Age-old mouse pointer bug is fixed as is UT2004 performance. Excellent release, let's get this tested ASAP! Release notes at http://www2.ati.com/drivers/linux/linux_8.12.10.html Download at https://support.ati.com/ics/support/KBAnswer.asp?questionID=1176 Reproducible: Always Steps to Reproduce:
I'll try to update it tomorrow at last.
these do NOT work for me AT ALL. i tried with the same kernel configs that i got the last version to work on with both the latest 2.6.11-r{5,6} kernels and GLX would not load for dog poo...testing old latest drivers against -r5 kernel now...
ok it seems it IS in fact these ati modules. older ones work fine with the gentoo-sources-2.6.11-r5 kernel :(
on x86 the latest drivers are working fine, I'll try to fix the amd64 module patch soon (not having an amd64 delays me a bit)
Created attachment 55654 [details] ati-drivers-extra-8.12.10.patch.bz2 New Patch for ati-drivers-extra. Nothing has changed in fireglcontrol, but after a small removed line we need a new patch, so it will be patched without any warning. I removed 2 files from last patch which were completely uneccessary for the ebuild and which should never be included in the last patch.
Compiling errors against Linux 2.6.11 give errors when loading the module: fglrx: Unknown symbol agp_bind_memory fglrx: Unknown symbol agp_enable fglrx: Unknown symbol agp_backend_acquire fglrx: Unknown symbol agp_free_memory fglrx: Unknown symbol agp_allocate_memory fglrx: Unknown symbol agp_unbind_memory fglrx: Unknown symbol agp_copy_info fglrx: Unknown symbol agp_backend_release
<quote> on x86 the latest drivers are working fine, I'll try to fix the amd64 module patch soon (not having an amd64 delays me a bit) </quote> I'm on a Pentium M laptop with ATi Mobility 9200 and its not working :( also to rid those errors Alex had, make sure agpgart is in the kernel or as a module. I had those same errors after i ran `make modules_install`. adding in agpgart as a module fixed it and modprobing fglrx brought in agpgart as a dependancy o.0
Could you open a bug related to your single issue and mark this bug as blocker? I'd like to solve all the problems within the week and looks like I cannot reproduce most of the issues right now.
Re #6: You need to enable AGP support (CONFIG_AGP) and the appropriate chipset. Doesn't the ebuild check for this (CONFIG_CHECK="AGP" or sth like that)?
to: Christoph M erhm, chipset? no just include agpgart and dri and it wont show those errors. i left out everything ati and radeon in the kernel. in fact, if you add in ati agp support, fglrx will NEVER load :/
After compiling agpgart and dri modules and loading fglrx it does indeed bring in agpgart, but I still get these errors in my dmesg. fglrx: Unknown symbol agp_bind_memory fglrx: Unknown symbol agp_enable fglrx: Unknown symbol agp_backend_acquire fglrx: Unknown symbol agp_free_memory fglrx: Unknown symbol agp_allocate_memory fglrx: Unknown symbol agp_unbind_memory fglrx: Unknown symbol agp_copy_info fglrx: Unknown symbol agp_backend_release Linux agpgart interface v0.100 (c) Dave Jones [fglrx] Maximum main memory to use for locked dma buffers: 927 MBytes. ACPI: PCI interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16 [fglrx] module loaded - fglrx 8.12.10 [Apr 4 2005] on minor 0 mtrr: 0xe0000000,0x8000000 overlaps existing 0xe0000000,0x1000000 [fglrx:firegl_addmap] *ERROR* mtrr allocation failed (-22) allocation failed: out of vmalloc space - use vmalloc=<size> to increase size. [fglrx] AGP detected, AgpState = 0x00000000 (hardware caps of chipset) [fglrx:firegl_unlock] *ERROR* Process 14739 using kernel context 0 mtrr: 0xe0000000,0x8000000 overlaps existing 0xe0000000,0x1000000 [fglrx:firegl_addmap] *ERROR* mtrr allocation failed (-22) allocation failed: out of vmalloc space - use vmalloc=<size> to increase size. [fglrx] AGP detected, AgpState = 0x00000000 (hardware caps of chipset) [fglrx:firegl_unlock] *ERROR* Process 14917 using kernel context 0
amd64 people please test it and tell me if there is something wrong or not since I don't have the hardware to test and I'm getting too many contraddicting informations.
Works fine here - amd64/Radeon9600M kernel=2.6.11-r6
Closing, thanks for the feedback
What happened to ati-drivers-extra-8.12.10? I know that nothing new is in there, but the ati-drivers-extra-8.10.19 ebuild depends on media-video/ati-drivers-8.12.10. This is very annoying when you have installed both and you cannot update media-video/ati-drivers-extra (ok with portage overlay, but i think not all people want to create a seperate ebuild), media-video/ati-drivers-extra-8.10.19 wants to downgrade to media-video/ati-drivers-8.10.19, the next time it wants to update to media-video/ati-drivers-8.12.10 and then downgrade......
In my upload cue, not yet in sync. (got a bit busy and that had low priority)