Bug 67002 - evergreen: after resume from suspend-to-ram operation is really slow with the latest DPM changes + gpu lockup
Summary: evergreen: after resume from suspend-to-ram operation is really slow with the...
Status: RESOLVED MOVED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Radeon (show other bugs)
Version: XOrg git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Default DRI bug account
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-07-17 15:10 UTC by jackdachef
Modified: 2019-11-19 08:36 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:

Attachments

Description jackdachef 2013-07-17 15:10:28 UTC 
Using the drm-code from drm-fixes-3.11 branch (http://cgit.freedesktop.org/~agd5f/linux/log/?h=drm-fixes-3.11) against a 3.10 final kernel as explained in following post (http://phoronix.com/forums/showthread.php?82357-drm-fixes-3-11-with-3-10-stable-kernel-patch-repo)

the card used is a HD5850, evergreen; ~amd64 Gentoo Linux, Intel Core i7 860, gcc 4.8.1, 2.17


I tested suspend-to-ram which in a prior state of drm-next would lock up the box (so much progress has been made - yay ! :) ):


via following command:

pm-suspend --quirk-dpms-on --quirk-s3-mode --quirk-vbe-post --quirk-vga-mode3

launched from VT1 (no X running)



first things noticable: 

it takes longer (compared to the non-DPM driver) to turn off the computer so I first thought that it had hardlocked (fans were still spinning, screen turned off, harddisks +/- off - after some time the box went to suspend/sleep)

it also takes longer to resume and for the screen to turn on




following behavior is observable after the resume from suspend-to-ram:

dmesg shows:

[  178.031724] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  178.430546] e1000e: eth0 NIC Link is Up 100 Mbps Full Duplex, Flow Control: Rx
[  178.430652] e1000e 0000:00:19.0 eth0: 10/100 speed: disabling TSO
[  179.052821] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  180.073918] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  181.095014] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  182.116110] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  182.120733] ata1: link is slow to respond, please be patient (ready=0)
[  182.122735] ata3: link is slow to respond, please be patient (ready=0)
[  182.128740] ata5: link is slow to respond, please be patient (ready=0)
[  182.130763] ata6: link is slow to respond, please be patient (ready=0)
[  183.137207] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  184.158303] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  185.179399] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  186.200496] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  186.817477] ata1: COMRESET failed (errno=-16)
[  186.819524] ata3: COMRESET failed (errno=-16)
[  186.825487] ata5: COMRESET failed (errno=-16)
[  186.827492] ata6: COMRESET failed (errno=-16)
[  187.221593] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  187.241615] [drm:r600_uvd_init] *ERROR* UVD not responding, giving up!!!
[  187.241618] [drm:evergreen_startup] *ERROR* radeon: error initializing UVD (-1).
[  187.241650] [drm] ib test on ring 0 succeeded in 0 usecs
[  187.241683] [drm] ib test on ring 3 succeeded in 1 usecs
[  187.246731] switching from power state:
[  187.246732] 	ui class: none
[  187.246733] 	internal class: boot 
[  187.246733] 	caps: video 
[  187.246734] 	uvd    vclk: 0 dclk: 0
[  187.246734] 		power level 0    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[  187.246735] 		power level 1    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[  187.246735] 		power level 2    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[  187.246736] 	status: c b 
[  187.246736] switching to power state:
[  187.246737] 	ui class: performance
[  187.246737] 	internal class: none
[  187.246738] 	caps: single_disp video 
[  187.246738] 	uvd    vclk: 0 dclk: 0
[  187.246739] 		power level 0    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[  187.246739] 		power level 1    sclk: 55000 mclk: 90000 vddc: 1038 vddci: 1100
[  187.246740] 		power level 2    sclk: 72500 mclk: 100000 vddc: 1088 vddci: 1100
[  187.246740] 	status: r 
[  188.350006] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  188.380064] ata3.00: configured for UDMA/133
[  188.380117] sd 2:0:0:0: [sdb] Starting disk
[  188.511188] ata6: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  188.523242] ata6.00: configured for UDMA/133
[  188.523276] sd 5:0:0:0: [sdd] Starting disk
[  190.500189] ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  190.513376] ata5.00: configured for UDMA/133
[  190.513455] sd 4:0:0:0: [sdc] Starting disk
[  191.462213] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  191.506493] ata1.00: configured for UDMA/133
[  191.506573] sd 0:0:0:0: [sda] Starting disk
[  192.356977] [drm:atom_op_jump] *ERROR* atombios stuck in loop for more than 5secs aborting
[  192.356978] [drm:atom_execute_table_locked] *ERROR* atombios stuck executing D3DE (len 55, WS 0, PS 0) @ 0xD401
[  192.376632] PM: resume of devices complete after 15600.251 msecs


so suspend & resume works but with caveats - now wanted to see how X is running:

launched up lightdm & logged into xfce4.10 :


compiz-fusion (0.8*)/compositing is very very slow: e.g. when grabbing a window at the top with left mouse button and trying to move it - it moves only for several millimeters on screen (22'' screen with 1920x1080 resolution) - thus unusable

when switching to xfwm4 (running Xfce4.10) moving windows around is somewhat acceptable - appears even fluid, launching up chromium and switching between tabs via ctrl + pgup/pgdown is noticably laggy

launching glxgears via gnome-terminal - it seemingly looks ok (gears smoothly turning) [FPS locked to around 60]

running glxgears via: R600_LLVM=0 vblank_mode=0 glxgears
there's noticable, "micro"-stuttering and the FPS is around 260-290 if I recall correctly

echo low > /sys/class/drm/card0/device/power_dpm_force_performance_level
or
echo high > /sys/class/drm/card0/device/power_dpm_force_performance_level

doesn't make a difference - also FPS is significantly lower compared to "normal" operation (around 10x faster, e.g. 2900-3600 FPS with the non-DPM and "high" profile)


next step was launching gnome-terminal from the below panel, then after sudo su -> dmesg
and wanted to do further tests (e.g. sensors output)
-> X crashed

[  973.217246] radeon 0000:01:00.0: GPU lockup CP stall for more than 10000msec
[  973.217254] radeon 0000:01:00.0: GPU lockup (waiting for 0x0000000000000199 last fence id 0x0000000000000198)


X (or lightdm) couldn't be launched after that anymore, VT output was still shown - so I restarted via shutdown -r now


hope that information helps in tracking down this issue

I really love this new driver - now only operation after resume needs to work also, since I used it quite a lot :)  (it does work with the non-DPM)



similar message and behavior was seen on july 3rd/4th after several hours of operation & uptime (no suspend and resume):

[  973.217246] radeon 0000:01:00.0: GPU lockup CP stall for more than 10000msec
[  973.217254] radeon 0000:01:00.0: GPU lockup (waiting for 0x0000000000000199 last fence id 0x0000000000000198)
Comment 1 Michel Dänzer 2013-07-19 12:54:22 UTC 
Can you retry with current drm-fixes? There has been a fix which might help at least for the UVD errors.
Comment 2 jackdachef 2013-07-19 15:35:50 UTC 
(In reply to comment #1)
> Can you retry with current drm-fixes? There has been a fix which might help
> at least for the UVD errors.

