Are there any other error messages in the same area ? That sounds like your build failed for the 3D driver...
Announcement
Collapse
No announcement yet.
KMS + Radeon quick mini guide
Collapse
X
-
-
Have you confirmed that you are getting hardware acceleration with direct rendering ? Remember that the "direct rendering: yes" message doesn't mean hardware acceleration these days, you can have direct-rendered software paths as well.
What does glxinfo | grep renderer give you ?Test signature
Comment
-
Originally posted by bridgman View PostHave you confirmed that you are getting hardware acceleration with direct rendering ? Remember that the "direct rendering: yes" message doesn't mean hardware acceleration these days, you can have direct-rendered software paths as well.
At the moment, the Xorg log is failing at the AIGLX stuff
Code:(II) RADEON(0): [DRI2] Setup complete ... (EE) AIGLX error: Calling driver entry point failed (EE) AIGLX: reverting to software rendering
Comment
-
Understood. You hadn't said anything about direct rendering, but these days every time I ask someone if they have hardware acceleration they reply "yes I have direct rendering" so I figured I would try to save one back-and-forth transaction and respond first
There was an ABI change on Sunday night / Monday morning which required a libdrm_radeon update as well - AFAIK nothing is broken right now but pulling and building while the transition was happening could cause problems. Suggest you pull and build again. I believe the ABI change was between libdrm_radeon and mesa/ddx, no change to drm (kernel) itself.Test signature
Comment
-
Originally posted by forum1793 View PostThis is what my kernel config looks like for that area and it worked for me with hd3200 (780g) (I think R600 but not sure).
Code:# # Generic Driver Options # CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug" # CONFIG_DEVTMPFS is not set # CONFIG_STANDALONE is not set # CONFIG_PREVENT_FIRMWARE_BUILD is not set CONFIG_FW_LOADER=y CONFIG_FIRMWARE_IN_KERNEL=y CONFIG_EXTRA_FIRMWARE=""
Edit: this was for 32-rc8. Haven't tried yet with the 32 release.
Edit2: I also didn't do initrd
I didn't see anything for audio in Xorg.0.log but guess what, audio over HDMI works now, using the radeon driver as opposed to radeonhd. So, good job devs.
Comment
-
Originally posted by forum1793 View PostWell, to get audio working I went with 2.6.33-rc2 kernel and tried to use exact same config but updated (make oldconfig). This compiled but died looking for firmware and locked up after 60 sec. I then copied firmware to /lib/firmware/radeon. Same probelm. So I then recompiled kernel and only change was making agpgart, drm, and radeon into modules as opposed to in kernel. I didn't have to blacklist anything and kms worked.
I didn't see anything for audio in Xorg.0.log but guess what, audio over HDMI works now, using the radeon driver as opposed to radeonhd. So, good job devs.
There are two ways you could have got it working in the kernel statically:
1. By using an initrd and copying the firmware files into it. I don't like initrds.
2. Put the firmware files into <kernel-src>/firmware/radeon subdir, then build kernel with
CONFIG_EXTRA_FIRMWARE="radeon/R600_rlc.bin radeon/R700_rlc.bin"
CONFIG_EXTRA_FIRMWARE_DIR="firmware"
I prefer to have no initrd and few modules because then with 'fastboot', stuff can load more in parallel, for a faster boot.
Also I don't like the additional dependency on /boot/initrd-foo. If it goes away, I may have to reach for the rescue CD...
HTH.
Comment
Comment