Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 114760 - xorg-xfree 6.8.99.15-r4: i810 driver not using more than 8MB, thus disabling xv and what-not
Summary: xorg-xfree 6.8.99.15-r4: i810 driver not using more than 8MB, thus disabling ...
Status: RESOLVED WONTFIX
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: New packages (show other bugs)
Hardware: x86 Linux
: High major (vote)
Assignee: Gentoo X packagers
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2005-12-07 08:15 UTC by Lutz Lehmann
Modified: 2006-01-30 11:50 UTC (History)
0 users

See Also:
Package list:
Runtime testing required: ---


Attachments
Patch to supposedly remove the errors in i830CheckAvailableMemory (agp2.patch,524 bytes, patch)
2005-12-07 08:17 UTC, Lutz Lehmann
Details | Diff
Patch to supposedly remove the errors in i830CheckAvailableMemory (agp2.patch,524 bytes, patch)
2005-12-07 08:18 UTC, Lutz Lehmann
Details | Diff
x11-drm patch for 'drmAddMap(backbuffer) failed' (drm_bufs.c.patch,1.42 KB, patch)
2005-12-07 13:39 UTC, Lutz Lehmann
Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Lutz Lehmann 2005-12-07 08:15:37 UTC
Messed around w/ my i915 card a while trying to get drm/dri working. XV and   
HW-Acceleration are not working, because the i810 driver won't reserve more   
than the bios-provided 8M. Found a patch (attached) on the net for the 8.99.15  
release that is supposed to do the magic. Anyone willing tell me how to do an 
overlay to test the patch and possibly even get it into portage?  
  
TIA  
  
Lutz  

Reproducible: Always
Steps to Reproduce:
1. 855resolution (to get 1400x1050) 
2. startx  
Actual Results:  
(II) Reloading /usr/lib/modules/linux/libint10.so  
(II) I810(0): initializing int10  
(WW) I810(0): Bad V_BIOS checksum  
(II) I810(0): Primary V_BIOS segment is: 0xc000  
(II) I810(0): VESA BIOS detected  
(II) I810(0): VESA VBE Version 3.0  
(II) I810(0): VESA VBE Total Mem: 7872 kB  
(II) I810(0): VESA VBE OEM: Intel(r)915GM/910ML/915MS Graphics Chip Accelerated  
VGA BIOS  
(II) I810(0): VESA VBE OEM Software Rev: 1.0  
(II) I810(0): VESA VBE OEM Vendor: Intel Corporation  
(II) I810(0): VESA VBE OEM Product: Intel(r)915GM/910ML/915MS Graphics  
Controller  
(II) I810(0): VESA VBE OEM Product Rev: Hardware Version 0.0  
(II) I810(0): Integrated Graphics Chipset: Intel(R) 915GM  
(--) I810(0): Chipset: "915GM"  
(--) I810(0): Linear framebuffer at 0xC0000000  
(--) I810(0): IO registers at addr 0xD0400000  
(II) I810(0): 2 display pipes available.  
(II) I810(0): detected 7932 kB stolen memory.  
(II) I810(0): I830CheckAvailableMemory: -545509856 kB available  
(WW) I810(0): /dev/agpgart is either not available, or no memory is available  
for allocation.  Using pre-allocated memory only.  
(II) I810(0): Monitoring connected displays enabled  
(WW) I810(0): VideoRAM reduced to 7932 kByte (limited to available sysmem)  
(--) I810(0): Pre-allocated VideoRAM: 7932 kByte  
(--) I810(0): VideoRAM: 7932 kByte  
(==) I810(0): video overlay key set to 0x101fe  
(**) I810(0): page flipping disabled  
(==) I810(0): Using gamma correction (1.0, 1.0, 1.0)  
(II) I810(0): BIOS Build: 3412  
 
I presume the bad BIOS checksum results from 855resolution... 
  

Expected Results:  
I'd love to see X to actually use more than 8MB to get xv and accel working
Comment 1 Lutz Lehmann 2005-12-07 08:17:16 UTC
Created attachment 74235 [details, diff]
Patch to supposedly remove the errors in i830CheckAvailableMemory
Comment 2 Lutz Lehmann 2005-12-07 08:18:19 UTC
Created attachment 74236 [details, diff]
Patch to supposedly remove the errors in i830CheckAvailableMemory
Comment 3 Joshua Baergen (RETIRED) gentoo-dev 2005-12-07 09:24:01 UTC
This is probably of no issue in modular.  You're using masked software, why not
switch? :P

Overlays:
http://www.gentoo.org/doc/en/handbook/handbook-x86.xml?part=3&chap=5#doc_chap2

Then you just need to make a new patch tarball.  If you unpack one of the
current ones, you'll see a bunch of patches with a pretty standard naming
structure.  Just pick a random number and stick it in, repack it, and run
'ebuild foo.ebuild digest' on the ebuild.

