putminor+0x20/0x110 [radeon] > kernel: eax: ebx: ecx: edx: > kernel: esi: edi: ffea ebp: dcab1e6c esp: dcab1e5c > kernel: ds: 007b es: 007b ss: 0068 > kernel: Process modprobe (pid: 2062, threadinfo=dcab task=df585220) > kernel: Stack: e1198003 07ff e11ac160 dcab1e84 e1195913 dcab1e84 > e11ab9e0 > kernel: e11ab9e0 dcab1ed0 e1191488 41ed dcab1eac > c017c056 > kernel:4122356e dc236654 db41c480 e11a9e18 c15c9c00 > dc236654 > kernel: Call Trace: > kernel: [show_stack+122/144] show_stack+0x7a/0x90 > kernel: [show_registers+329/416] show_registers+0x149/0x1a0 > kernel: [die+141/256] die+0x8d/0x100 > kernel: [do_page_fault+485/1383] do_page_fault+0x1e5/0x567 > kernel: [error_code+45/56] error_code+0x2d/0x38 > kernel: [__crc_snd_ctl_free_one+3748585/4594735] radeon_stub_unregister+0x33/0x60 > [radeon] > kernel: [__crc_snd_ctl_free_one+3731038/4594735] drm_probe+0x208/0x270 [radeon] > kernel: [pci_device_probe_static+73/96] pci_device_probe_static+0x49/0x60 > kernel: [__pci_device_probe+55/80] __pci_device_probe+0x37/0x50 > kernel: [pci_device_probe+38/80] pci_device_probe+0x26/0x50 > kernel: [bus_match+53/112] bus_match+0x35/0x70 > kernel: [driver_attach+89/144] driver_attach+0x59/0x90 > kernel: [bus_add_driver+140/176] bus_add_driver+0x8c/0xb0 > kernel: [driver_register+43/48] driver_register+0x2b/0x30 > kernel: [pci_register_driver+89/128] pci_register_driver+0x59/0x80 > kernel: [__crc_snd_ctl_free_one+2013811/4594735] drm_init+0x9d/0xd7 [radeon] > kernel: [sys_init_module+248/528] sys_init_module+0xf8/0x210 > kernel: [syscall_call+7/11] syscall_call+0x7/0xb > kernel: Code: c7 04 10 00 00 00 00 a1 f0 d7 1a e1 89 4c 10 04 a1 f0 d7 1a > > this is on a thinkpad t40 using a radeon mobility 9000 card. i can > provide more info if needed. > > --alex-- > > -- David Airlie, Software Engineer http://www.skynet.ie/~airlied / airlied at skynet.ie pam_smb / Linux DECstation / Linux VAX / ILUG person --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: Adding extension implemented by Mesa in software to driver
Philipp Klaus Krause wrote: How is functionality already implemented by Mesa added to a driver? I want to try adding vertex program support to the r200 driver. I took a look at the i915 driver to see what is done there. The only things I could find where the added extensions in the extension string and the added vertex program stage in the pipeline struct. Where else should I look and modify the r200 driver? You normally would only need to turn the extension on. However, there is a conflict in the r200 driver, as that hardware implements TNL but not vertex programs. So, you'll need to modify the driver to recognize when a vertex program is active, and make sure that a TNL fallback takes place so that Mesa gets to handle the program. Keith --- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 -- ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
