Bug 66519

Created attachment 81907 [details]
Xorg.0.log.old

Created attachment 81908 [details]
dmesg

Does it work if build the driver as a module and load it manually after the system has booted to a non-X runlevel?

(In reply to comment #3)
> Does it work if build the driver as a module and load it manually after the
> system has booted to a non-X runlevel?

No, same problem:

[   13.533427] igb: eth3 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX/TX
[   13.533706] IPv6: ADDRCONF(NETDEV_CHANGE): eth3: link becomes ready

(end of boot single mode)

Then, load module (modprobe radeon):

[  189.052514] [drm] radeon kernel modesetting enabled.
[  189.052851] [drm] initializing kernel modesetting (CEDAR 0x1002:0x68E1 0x1787:0x3000).
[  189.052936] [drm] register mmio base: 0xFBC20000
[  189.052984] [drm] register mmio size: 131072
[  189.053098] ATOM BIOS: PARK
[  189.053195] radeon 0000:05:00.0: VRAM: 1024M 0x0000000000000000 - 0x000000003FFFFFFF (1024M used)
[  189.053251] radeon 0000:05:00.0: GTT: 512M 0x0000000040000000 - 0x000000005FFFFFFF
[  189.070227] [drm] Detected VRAM RAM=1024M, BAR=256M
[  189.070279] [drm] RAM width 64bits DDR
[  189.070449] [TTM] Zone  kernel: Available graphics memory: 33022834 kiB
[  189.070499] [TTM] Zone   dma32: Available graphics memory: 2097152 kiB
[  189.070548] [TTM] Initializing pool allocator
[  189.070599] [TTM] Initializing DMA pool allocator
[  189.070677] [drm] radeon: 1024M of VRAM memory ready
[  189.070729] [drm] radeon: 512M of GTT memory ready.
[  189.070855] radeon 0000:05:00.0: ffff88103d246c00 unpin not necessary
[  189.189112] radeon 0000:05:00.0: fence driver on ring 5 use gpu addr 0x000000000005c418 and cpu addr 0xffffc900159ba418
[  189.189170] [drm] GART: num cpu pages 131072, num gpu pages 131072
[  189.189646] [drm] enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0
[  189.189739] [drm] Loading CEDAR Microcode
[  189.208711] [drm] PCIE GART of 512M enabled (table at 0x0000000000040000).
[  189.208891] radeon 0000:05:00.0: WB enabled
[  189.208947] radeon 0000:05:00.0: fence driver on ring 0 use gpu addr 0x0000000040000c00 and cpu addr 0xffff88103a3eec00
[  189.209018] radeon 0000:05:00.0: fence driver on ring 3 use gpu addr 0x0000000040000c0c and cpu addr 0xffff88103a3eec0c
[  189.222481] radeon 0000:05:00.0: fence driver on ring 5 use gpu addr 0x000000000015e418 and cpu addr 0xffffc9001621c418
[  189.222554] [drm] Supports vblank timestamp caching Rev 1 (10.10.2010).
[  189.222612] [drm] Driver supports precise vblank timestamp query.
[  189.222691] radeon 0000:05:00.0: irq 130 for MSI/MSI-X
[  189.222701] radeon 0000:05:00.0: radeon: using MSI.
[  189.222784] [drm] radeon: irq initialized.
[  189.239494] [drm] ring test on 0 succeeded in 1 usecs
[  189.239608] [drm] ring test on 3 succeeded in 1 usecs
[  190.415928] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  191.436153] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  192.456377] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  193.476604] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  194.496828] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  195.517052] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  196.537273] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  197.557497] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  198.577725] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  199.597950] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  199.618022] [drm:r600_uvd_init] *ERROR* UVD not responding, giving up!!!
[  199.618083] [drm:evergreen_startup] *ERROR* radeon: error initializing UVD (-1).
[  199.618369] [drm] ib test on ring 0 succeeded in 0 usecs
[  199.618452] [drm] ib test on ring 3 succeeded in 0 usecs
[  199.619870] [drm] Radeon Display Connectors
[  199.619930] [drm] Connector 0:
[  199.619985] [drm]   DP-1
[  199.620039] [drm]   HPD2
[  199.620095] [drm]   DDC: 0x6460 0x6460 0x6464 0x6464 0x6468 0x6468 0x646c 0x646c
[  199.620163] [drm]   Encoders:
[  199.620218] [drm]     DFP1: INTERNAL_UNIPHY1
[  199.620274] [drm] Connector 1:
[  199.620330] [drm]   DVI-I-1
[  199.620384] [drm]   HPD4
[  199.620439] [drm]   DDC: 0x6450 0x6450 0x6454 0x6454 0x6458 0x6458 0x645c 0x645c
[  199.620508] [drm]   Encoders:
[  199.620563] [drm]     DFP2: INTERNAL_UNIPHY
[  199.620623] [drm]     CRT1: INTERNAL_KLDSCP_DAC1
[  199.620680] [drm] Connector 2:
[  199.620735] [drm]   VGA-1
[  199.620790] [drm]   DDC: 0x6430 0x6430 0x6434 0x6434 0x6438 0x6438 0x643c 0x643c
[  199.620859] [drm]   Encoders:
[  199.620913] [drm]     CRT2: INTERNAL_KLDSCP_DAC2
[  199.621012] [drm] Internal thermal controller with fan control
[  199.621137] [drm] radeon: power management initialized
[  199.715497] [drm] fb mappable at 0x3C0FE035F000
[  199.715554] [drm] vram apper at 0x3C0FE0000000
[  199.715608] [drm] size 9216000
[  199.715661] [drm] fb depth is 24
[  199.715714] [drm]    pitch is 7680
[  199.715797] fbcon: radeondrmfb (fb0) is primary device
[  199.990513] Console: switching to colour frame buffer device 240x75
[  200.042891] radeon 0000:05:00.0: fb0: radeondrmfb frame buffer device
[  200.043109] radeon 0000:05:00.0: registered panic notifier
[  200.043296] [drm] Initialized radeon 2.33.0 20080528 for 0000:05:00.0 on minor 0