Let me know how it turns out.  We're not really adding many patches to the
monoliths anymore, but if we do I'll tag this for addition.  Do you know if this
was a problem in 6.8.2 as well?
Comment 4 Lutz Lehmann 2005-12-07 13:38:14 UTC
(In reply to comment #3) 
> This is probably of no issue in modular.  You're using masked software, why 
not 
> switch? :P 
>  
 
Will (probably) do next week or so. Still trying to get the switching-howto 
into my head :) 
 
> Overlays: 
> 
http://www.gentoo.org/doc/en/handbook/handbook-x86.xml?part=3&chap=5#doc_chap2 
>  
> Then you just need to make a new patch tarball.  If you unpack one of the 
> current ones, you'll see a bunch of patches with a pretty standard naming 
> structure.  Just pick a random number and stick it in, repack it, and run 
> 'ebuild foo.ebuild digest' on the ebuild. 
 
After a little more googling, I found out that in  
http://lists.freedesktop.org/pipermail/xorg/2005-July/008735.html is stated, 
that the patch merely hides the problem but does not resolve it. 
 
Still gave the overlay and the patch a try, and it turned out to be working 
perfectly as far as the patching goes, even though I used epatch. However, it 
still didn't do for dri. 
Xorg.0.log said something about drmAddMap(backbuffer) failing.  
 
Found that one on https://bugs.freedesktop.org/show_bug.cgi?id=4908, so I made 
an overlay for x11-drm using patch #3, rebuilt that one, and got confused.  
 
First off, using 1400x1050-24 glxgears is running (or should I rather say 
creeping) at 250 and some frames. Using in-kernel dri gave approx. 750 fps 
 
On the other hand, Chromium B.S.U. is now at ~45 fps, it used to creep so slow 
that it took nearly 10sec for the menu to unfold.  
 
glxinfo confirms dri to be enabled. 
 
> Let me know how it turns out.  We're not really adding many patches to the 
> monoliths anymore, but if we do I'll tag this for addition.  Do you know if 
this 
> was a problem in 6.8.2 as well? 
 
Can't say right now, will try a downgrade tomorrow. 
Comment 5 Lutz Lehmann 2005-12-07 13:39:22 UTC
Created attachment 74250 [details, diff]
x11-drm patch for 'drmAddMap(backbuffer) failed'
Comment 6 Joshua Baergen (RETIRED) gentoo-dev 2005-12-07 14:18:47 UTC
I just realized this was assigned wrong.  Re-assigning to x11@.

Let us know how that turns out.
Comment 7 Lutz Lehmann 2005-12-08 03:40:14 UTC
(In reply to comment #4)
> > Let me know how it turns out.  We're not really adding many patches to the 
> > monoliths anymore, but if we do I'll tag this for addition.  Do you 
> > know if this was a problem in 6.8.2 as well? 
>  
> Can't say right now, will try a downgrade tomorrow. 

Just did the downgrade. Apparently, i810 driver from 6.8.2 does not like the 
idea of the chip having two subsequent BusIDs (PCI:0:2:0 and PCI:0:2:1). Tried 
several setups, but X won't start. Anyone actually running 6.8.2 on 915GM and 
NOT using the vesa driver?!?
Comment 8 Lutz Lehmann 2005-12-15 04:42:33 UTC
(In reply to comment #3)  
> This is probably of no issue in modular.  You're using masked software, why  
> not switch? :P  
  
I tried to switch to modular yesterday. I never realized how much of a pain in  
the a** it is to google for help with lynx :)  
  
Using spyderous' switching-howto gave a good guideline (thank you for that),  
but unfortunately X did not start up: could not open default font 'fixed'. I  
emerged pretty much everything under media-fonts that sounded right to me,  
especially font-alias, but to no avail. I'm not really comfortable with the X 
font system yet, but with someone taking my hand regarding the font problems, 
I'm willing to give modular another try.  
Comment 9 Donnie Berkholz (RETIRED) gentoo-dev 2006-01-19 12:06:02 UTC
(In reply to comment #8)
> Using spyderous' switching-howto gave a good guideline (thank you for that),  
> but unfortunately X did not start up: could not open default font 'fixed'. I  
> emerged pretty much everything under media-fonts that sounded right to me,  
> especially font-alias, but to no avail. I'm not really comfortable with the X 
> font system yet, but with someone taking my hand regarding the font problems, 
> I'm willing to give modular another try.  

The fixed font comes from font-misc-misc, and font-alias and encodings are also necessary.
Comment 10 Lutz Lehmann 2006-01-30 11:50:11 UTC
(In reply to comment #9)

> The fixed font comes from font-misc-misc, and font-alias and encodings are 
> also necessary.


Had to re-install my entire system last weekend due to a destroyed root-fs (but that's another story). Used 7.0-rc1, and it (now) works like a charm. I only had to patch x11-drm for the pci_pretty_name symbol as in #109757

Thanks to everyone. Please close this bug as you see fit