sure:

uvd seems to have initialized better/correctly now:

[   33.640580] [drm] radeon kernel modesetting enabled.
[   33.640868] [drm] initializing kernel modesetting (CYPRESS 0x1002:0x6899 0x1002:0x0B00).
[   33.640886] [drm] register mmio base: 0xFBCE0000
[   33.640887] [drm] register mmio size: 131072
[   33.640931] ATOM BIOS: CYPRESS
[   33.641006] radeon 0000:01:00.0: VRAM: 1024M 0x0000000000000000 - 0x000000003FFFFFFF (1024M used)
[   33.641008] radeon 0000:01:00.0: GTT: 512M 0x0000000040000000 - 0x000000005FFFFFFF
[   33.641009] [drm] Detected VRAM RAM=1024M, BAR=256M
[   33.641010] [drm] RAM width 256bits DDR
[   33.641089] [TTM] Zone  kernel: Available graphics memory: 4077798 kiB
[   33.641090] [TTM] Zone   dma32: Available graphics memory: 2097152 kiB
[   33.641091] [TTM] Initializing pool allocator
[   33.641094] [TTM] Initializing DMA pool allocator
[   33.641105] [drm] radeon: 1024M of VRAM memory ready
[   33.641106] [drm] radeon: 512M of GTT memory ready.
[   33.683758] [drm] GART: num cpu pages 131072, num gpu pages 131072
[   33.684468] [drm] enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0
[   33.684507] [drm] Loading CYPRESS Microcode
[   33.761563] [drm] PCIE GART of 512M enabled (table at 0x000000000025D000).
[   33.761693] radeon 0000:01:00.0: WB enabled
[   33.761695] radeon 0000:01:00.0: fence driver on ring 0 use gpu addr 0x0000000040000c00 and cpu addr 0xffff880234b81c00
[   33.761696] radeon 0000:01:00.0: fence driver on ring 3 use gpu addr 0x0000000040000c0c and cpu addr 0xffff880234b81c0c
[   33.761699] radeon 0000:01:00.0: fence driver on ring 5 use gpu addr 0x000000000005c418 and cpu addr 0xffffc9001351c418
[   33.761701] [drm] Supports vblank timestamp caching Rev 1 (10.10.2010).
[   33.761702] [drm] Driver supports precise vblank timestamp query.
[   33.761717] radeon 0000:01:00.0: irq 57 for MSI/MSI-X
[   33.761728] radeon 0000:01:00.0: radeon: using MSI.
[   33.761750] [drm] radeon: irq initialized.
[   33.778505] [drm] ring test on 0 succeeded in 1 usecs
[   33.778563] [drm] ring test on 3 succeeded in 1 usecs
[   33.967467] [drm] ring test on 5 succeeded in 1 usecs
[   33.967471] [drm] UVD initialized successfully.
[   33.967667] [drm] ib test on ring 0 succeeded in 0 usecs
[   33.967695] [drm] ib test on ring 3 succeeded in 0 usecs
[   34.123160] [drm] ib test on ring 5 succeeded
[   34.123540] [drm] Radeon Display Connectors
[   34.123541] [drm] Connector 0:
[   34.123542] [drm]   DP-1
[   34.123543] [drm]   HPD4
[   34.123544] [drm]   DDC: 0x6430 0x6430 0x6434 0x6434 0x6438 0x6438 0x643c 0x643c
[   34.123545] [drm]   Encoders:
[   34.123546] [drm]     DFP1: INTERNAL_UNIPHY2
[   34.123547] [drm] Connector 1:
[   34.123548] [drm]   HDMI-A-1
[   34.123548] [drm]   HPD5
[   34.123550] [drm]   DDC: 0x6460 0x6460 0x6464 0x6464 0x6468 0x6468 0x646c 0x646c
[   34.123550] [drm]   Encoders:
[   34.123551] [drm]     DFP2: INTERNAL_UNIPHY2
[   34.123552] [drm] Connector 2:
[   34.123553] [drm]   DVI-I-1
[   34.123553] [drm]   HPD1
[   34.123554] [drm]   DDC: 0x6450 0x6450 0x6454 0x6454 0x6458 0x6458 0x645c 0x645c
[   34.123555] [drm]   Encoders:
[   34.123556] [drm]     DFP3: INTERNAL_UNIPHY1
[   34.123557] [drm]     CRT2: INTERNAL_KLDSCP_DAC2
[   34.123558] [drm] Connector 3:
[   34.123558] [drm]   DVI-I-2
[   34.123559] [drm]   HPD6
[   34.123560] [drm]   DDC: 0x6470 0x6470 0x6474 0x6474 0x6478 0x6478 0x647c 0x647c
[   34.123561] [drm]   Encoders:
[   34.123562] [drm]     DFP4: INTERNAL_UNIPHY
[   34.123563] [drm]     CRT1: INTERNAL_KLDSCP_DAC1
[   34.123606] [drm] Internal thermal controller with fan control
[   34.123628] == power state 0 ==
[   34.123630] 	ui class: none
[   34.123631] 	internal class: boot 
[   34.123632] 	caps: video 
[   34.123634] 	uvd    vclk: 0 dclk: 0
[   34.123635] 		power level 0    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   34.123636] 		power level 1    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   34.123637] 		power level 2    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   34.123638] 	status: c r b 
[   34.123639] == power state 1 ==
[   34.123640] 	ui class: performance
[   34.123641] 	internal class: none
[   34.123642] 	caps: single_disp video 
[   34.123644] 	uvd    vclk: 0 dclk: 0
[   34.123645] 		power level 0    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[   34.123646] 		power level 1    sclk: 55000 mclk: 90000 vddc: 1038 vddci: 1100
[   34.123647] 		power level 2    sclk: 72500 mclk: 100000 vddc: 1088 vddci: 1100
[   34.123647] 	status: 
[   34.123648] == power state 2 ==
[   34.123649] 	ui class: none
[   34.123650] 	internal class: uvd 
[   34.123651] 	caps: video 
[   34.123652] 	uvd    vclk: 54000 dclk: 40000
[   34.123653] 		power level 0    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 1100
[   34.123654] 		power level 1    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 1100
[   34.123655] 		power level 2    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 1100
[   34.123656] 	status: 
[   34.123657] == power state 3 ==
[   34.123657] 	ui class: performance
[   34.123658] 	internal class: none
[   34.123659] 	caps: video 
[   34.123660] 	uvd    vclk: 0 dclk: 0
[   34.123661] 		power level 0    sclk: 40000 mclk: 100000 vddc: 1000 vddci: 1100
[   34.123662] 		power level 1    sclk: 55000 mclk: 100000 vddc: 1038 vddci: 1100
[   34.123663] 		power level 2    sclk: 72500 mclk: 100000 vddc: 1088 vddci: 1100
[   34.123664] 	status: 
[   34.123665] == power state 4 ==
[   34.123666] 	ui class: none
[   34.123666] 	internal class: acpi 
[   34.123667] 	caps: 
[   34.123669] 	uvd    vclk: 0 dclk: 0
[   34.123670] 		power level 0    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[   34.123671] 		power level 1    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[   34.123672] 		power level 2    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[   34.123672] 	status: 
[   34.128809] switching from power state:
[   34.128811] 	ui class: none
[   34.128812] 	internal class: boot 
[   34.128813] 	caps: video 
[   34.128814] 	uvd    vclk: 0 dclk: 0
[   34.128815] 		power level 0    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   34.128816] 		power level 1    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   34.128817] 		power level 2    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   34.128818] 	status: c b 
[   34.128819] switching to power state:
[   34.128820] 	ui class: performance
[   34.128821] 	internal class: none
[   34.128822] 	caps: single_disp video 
[   34.128823] 	uvd    vclk: 0 dclk: 0
[   34.128824] 		power level 0    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[   34.128825] 		power level 1    sclk: 55000 mclk: 90000 vddc: 1038 vddci: 1100
[   34.128826] 		power level 2    sclk: 72500 mclk: 100000 vddc: 1088 vddci: 1100
[   34.128827] 	status: r 
[   34.134851] [drm] radeon: dpm initialized
[   34.214869] [drm] fb mappable at 0xD035F000
[   34.214871] [drm] vram apper at 0xD0000000
[   34.214873] [drm] size 8294400
[   34.214873] [drm] fb depth is 24
[   34.214874] [drm]    pitch is 7680
[   34.214909] fbcon: radeondrmfb (fb0) is primary device
[   34.239815] Console: switching to colour frame buffer device 240x67
[   34.243830] radeon 0000:01:00.0: fb0: radeondrmfb frame buffer device
[   34.243831] radeon 0000:01:00.0: registered panic notifier
[   34.243834] [drm] Initialized radeon 2.34.0 20080528 for 0000:01:00.0 on minor 0