Justin.

Created attachment 81945 [details] [review]
Debugging patch

Please try the attached patch, it shouldn't fix the issue but instead outputs some debug messages ("IH: CP....") into dmesg.

If you don't see those messages try to but the system under some graphics load, for example start glxgears ten times simultaneously or something like that.

attach(In reply to comment #5)
> Created attachment 81945 [details] [review] [review]
> Debugging patch
> 
> Please try the attached patch, it shouldn't fix the issue but instead
> outputs some debug messages ("IH: CP....") into dmesg.
> 
> If you don't see those messages try to but the system under some graphics
> load, for example start glxgears ten times simultaneously or something like
> that.

Yes, that is what happened:

[  334.326551] [drm:evergreen_irq_process] *ERROR* IH: CP EOP
[  334.326700] [drm:evergreen_irq_process] *ERROR* IH: CP EOP
[  334.326922] [drm:evergreen_irq_process] *ERROR* IH: CP EOP
[  334.327071] [drm:evergreen_irq_process] *ERROR* IH: CP EOP
[  334.327190] [drm:evergreen_irq_process] *ERROR* IH: CP EOP

Attaching full dmesg as well.

Created attachment 81952 [details]
dmesg with patch

Please do a "make clean" then and fully recompile your kernel.

The kernel makefile doesn't recognize it when you change the firmware on the disk and still builds the old firmware file into the kernel.

(In reply to comment #8)
> Please do a "make clean" then and fully recompile your kernel.
> 
> The kernel makefile doesn't recognize it when you change the firmware on the
> disk and still builds the old firmware file into the kernel.


$ git clone git://git.kernel.org/pub/scm/linux/kernel/git/dwmw2/linux-firmware.git
Cloning into 'linux-firmware'...
remote: Counting objects: 2210, done.
remote: Compressing objects: 100% (1102/1102), done.
remote: Total 2210 (delta 1121), reused 2114 (delta 1073)
Receiving objects: 100% (2210/2210), 39.99 MiB | 6.39 MiB/s, done.
Resolving deltas: 100% (1121/1121), done.

$ date
Wed Jul  3 08:53:01 EDT 2013

# mv /lib/firmware /lib/firmware.old
# mv linux-firmware/ /lib/firmware
# cd /lib/firmware/radeon
# md5sum CEDAR* CYPRESS_uvd.bin
2b244d41832f46382bfbb8994522dcdd  CEDAR_me.bin
23915e382ea0d2f2491a19146ca3001c  CEDAR_pfp.bin
e8770d3d588f24dc6f1a8609c9db3467  CEDAR_rlc.bin
fb23b281dcc94a035d374e709c9842bd  CYPRESS_uvd.bin

Check firmware:

/lib/firmware/radeon# md5sum CEDAR* CYPRESS_uvd.bin
2b244d41832f46382bfbb8994522dcdd  CEDAR_me.bin
23915e382ea0d2f2491a19146ca3001c  CEDAR_pfp.bin
e8770d3d588f24dc6f1a8609c9db3467  CEDAR_rlc.bin
fb23b281dcc94a035d374e709c9842bd  CYPRESS_uvd.bin

Use fresh tree:

# cp linux-3.10/.config oldconfig
# rm -rf linux-3.10
# tar jxf linux-3.10.tar.bz2
# rm linux
# ln -s linux-3.10 linux
# cp oldconfig linux/.config
# cd linux
# make oldconfig
  HOSTCC  scripts/basic/fixdep
  HOSTCC  scripts/kconfig/conf.o
  SHIPPED scripts/kconfig/zconf.tab.c
  SHIPPED scripts/kconfig/zconf.lex.c
  SHIPPED scripts/kconfig/zconf.hash.c
  HOSTCC  scripts/kconfig/zconf.tab.o
  HOSTLD  scripts/kconfig/conf
scripts/kconfig/conf --oldconfig Kconfig
warning: (DRM_RADEON && DRM_I915 && DRM_GMA500 && DRM_TILCDC && FB_BACKLIGHT && USB_APPLEDISPLAY && FB_OLPC_DCON && ASUS_LAPTOP && SONY_LAPTOP && THINKPAD_ACPI && EEEPC_LAPTOP && ACPI_CMPC && SAMSUNG_Q10) selects BACKLIGHT_CLASS_DEVICE which has unmet direct dependencies (HAS_IOMEM && BACKLIGHT_LCD_SUPPORT)
warning: (HAVE_TEXT_POKE_SMP) selects STOP_MACHINE which has unmet direct dependencies (SMP && MODULE_UNLOAD || HOTPLUG_CPU)
warning: (DRM_RADEON && DRM_I915 && DRM_GMA500 && DRM_TILCDC && FB_BACKLIGHT && USB_APPLEDISPLAY && FB_OLPC_DCON && ASUS_LAPTOP && SONY_LAPTOP && THINKPAD_ACPI && EEEPC_LAPTOP && ACPI_CMPC && SAMSUNG_Q10) selects BACKLIGHT_CLASS_DEVICE which has unmet direct dependencies (HAS_IOMEM && BACKLIGHT_LCD_SUPPORT)
warning: (HAVE_TEXT_POKE_SMP) selects STOP_MACHINE which has unmet direct dependencies (SMP && MODULE_UNLOAD || HOTPLUG_CPU)
#
# configuration written to .config
#

Compile kernel/install/reboot.

It is happy now, I've removed my distribution's firmware package as well.


[    0.817384] [drm] GART: num cpu pages 131072, num gpu pages 131072
[    0.817917] [drm] enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0
[    0.818047] [drm] Loading CEDAR Microcode
[    0.826262] [drm] PCIE GART of 512M enabled (table at 0x0000000000040000).

[    0.840416] [drm] radeon: irq initialized.
[    0.857152] [drm] ring test on 0 succeeded in 2 usecs
[    0.857267] [drm] ring test on 3 succeeded in 1 usecs
[    1.043474] [drm] ring test on 5 succeeded in 4 usecs
[    1.043537] [drm] UVD initialized successfully.
[    1.043826] [drm] ib test on ring 0 succeeded in 0 usecs
[    1.043912] [drm] ib test on ring 3 succeeded in 0 usecs
[    1.195406] [drm] ib test on ring 5 succeeded
[    1.196322] [drm] Radeon Display Connectors
[    1.196378] [drm] Connector 0:
[    1.196432] [drm]   DP-1
[    1.196486] [drm]   HPD2
[    1.196545] [drm]   DDC: 0x6460 0x6460 0x6464 0x6464 0x6468 0x6468 0x646c 0x646c
[    1.196613] [drm]   Encoders:
[    1.196667] [drm]     DFP1: INTERNAL_UNIPHY1
[    1.196723] [drm] Connector 1:
[    1.196777] [drm]   DVI-I-1
[    1.196830] [drm]   HPD4
[    1.196884] [drm]   DDC: 0x6450 0x6450 0x6454 0x6454 0x6458 0x6458 0x645c 0x645c
[    1.196952] [drm]   Encoders:
[    1.197006] [drm]     DFP2: INTERNAL_UNIPHY
[    1.197061] [drm]     CRT1: INTERNAL_KLDSCP_DAC1
[    1.197117] [drm] Connector 2:
[    1.197171] [drm]   VGA-1
[    1.197225] [drm]   DDC: 0x6430 0x6430 0x6434 0x6434 0x6438 0x6438 0x643c 0x643c
[    1.197293] [drm]   Encoders:
[    1.197347] [drm]     CRT2: INTERNAL_KLDSCP_DAC2
[    1.197445] [drm] Internal thermal controller with fan control
[    1.197570] [drm] radeon: power management initialized
[    1.291637] [drm] fb mappable at 0x3C0FE035F000
[    1.291695] [drm] vram apper at 0x3C0FE0000000
[    1.291750] [drm] size 9216000
[    1.291804] [drm] fb depth is 24
[    1.291858] [drm]    pitch is 7680


Thanks,

Justin.

Thanks for the info, closing it.