I gave it some testing (several videos played back, mkv, Avatar and Tron Legacy for some minutes) via vdpau (smplayer), 

with

echo auto > /sys/class/drm/card0/device/power_dpm_force_performance_level

and

echo low > /sys/class/drm/card0/device/power_dpm_force_performance_level


used the composited desktop, and did some web-browsing with chromium (pepperflash) and firefox 64bit (adobe-flash via vdpau) and watched some youtube videos


looked stable to me,

[note (if I forget it to poster later in this entry): chromium is using gpu compositing - which might trigger and might have triggered the gpu lock & hardlock this time]


so far - so good


for data integrity & safety reasons I quit X and tried out suspend-to-ram and resume:

from VT1:

pm-suspend --quirk-dpms-on --quirk-s3-mode --quirk-vbe-post --quirk-vga-mode3 & vlock


delayed suspend switch and wakeup (as usual & observed from last time):


output from dmesg:

[ 4487.786933] [drm] enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0
[ 4487.789211] [drm] PCIE GART of 512M enabled (table at 0x000000000025D000).
[ 4487.789561] radeon 0000:01:00.0: WB enabled
[ 4487.789563] radeon 0000:01:00.0: fence driver on ring 0 use gpu addr 0x0000000040000c00 and cpu addr 0xffff880234b81c00
[ 4487.789564] radeon 0000:01:00.0: fence driver on ring 3 use gpu addr 0x0000000040000c0c and cpu addr 0xffff880234b81c0c
[ 4487.790405] radeon 0000:01:00.0: fence driver on ring 5 use gpu addr 0x000000000005c418 and cpu addr 0xffffc9001351c418
[ 4487.806998] [drm] ring test on 0 succeeded in 1 usecs
[ 4487.807056] [drm] ring test on 3 succeeded in 1 usecs
[ 4488.017296] ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
[ 4488.021126] ata2.00: configured for UDMA/100
[ 4488.022363] ata4: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
[ 4488.028358] ata8: SATA link down (SStatus 0 SControl 300)
[ 4488.028390] ata7: SATA link down (SStatus 0 SControl 300)
[ 4488.050196] ata4.00: configured for UDMA/100
[ 4488.984442] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[ 4489.333196] e1000e: eth0 NIC Link is Up 100 Mbps Full Duplex, Flow Control: Rx
[ 4489.333302] e1000e 0000:00:19.0 eth0: 10/100 speed: disabling TSO
[ 4490.005567] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[ 4491.026691] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[ 4492.047816] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[ 4493.068940] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[ 4493.074353] ata1: link is slow to respond, please be patient (ready=0)
[ 4493.078357] ata3: link is slow to respond, please be patient (ready=0)
[ 4493.080359] ata5: link is slow to respond, please be patient (ready=0)
[ 4493.082360] ata6: link is slow to respond, please be patient (ready=0)
[ 4494.090065] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[ 4495.111189] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[ 4496.132314] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[ 4497.153481] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[ 4497.771073] ata1: COMRESET failed (errno=-16)
[ 4497.775081] ata3: COMRESET failed (errno=-16)
[ 4497.777085] ata5: COMRESET failed (errno=-16)
[ 4497.779160] ata6: COMRESET failed (errno=-16)
[ 4498.174616] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[ 4498.194639] [drm:r600_uvd_init] *ERROR* UVD not responding, giving up!!!
[ 4498.194642] [drm:evergreen_startup] *ERROR* radeon: error initializing UVD (-1).
[ 4498.194675] [drm] ib test on ring 0 succeeded in 0 usecs
[ 4498.194707] [drm] ib test on ring 3 succeeded in 1 usecs
[ 4498.199776] switching from power state:
[ 4498.199777] 	ui class: none
[ 4498.199778] 	internal class: boot 
[ 4498.199779] 	caps: video 
[ 4498.199779] 	uvd    vclk: 0 dclk: 0
[ 4498.199780] 		power level 0    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[ 4498.199781] 		power level 1    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[ 4498.199781] 		power level 2    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[ 4498.199782] 	status: c b 
[ 4498.199782] switching to power state:
[ 4498.199782] 	ui class: performance
[ 4498.199783] 	internal class: none
[ 4498.199784] 	caps: single_disp video 
[ 4498.199784] 	uvd    vclk: 0 dclk: 0
[ 4498.199785] 		power level 0    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[ 4498.199785] 		power level 1    sclk: 55000 mclk: 90000 vddc: 1038 vddci: 1100
[ 4498.199786] 		power level 2    sclk: 72500 mclk: 100000 vddc: 1088 vddci: 1100
[ 4498.199786] 	status: r 
[ 4499.462782] ata6: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 4499.475045] ata6.00: configured for UDMA/133
[ 4499.485899] sd 5:0:0:0: [sdd] Starting disk
[ 4499.662989] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 4499.692137] ata3.00: configured for UDMA/133
[ 4499.703084] sd 2:0:0:0: [sdb] Starting disk
[ 4501.502837] ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 4501.523748] ata5.00: configured for UDMA/133
[ 4501.533954] sd 4:0:0:0: [sdc] Starting disk
[ 4502.364726] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 4502.407461] ata1.00: configured for UDMA/133
[ 4502.417816] sd 0:0:0:0: [sda] Starting disk
[ 4503.310597] [drm:atom_op_jump] *ERROR* atombios stuck in loop for more than 5secs aborting
[ 4503.310599] [drm:atom_execute_table_locked] *ERROR* atombios stuck executing D3DE (len 55, WS 0, PS 0) @ 0xD401
[ 4503.330289] PM: resume of devices complete after 15602.288 msecs


<== note the error messages in dmesg output after resume about atombios & uvd

several times typed in dmesg to make sure that showing dmesg-output doesn't lead to X crash, gpu lock, etc.


following things could be observed:

*) dmesg didn't lead to gpu lock or X crash anymore
*) composited desktop (compiz-fusion) was running fluid after the resume: exposé-like feature worked, zooming in, zooming out, invert screen/colors - worked fine so far (can't remember if I minimized windows - at least maximizing also worked)

*) fired up chromium, did some little browsing and wanted to surf over to this bug-entry to report - then it hardlocked

the box wouldn't react to magic sysrq key, toggling num lock, caps lock, etc. - no sign of life from the keyboard either

fans on the gpu didn't spin up faster (e.g. like with old non-DPM driver, static profiles when it locked up), fan of the PC/power supply or cpu fan also didn't seem to spin up faster)


currently I'm more worried about it locking up hard than the UVD issues (I'm not fully trusting brfs on my /home partition yet - otherwise I would try suspending from within X & Xfce)


hope you can fix the issue with atombios looping & *ERROR* atombios stuck executing D3DE (len 55, WS 0, PS 0) @ 0xD401 - that currently worries me the most, as far as I remember that didn't occur with the old driver static implementation (might have to investigate this ...)



looking through this text again: could gpu compositing from chromium be the reason the it hardlocks ?

figuratively speaking: that finishes off the gpu after these errors ? or is it completely unrelated


anyway - thanks for your continued efforts :)
Comment 3 jackdachef 2013-07-19 15:37:50 UTC 
at least according to the kernel logs the atombios error seems to have been introduced, is new with the new DPM code (running the new firmware both with old and new driver)
Comment 4 Alex Deucher 2013-07-19 16:51:05 UTC 
The atombios messages are due to bug 67043.  You can revert the patch referenced in that bug to fix that.  danvet supposed has a v2 of that patch which doesn't break resume.

Also, does it work any better if you drop all the pm-suspend quirks?  You shouldn't need them for KMS.
Comment 5 jackdachef 2013-07-19 17:26:13 UTC 
(In reply to comment #4)
> The atombios messages are due to bug 67043.  You can revert the patch
> referenced in that bug to fix that.  danvet supposed has a v2 of that patch
> which doesn't break resume.
> 
> Also, does it work any better if you drop all the pm-suspend quirks?  You
> shouldn't need them for KMS.

ok, so reverting the original patch and applying the attached one is the correct order ? or shall it be applied to current source ? - (probably the former) - will try


I hesitantly will give it a try without the quirks - up until now the box always would hardlock without those quirks - both on xf86-video-ati or fglrx

but ... it might work already without quirks since via Xfce it suspended and resumed successfully to ram 2-3 times already from the panel menu with non(new)-DPM so far


thanks !
Comment 6 Alex Deucher 2013-07-19 17:44:45 UTC 
(In reply to comment #5)
> (In reply to comment #4)
> > The atombios messages are due to bug 67043.  You can revert the patch
> > referenced in that bug to fix that.  danvet supposed has a v2 of that patch
> > which doesn't break resume.
> > 
> > Also, does it work any better if you drop all the pm-suspend quirks?  You
> > shouldn't need them for KMS.
> 
> ok, so reverting the original patch and applying the attached one is the
> correct order ? or shall it be applied to current source ? - (probably the
> former) - will try
> 

Either revert the original or apply the new patch, but don't do both.  The new patch contains the revert.
Comment 7 jackdachef 2013-07-20 15:42:51 UTC 
(In reply to comment #6)
> (In reply to comment #5)
> > (In reply to comment #4)
> > > The atombios messages are due to bug 67043.  You can revert the patch
> > > referenced in that bug to fix that.  danvet supposed has a v2 of that patch
> > > which doesn't break resume.
> > > 
> > > Also, does it work any better if you drop all the pm-suspend quirks?  You
> > > shouldn't need them for KMS.
> > 
> > ok, so reverting the original patch and applying the attached one is the
> > correct order ? or shall it be applied to current source ? - (probably the
> > former) - will try
> > 
> 
> Either revert the original or apply the new patch, but don't do both.  The
> new patch contains the revert.


ok, applied the patch - this is output from during boot

[   32.664459] [drm] radeon kernel modesetting enabled.
[   32.664811] [drm] initializing kernel modesetting (CYPRESS 0x1002:0x6899 0x1002:0x0B00).
[   32.664830] [drm] register mmio base: 0xFBCE0000
[   32.664832] [drm] register mmio size: 131072
[   32.664880] ATOM BIOS: CYPRESS
[   32.664954] radeon 0000:01:00.0: VRAM: 1024M 0x0000000000000000 - 0x000000003FFFFFFF (1024M used)
[   32.664955] radeon 0000:01:00.0: GTT: 512M 0x0000000040000000 - 0x000000005FFFFFFF
[   32.664956] [drm] Detected VRAM RAM=1024M, BAR=256M
[   32.664957] [drm] RAM width 256bits DDR
[   32.665054] [TTM] Zone  kernel: Available graphics memory: 4077798 kiB
[   32.665055] [TTM] Zone   dma32: Available graphics memory: 2097152 kiB
[   32.665056] [TTM] Initializing pool allocator
[   32.665058] [TTM] Initializing DMA pool allocator
[   32.665074] [drm] radeon: 1024M of VRAM memory ready
[   32.665075] [drm] radeon: 512M of GTT memory ready.
[   32.690308] [drm] GART: num cpu pages 131072, num gpu pages 131072
[   32.690956] [drm] enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0
[   32.690994] [drm] Loading CYPRESS Microcode
[   32.777410] [drm] PCIE GART of 512M enabled (table at 0x000000000025D000).
[   32.777562] radeon 0000:01:00.0: WB enabled
[   32.777566] radeon 0000:01:00.0: fence driver on ring 0 use gpu addr 0x0000000040000c00 and cpu addr 0xffff88023043cc00
[   32.777568] radeon 0000:01:00.0: fence driver on ring 3 use gpu addr 0x0000000040000c0c and cpu addr 0xffff88023043cc0c
[   32.777571] radeon 0000:01:00.0: fence driver on ring 5 use gpu addr 0x000000000005c418 and cpu addr 0xffffc90010d9c418
[   32.777573] [drm] Supports vblank timestamp caching Rev 1 (10.10.2010).
[   32.777574] [drm] Driver supports precise vblank timestamp query.
[   32.777594] radeon 0000:01:00.0: irq 56 for MSI/MSI-X
[   32.777606] radeon 0000:01:00.0: radeon: using MSI.
[   32.777633] [drm] radeon: irq initialized.
[   32.786487] input: HDA Intel MID Front Headphone as /devices/pci0000:00/0000:00:1b.0/sound/card0/input4
[   32.786609] input: HDA Intel MID Line Out CLFE as /devices/pci0000:00/0000:00:1b.0/sound/card0/input5
[   32.786689] input: HDA Intel MID Line Out Surround as /devices/pci0000:00/0000:00:1b.0/sound/card0/input6
[   32.786758] input: HDA Intel MID Line Out Front as /devices/pci0000:00/0000:00:1b.0/sound/card0/input7
[   32.786826] input: HDA Intel MID Line as /devices/pci0000:00/0000:00:1b.0/sound/card0/input8
[   32.786895] input: HDA Intel MID Rear Mic as /devices/pci0000:00/0000:00:1b.0/sound/card0/input9
[   32.786963] input: HDA Intel MID Front Mic as /devices/pci0000:00/0000:00:1b.0/sound/card0/input10
[   32.794513] [drm] ring test on 0 succeeded in 1 usecs
[   32.794572] [drm] ring test on 3 succeeded in 1 usecs
[   32.987493] [drm] ring test on 5 succeeded in 1 usecs
[   32.987498] [drm] UVD initialized successfully.
[   32.987742] [drm] ib test on ring 0 succeeded in 0 usecs
[   32.987773] [drm] ib test on ring 3 succeeded in 0 usecs
[   33.142552] [drm] ib test on ring 5 succeeded
[   33.143060] [drm] Radeon Display Connectors
[   33.143062] [drm] Connector 0:
[   33.143063] [drm]   DP-1
[   33.143064] [drm]   HPD4
[   33.143065] [drm]   DDC: 0x6430 0x6430 0x6434 0x6434 0x6438 0x6438 0x643c 0x643c
[   33.143066] [drm]   Encoders:
[   33.143067] [drm]     DFP1: INTERNAL_UNIPHY2
[   33.143068] [drm] Connector 1:
[   33.143069] [drm]   HDMI-A-1
[   33.143070] [drm]   HPD5
[   33.143071] [drm]   DDC: 0x6460 0x6460 0x6464 0x6464 0x6468 0x6468 0x646c 0x646c
[   33.143071] [drm]   Encoders:
[   33.143072] [drm]     DFP2: INTERNAL_UNIPHY2
[   33.143073] [drm] Connector 2:
[   33.143074] [drm]   DVI-I-1
[   33.143074] [drm]   HPD1
[   33.143076] [drm]   DDC: 0x6450 0x6450 0x6454 0x6454 0x6458 0x6458 0x645c 0x645c
[   33.143076] [drm]   Encoders:
[   33.143077] [drm]     DFP3: INTERNAL_UNIPHY1
[   33.143078] [drm]     CRT2: INTERNAL_KLDSCP_DAC2
[   33.143079] [drm] Connector 3:
[   33.143080] [drm]   DVI-I-2
[   33.143080] [drm]   HPD6
[   33.143082] [drm]   DDC: 0x6470 0x6470 0x6474 0x6474 0x6478 0x6478 0x647c 0x647c
[   33.143082] [drm]   Encoders:
[   33.143083] [drm]     DFP4: INTERNAL_UNIPHY
[   33.143084] [drm]     CRT1: INTERNAL_KLDSCP_DAC1
[   33.143130] [drm] Internal thermal controller with fan control
[   33.143151] == power state 0 ==
[   33.143152] 	ui class: none
[   33.143153] 	internal class: boot 
[   33.143155] 	caps: video 
[   33.143156] 	uvd    vclk: 0 dclk: 0
[   33.143157] 		power level 0    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   33.143158] 		power level 1    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   33.143159] 		power level 2    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   33.143159] 	status: c r b 
[   33.143161] == power state 1 ==
[   33.143161] 	ui class: performance
[   33.143162] 	internal class: none
[   33.143163] 	caps: single_disp video 
[   33.143164] 	uvd    vclk: 0 dclk: 0
[   33.143165] 		power level 0    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[   33.143166] 		power level 1    sclk: 55000 mclk: 90000 vddc: 1038 vddci: 1100
[   33.143167] 		power level 2    sclk: 72500 mclk: 100000 vddc: 1088 vddci: 1100
[   33.143168] 	status: 
[   33.143168] == power state 2 ==
[   33.143169] 	ui class: none
[   33.143170] 	internal class: uvd 
[   33.143171] 	caps: video 
[   33.143172] 	uvd    vclk: 54000 dclk: 40000
[   33.143173] 		power level 0    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 1100
[   33.143174] 		power level 1    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 1100
[   33.143175] 		power level 2    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 1100
[   33.143175] 	status: 
[   33.143176] == power state 3 ==
[   33.143176] 	ui class: performance
[   33.143177] 	internal class: none
[   33.143178] 	caps: video 
[   33.143179] 	uvd    vclk: 0 dclk: 0
[   33.143180] 		power level 0    sclk: 40000 mclk: 100000 vddc: 1000 vddci: 1100
[   33.143181] 		power level 1    sclk: 55000 mclk: 100000 vddc: 1038 vddci: 1100
[   33.143182] 		power level 2    sclk: 72500 mclk: 100000 vddc: 1088 vddci: 1100
[   33.143182] 	status: 
[   33.143183] == power state 4 ==
[   33.143184] 	ui class: none
[   33.143184] 	internal class: acpi 
[   33.143185] 	caps: 
[   33.143186] 	uvd    vclk: 0 dclk: 0
[   33.143187] 		power level 0    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[   33.143188] 		power level 1    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[   33.143189] 		power level 2    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[   33.143190] 	status: 
[   33.148444] switching from power state:
[   33.148445] 	ui class: none
[   33.148446] 	internal class: boot 
[   33.148447] 	caps: video 
[   33.148448] 	uvd    vclk: 0 dclk: 0
[   33.148449] 		power level 0    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   33.148450] 		power level 1    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   33.148451] 		power level 2    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[   33.148451] 	status: c b 
[   33.148452] switching to power state:
[   33.148453] 	ui class: performance
[   33.148454] 	internal class: none
[   33.148455] 	caps: single_disp video 
[   33.148456] 	uvd    vclk: 0 dclk: 0
[   33.148457] 		power level 0    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[   33.148458] 		power level 1    sclk: 55000 mclk: 90000 vddc: 1038 vddci: 1100
[   33.148459] 		power level 2    sclk: 72500 mclk: 100000 vddc: 1088 vddci: 1100
[   33.148459] 	status: r 
[   33.154479] [drm] radeon: dpm initialized
[   33.234644] [drm] fb mappable at 0xD035F000
[   33.234646] [drm] vram apper at 0xD0000000
[   33.234647] [drm] size 8294400
[   33.234647] [drm] fb depth is 24
[   33.234648] [drm]    pitch is 7680
[   33.234758] fbcon: radeondrmfb (fb0) is primary device
[   33.259491] Console: switching to colour frame buffer device 240x67
[   33.263611] radeon 0000:01:00.0: fb0: radeondrmfb frame buffer device
[   33.263612] radeon 0000:01:00.0: registered panic notifier
[   33.263615] [drm] Initialized radeon 2.34.0 20080528 for 0000:01:00.0 on minor 0


and tried to suspend-to-ram without quirks (only via pm-suspend from VT1 - no X yet) and it worked,resume also worked - so no hardlock or crash in that process [yet]

but after resume, starting up X (lightdm) and logging into Xfce 

the composited desktop was very very slow again (I tested "inverted colors" effect and tab-switching between apps which worked)

output of dmesg also was displayed

switched from compiz(fusion) to xfwm4 - moving the gnome-terminal window via pressed left mouse-button & mouse was laggy and with delay / trail

output of dmesg still was displayed without crash or hardlock

switched back to compiz-fusion (switching is done via fusion-icon)


now I wanted to post some preliminary info here - so fired up firefox and screen went black

--> totally black, after several seconds and magic sysrq key + r & switching to VT things got displayed again



dmesg from after resume:



[  141.323468] PM: noirq suspend of devices complete after 43.787 msecs
[  141.323844] ACPI: Preparing to enter system sleep state S3
[  141.324917] PM: Saving platform NVS memory
[  141.325103] Disabling non-boot CPUs ...
[  141.340593] Broke affinity for irq 41
[  141.341717] smpboot: CPU 1 is now offline
[  141.360603] Broke affinity for irq 42
[  141.361724] smpboot: CPU 2 is now offline
[  141.380611] Broke affinity for irq 43
[  141.381959] smpboot: CPU 3 is now offline
[  141.400604] Broke affinity for irq 44
[  141.401720] smpboot: CPU 4 is now offline
[  141.421728] smpboot: CPU 5 is now offline
[  141.441740] smpboot: CPU 6 is now offline
[  141.460798] smpboot: CPU 7 is now offline
[  141.461880] ACPI: Low-level resume complete
[  141.461911] PM: Restoring platform NVS memory
[  141.462213] microcode: CPU0 sig=0x106e5, pf=0x2, revision=0x5
[  141.462277] Enabling non-boot CPUs ...
[  141.462320] smpboot: Booting Node 0 Processor 1 APIC 0x2
[  141.472917] CPU1 microcode updated early to revision 0x5, date = 2011-09-01
[  141.475545] microcode: CPU1 sig=0x106e5, pf=0x2, revision=0x5
[  141.506986] hpet: hpet3 irq 41 for MSI
[  141.506993] CPU1 is up
[  141.507014] smpboot: Booting Node 0 Processor 2 APIC 0x4
[  141.517596] CPU2 microcode updated early to revision 0x5, date = 2011-09-01
[  141.520352] microcode: CPU2 sig=0x106e5, pf=0x2, revision=0x5
[  141.554033] hpet: hpet4 irq 42 for MSI
[  141.554039] CPU2 is up
[  141.554059] smpboot: Booting Node 0 Processor 3 APIC 0x6
[  141.564641] CPU3 microcode updated early to revision 0x5, date = 2011-09-01
[  141.567394] microcode: CPU3 sig=0x106e5, pf=0x2, revision=0x5
[  141.602082] hpet: hpet5 irq 43 for MSI
[  141.602089] CPU3 is up
[  141.602109] smpboot: Booting Node 0 Processor 4 APIC 0x1
[  141.612682] CPU4 microcode updated early to revision 0x5, date = 2011-09-01
[  141.615460] microcode: CPU4 sig=0x106e5, pf=0x2, revision=0x5
[  141.650130] hpet: hpet6 irq 44 for MSI
[  141.650136] CPU4 is up
[  141.650153] smpboot: Booting Node 0 Processor 5 APIC 0x3
[  141.663494] microcode: CPU5 sig=0x106e5, pf=0x2, revision=0x5
[  141.698142] CPU5 is up
[  141.698160] smpboot: Booting Node 0 Processor 6 APIC 0x5
[  141.711514] microcode: CPU6 sig=0x106e5, pf=0x2, revision=0x5
[  141.746187] CPU6 is up
[  141.746205] smpboot: Booting Node 0 Processor 7 APIC 0x7
[  141.759565] microcode: CPU7 sig=0x106e5, pf=0x2, revision=0x5
[  141.794241] CPU7 is up
[  141.798872] ACPI: Waking up from system sleep state S3
[  141.821165] ehci-pci 0000:00:1a.0: System wakeup disabled by ACPI
[  141.843184] ehci-pci 0000:00:1d.0: System wakeup disabled by ACPI
[  141.959353] pcieport 0000:00:1c.2: System wakeup disabled by ACPI
[  141.959528] PM: noirq resume of devices complete after 159.807 msecs
[  141.959652] PM: early resume of devices complete after 0.085 msecs
[  141.959673] e1000e 0000:00:19.0: setting latency timer to 64
[  141.959688] pci 0000:00:1e.0: setting latency timer to 64
[  141.959695] ahci 0000:00:1f.2: setting latency timer to 64
[  141.959732] e1000e 0000:00:19.0: irq 54 for MSI/MSI-X
[  141.959745] usb usb3: root hub lost power or was reset
[  141.959746] usb usb4: root hub lost power or was reset
[  141.959816] ehci-pci 0000:00:1a.0: setting latency timer to 64
[  141.959836] usb usb1: usb resume
[  141.959838] ehci-pci 0000:00:1a.0: resume root hub
[  141.959935] snd_hda_intel 0000:00:1b.0: irq 46 for MSI/MSI-X
[  141.959978] ehci-pci 0000:00:1d.0: setting latency timer to 64
[  141.959979] xhci_hcd 0000:04:00.0: irq 47 for MSI/MSI-X
[  141.959984] xhci_hcd 0000:04:00.0: irq 48 for MSI/MSI-X
[  141.959989] xhci_hcd 0000:04:00.0: irq 49 for MSI/MSI-X
[  141.959993] usb usb2: usb resume
[  141.959994] xhci_hcd 0000:04:00.0: irq 50 for MSI/MSI-X
[  141.959996] ehci-pci 0000:00:1d.0: resume root hub
[  141.959999] xhci_hcd 0000:04:00.0: irq 51 for MSI/MSI-X
[  141.960003] xhci_hcd 0000:04:00.0: irq 52 for MSI/MSI-X
[  141.960008] xhci_hcd 0000:04:00.0: irq 53 for MSI/MSI-X
[  141.960011] xhci_hcd 0000:04:00.0: irq 55 for MSI/MSI-X
[  141.960055] pata_jmicron 0000:03:00.1: setting latency timer to 64
[  141.960062] i8042 kbd 00:05: System wakeup disabled by ACPI
[  141.960063] snd_hda_intel 0000:01:00.1: irq 57 for MSI/MSI-X
[  141.963219] usb usb3: usb resume
[  141.963300] usb usb4: usb resume
[  141.963307] hub 3-0:1.0: hub_reset_resume
[  141.963308] hub 3-0:1.0: enabling power on all ports
[  141.963328] hub 4-0:1.0: hub_reset_resume
[  141.963330] hub 4-0:1.0: enabling power on all ports
[  141.991556] hub 2-0:1.0: hub_resume
[  141.991563] hub 1-0:1.0: hub_resume
[  141.991565] hub 2-0:1.0: port 1: status 0503 change 0000
[  141.991571] hub 1-0:1.0: port 1: status 0503 change 0000
[  141.991603] usb 2-1: finish resume
[  141.991608] usb 1-1: finish resume
[  141.991732] hub 2-1:1.0: hub_resume
[  141.991743] hub 1-1:1.0: hub_resume
[  141.991970] hub 2-1:1.0: port 2: status 0103 change 0000
[  141.992287] hub 1-1:1.0: port 5: status 0503 change 0000
[  141.992453] ehci-pci 0000:00:1a.0: reused qh ffff8802362a2280 schedule
[  141.992455] usb 1-1: link qh256-0001/ffff8802362a2280 start 1 [1/0 us]
[  141.992530] hub 2-1:1.0: port 7: status 0501 change 0000
[  141.992724] ehci-pci 0000:00:1d.0: reused qh ffff8802362a2880 schedule
[  141.992728] usb 2-1: link qh256-0001/ffff8802362a2880 start 1 [1/0 us]
[  141.992729] usb 1-1.5: finish resume
[  141.992972] usb 2-1.2: finish resume
[  141.992978] hub 1-1.5:1.0: hub_resume
[  141.993263] ehci-pci 0000:00:1d.0: reused qh ffff880236a39600 schedule
[  141.993265] usb 2-1.2: link qh2-0e01/ffff880236a39600 start 0 [1/3 us]
[  141.993318] hub 1-1.5:1.0: port 1: status 0103 change 0000
[  141.994099] ehci-pci 0000:00:1a.0: reused qh ffff8802362a2480 schedule
[  141.994104] usb 1-1.5: link qh256-0001/ffff8802362a2480 start 2 [1/0 us]
[  141.994603] usb 1-1.5.1: finish resume
[  141.995225] ehci-pci 0000:00:1a.0: reused qh ffff880231d1bd80 schedule
[  141.995227] usb 1-1.5.1: link qh1-0e01/ffff880231d1bd80 start 0 [1/2 us]
[  141.995292] ehci-pci 0000:00:1a.0: reused qh ffff880236979180 schedule
[  141.995294] usb 1-1.5.1: link qh1-0e01/ffff880236979180 start 0 [1/2 us]
[  142.033104] [drm] enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0
[  142.035367] [drm] PCIE GART of 512M enabled (table at 0x000000000025D000).
[  142.035505] radeon 0000:01:00.0: WB enabled
[  142.035507] radeon 0000:01:00.0: fence driver on ring 0 use gpu addr 0x0000000040000c00 and cpu addr 0xffff88023043cc00
[  142.035508] radeon 0000:01:00.0: fence driver on ring 3 use gpu addr 0x0000000040000c0c and cpu addr 0xffff88023043cc0c
[  142.036337] radeon 0000:01:00.0: fence driver on ring 5 use gpu addr 0x000000000005c418 and cpu addr 0xffffc90010d9c418
[  142.052912] [drm] ring test on 0 succeeded in 1 usecs
[  142.052970] [drm] ring test on 3 succeeded in 1 usecs
[  142.266592] ata4: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
[  142.270595] ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
[  142.274144] ata2.00: configured for UDMA/100
[  142.277610] ata8: SATA link down (SStatus 0 SControl 300)
[  142.277699] ata7: SATA link down (SStatus 0 SControl 300)
[  142.294309] ata4.00: configured for UDMA/100
[  143.230272] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  143.681632] e1000e: eth0 NIC Link is Up 100 Mbps Full Duplex, Flow Control: Rx
[  143.681738] e1000e 0000:00:19.0 eth0: 10/100 speed: disabling TSO
[  144.251358] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  145.272446] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  146.293535] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  147.314624] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  147.318664] ata1: link is slow to respond, please be patient (ready=0)
[  147.322668] ata5: link is slow to respond, please be patient (ready=0)
[  147.325732] ata6: link is slow to respond, please be patient (ready=0)
[  147.328674] ata3: link is slow to respond, please be patient (ready=0)
[  148.335792] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  149.356881] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  150.377968] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  151.399055] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  152.015389] ata1: COMRESET failed (errno=-16)
[  152.019392] ata5: COMRESET failed (errno=-16)
[  152.022458] ata6: COMRESET failed (errno=-16)
[  152.025398] ata3: COMRESET failed (errno=-16)
[  152.420143] [drm:r600_uvd_init] *ERROR* UVD not responding, trying to reset the VCPU!!!
[  152.440166] [drm:r600_uvd_init] *ERROR* UVD not responding, giving up!!!
[  152.440169] [drm:evergreen_startup] *ERROR* radeon: error initializing UVD (-1).
[  152.440201] [drm] ib test on ring 0 succeeded in 0 usecs
[  152.440234] [drm] ib test on ring 3 succeeded in 1 usecs
[  152.445310] switching from power state:
[  152.445311] 	ui class: none
[  152.445312] 	internal class: boot 
[  152.445312] 	caps: video 
[  152.445313] 	uvd    vclk: 0 dclk: 0
[  152.445314] 		power level 0    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[  152.445314] 		power level 1    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[  152.445315] 		power level 2    sclk: 40000 mclk: 90000 vddc: 1000 vddci: 0
[  152.445316] 	status: c b 
[  152.445316] switching to power state:
[  152.445316] 	ui class: performance
[  152.445317] 	internal class: none
[  152.445318] 	caps: single_disp video 
[  152.445318] 	uvd    vclk: 0 dclk: 0
[  152.445319] 		power level 0    sclk: 15700 mclk: 30000 vddc: 950 vddci: 1100
[  152.445319] 		power level 1    sclk: 55000 mclk: 90000 vddc: 1038 vddci: 1100
[  152.445320] 		power level 2    sclk: 72500 mclk: 100000 vddc: 1088 vddci: 1100
[  152.445320] 	status: r 





after X crash / gpu hang:

[  855.159690] radeon 0000:01:00.0: GPU lockup CP stall for more than 10000msec
[  855.159697] radeon 0000:01:00.0: GPU lockup (waiting for 0x00000000000000eb last fence id 0x00000000000000ea)
[  870.270720] SysRq : Keyboard mode set to system default
[  896.490743] reiserfs: enabling write barrier flush mode

(dmesg output and stuff is saved on /boot, a reiserfs partition)
Comment 8 jackdachef 2013-07-24 10:07:47 UTC 
ok,so the atombios error is fixed now

what would the next step be to gradually fix this ? (so the uvd error after resume & gpu lock/hang)

switch to gcc 4.7* ? would that make a change ?

the gcc 4.8*/4.8.1* related issues are already solved, right ?



do you need more info ?
Comment 9 Andre Heider 2013-07-24 14:45:11 UTC 
Maybe unrelated to this bug report, but with today's Linus' master (04012e3076943f17c8cfc77a196701c8f0b8964f) the only suspend/resume issue left is that I get:

[drm:rv770_stop_dpm] *ERROR* Could not force DPM to low.

in dmesg upon resume.

That is before the 'ring' and 'uvd' messages. It seems harmless, I didn't notice any problems. It can switch power levels later on, play videos via uvd... it just seems to try to do something with dpm too early in the resume process.
Comment 10 jackdachef 2013-07-24 20:05:31 UTC 
(In reply to comment #9)
> Maybe unrelated to this bug report, but with today's Linus' master
> (04012e3076943f17c8cfc77a196701c8f0b8964f) the only suspend/resume issue
> left is that I get:
> 
> [drm:rv770_stop_dpm] *ERROR* Could not force DPM to low.
> 
> in dmesg upon resume.
> 
> That is before the 'ring' and 'uvd' messages. It seems harmless, I didn't
> notice any problems. It can switch power levels later on, play videos via
> uvd... it just seems to try to do something with dpm too early in the resume
> process.

indeed - the recent changes made a difference

I'm currently posting from the kernel running from resumed suspend-to-ram

the only changes were

that I set to "auto"

 echo auto > /sys/class/drm/card0/device/power_dpm_force_performance_level

and am running that new kernel posted over at phoronix forums


probably will try out with forced "low" profile/performance_level in the weekend and see if I get a similar message to yours

now I need the box up and running for work

will continue using it and see whether something strange occurs

if that's the case I'll post it here

otherwise it's running fine so far ...


Thanks a lot to everyone for the fixes & work ! 


on a side-note: the hardlocks I experienced in the past might have been due to BFQ & some exotic setup/tweaks/additional patches that I'm using

so perhaps no hardlocks with the DPM code so far :)
Comment 11 jackdachef 2013-07-24 21:30:54 UTC 
I had to restart due to some other issue

and it wouldn't "successfully" resume anymore:

the lag/slowdown was there again


forcing performance_level to "low" or "auto" didn't seem to make a difference


so I'll try a 3.11-rc2* kernel in the weekend then - provided of course that it runs without other problems


could it be related to reboot from Windows (windows 8 -> linux) that resume worked properly ?
Comment 12 jackdachef 2013-07-27 21:24:33 UTC 
(In reply to comment #11)
> I had to restart due to some other issue
> 
> and it wouldn't "successfully" resume anymore:
> 
> the lag/slowdown was there again
> 
> 
> forcing performance_level to "low" or "auto" didn't seem to make a difference
> 
> 
> so I'll try a 3.11-rc2* kernel in the weekend then - provided of course that
> it runs without other problems
> 
> 
> could it be related to reboot from Windows (windows 8 -> linux) that resume
> worked properly ?

had the chance to test it on latest 3.11-rc2-something kernel with most recent drm-fixes

and exactly the same:

after resume (from X - suspended from the Xfce menu at the right corner) the composited desktop is really really slow + the following behavior (I try to make the description of the "slowness" as clear as possible so you can draw conclusions easily) 

and after some time of slow browsing via chromium, dmesg-output without crashing, etc. - surfing over and attempting to type in some data into this bug entry - X crashed and xdm/lightdm attempted several times to reload until screen stayed black with only one cursor point blinking at the left upper corner - attempted to switch to VT1, VT2, etc. didn't work, magic sysrq + r and then attempt to switching to VT (which had worked with patched up 3.10 kernel) for a short moment showed the logon line on VT1 and then screen stayed black, tried that several times - then it stayed completely black only with the cursor at the left upper corner blinking

Jul 27 22:43:57 localhost kernel: [ 1380.040292] radeon 0000:01:00.0: GPU lockup CP stall for more than 10000msec
Jul 27 22:43:57 localhost kernel: [ 1380.040300] radeon 0000:01:00.0: GPU lockup (waiting for 0x0000000000000431 last fence id 0x0000000000000430)
Jul 27 22:44:06 localhost kernel: [ 1388.818320] SysRq : Keyboard mode set to system default
Jul 27 22:44:09 localhost kernel: [ 1391.668196] SysRq : Keyboard mode set to system default
Jul 27 22:44:25 localhost kernel: [ 1407.683331] SysRq : Keyboard mode set to system default
Jul 27 22:44:28 localhost kernel: [ 1410.906535] SysRq : Keyboard mode set to system default
Jul 27 22:44:29 localhost kernel: [ 1411.308996] SysRq : Terminate All Tasks

after that rebooted via magic sysrq + reisub


"slowness":

*) my default 2 gnome-terminal windows (the left one attached to the left side of the monitor output), so that at the right side there's room for almost a third window (http://img845.imageshack.us/img845/8577/kfe8.jpg)

*) when left-clicking on the middle of the left gnome-terminal and then moving the window slowly (keeping left button pressed) the window slowly (with stuttering) is shown on the movement, where it normally is shown during the whole movement at every position - so it rather resembles a slide-show than smooth movement with approx. 0.5 cm distance between the corners of the window on a 22" screen

*) when attempting to fast moving the window (accelerated mouse movement) the cursor - like in a loop - is moving diagonally from the upper left corner of the window across the right lower corner of the window (a little bit further) - seems like the driver (is it even the driver ?) has issues working with the screen resolution



also: video acceleration of the pepperflash flash in chromium/google-chrome doesn't seem to work properly with the new DPM driver - every video seems to stutter (either "low" or "auto")

with the old driver even at "low" profile it appeared smooth


latest attempts on 3.10-patched kernel and 3.11-rc2* were compiled with gcc 4.7.2
Comment 13 jackdachef 2013-07-27 21:37:53 UTC 
I'll give radeon.aspm=0 a try - are there any more things I could test ?
Comment 14 Martin Peres 2019-11-19 08:36:18 UTC 
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/drm/amd/issues/353.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.