[opensuse] bad thigs _DO_ happen!
I don't normally see the early part of the boot sequence, but today I caught something and decided to look deeper. In /var/log/boot.msg I find <6>[ 1.726463] [drm] Initialized drm 1.1.0 20060810 <6>[ 1.882467] [drm] radeon defaulting to kernel modesetting. <6>[ 1.882473] [drm] radeon kernel modesetting enabled. <3>[ 1.882599] fb: conflicting fb hw usage radeondrmfb vs VESA VGA - removing generic dr iver <6>[ 1.883552] radeon 0000:01:00.0: power state changed by ACPI to D0 <6>[ 1.883580] radeon 0000:01:00.0: power state changed by ACPI to D0 <6>[ 1.883598] radeon 0000:01:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16 <7>[ 1.883611] radeon 0000:01:00.0: setting latency timer to 64 <6>[ 1.887134] [drm] initializing kernel modesetting (RV380 0x1002:0x5460). <6>[ 1.887640] [drm] register mmio base: 0xC8100000 <6>[ 1.887645] [drm] register mmio size: 65536 <6>[ 1.889589] [drm] Generation 2 PCI interface, using max accessible memory <6>[ 1.889603] radeon 0000:01:00.0: VRAM: 128M 0x00000000D0000000 - 0x00000000D7FFFFFF ( 64M used) <6>[ 1.889613] radeon 0000:01:00.0: GTT: 512M 0x00000000B0000000 - 0x00000000CFFFFFFF <7>[ 1.889708] radeon 0000:01:00.0: irq 40 for MSI/MSI-X <6>[ 1.889721] radeon 0000:01:00.0: radeon: using MSI. <6>[ 1.889766] [drm] radeon: irq initialized. <6>[ 1.896415] [drm] Detected VRAM RAM=128M, BAR=128M <6>[ 1.896424] [drm] RAM width 128bits DDR <6>[ 1.902149] [drm] radeon: 64M of VRAM memory ready <6>[ 1.902159] [drm] radeon: 512M of GTT memory ready. <6>[ 1.902202] [drm] GART: num cpu pages 131072, num gpu pages 131072 <6>[ 1.904701] [drm] radeon: 1 quad pipes, 1 Z pipes initialized. <6>[ 1.908274] [drm] PCIE GART of 512M enabled (table at 0xD0040000). <6>[ 1.908324] radeon 0000:01:00.0: WB enabled <6>[ 1.908458] [drm] Loading R300 Microcode <6>[ 1.913683] [drm] radeon: ring at 0x00000000B0001000 <3>[ 2.129149] [drm:r100_ring_test] *ERROR* radeon: ring test failed (sracth(0x15E4)=0xC AFEDEAD) <3>[ 2.129168] [drm:r100_cp_init] *ERROR* radeon: cp isn't working (-22). <3>[ 2.129181] radeon 0000:01:00.0: failled initializing CP (-22). <3>[ 2.129192] radeon 0000:01:00.0: Disabling GPU acceleration <3>[ 2.345155] [drm:r100_cp_fini] *ERROR* Wait for CP idle timeout, shutting down CP. <4>[ 2.560986] Failed to wait GUI idle while programming pipes. Bad things might happen. <6>[ 2.561169] [drm] radeon: cp finalized Well, there is it "Failed to wait GUI idle while programming pipes. Bad things might happen" Yes, bad things do happen. A variety of them. Most common is the machine freezing, probably a GPU related panic. No panic message in the logs, just freeze. Sometimes the cursor freezes as well. Sometimes its preceded by what I can only call 'reticulation', if you know the effect on photographic emulsion -- a sort of overlay of green pixelation that renders things unreadable. Sometimes, not always. The freeze is often total, the laptop doesn't respond to pings. But not always. And sometimes I can ssh in and do a proper shutdown. But most of the time it takes the hardware OFF button to fix. Software or hardware? I'm using Linux BigBoy 2.6.37.6-0.5-pae #1 SMP 2011-04-25 21:48:33 +0200 i686 i686 i386 GNU/Linux lsmod tells me 'radeon' and /etc/X11/xorg.conf.d/50-device.conf says 'Driver "radeon"' When things work they work OK. -- "Good judgment comes from experience, and experience -- well, that comes from poor judgment." -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 20/07/11 21:43, Anton Aylward wrote:
I don't normally see the early part of the boot sequence, but today I caught something and decided to look deeper.
In /var/log/boot.msg I find
<6>[ 1.726463] [drm] Initialized drm 1.1.0 20060810 <6>[ 1.882467] [drm] radeon defaulting to kernel modesetting. <6>[ 1.882473] [drm] radeon kernel modesetting enabled. <3>[ 1.882599] fb: conflicting fb hw usage radeondrmfb vs VESA VGA - removing generic dr iver <6>[ 1.883552] radeon 0000:01:00.0: power state changed by ACPI to D0 <6>[ 1.883580] radeon 0000:01:00.0: power state changed by ACPI to D0 <6>[ 1.883598] radeon 0000:01:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16 <7>[ 1.883611] radeon 0000:01:00.0: setting latency timer to 64 <6>[ 1.887134] [drm] initializing kernel modesetting (RV380 0x1002:0x5460). <6>[ 1.887640] [drm] register mmio base: 0xC8100000 <6>[ 1.887645] [drm] register mmio size: 65536 <6>[ 1.889589] [drm] Generation 2 PCI interface, using max accessible memory <6>[ 1.889603] radeon 0000:01:00.0: VRAM: 128M 0x00000000D0000000 - 0x00000000D7FFFFFF ( 64M used) <6>[ 1.889613] radeon 0000:01:00.0: GTT: 512M 0x00000000B0000000 - 0x00000000CFFFFFFF <7>[ 1.889708] radeon 0000:01:00.0: irq 40 for MSI/MSI-X <6>[ 1.889721] radeon 0000:01:00.0: radeon: using MSI. <6>[ 1.889766] [drm] radeon: irq initialized. <6>[ 1.896415] [drm] Detected VRAM RAM=128M, BAR=128M <6>[ 1.896424] [drm] RAM width 128bits DDR <6>[ 1.902149] [drm] radeon: 64M of VRAM memory ready <6>[ 1.902159] [drm] radeon: 512M of GTT memory ready. <6>[ 1.902202] [drm] GART: num cpu pages 131072, num gpu pages 131072 <6>[ 1.904701] [drm] radeon: 1 quad pipes, 1 Z pipes initialized. <6>[ 1.908274] [drm] PCIE GART of 512M enabled (table at 0xD0040000). <6>[ 1.908324] radeon 0000:01:00.0: WB enabled <6>[ 1.908458] [drm] Loading R300 Microcode <6>[ 1.913683] [drm] radeon: ring at 0x00000000B0001000 <3>[ 2.129149] [drm:r100_ring_test] *ERROR* radeon: ring test failed (sracth(0x15E4)=0xC AFEDEAD) <3>[ 2.129168] [drm:r100_cp_init] *ERROR* radeon: cp isn't working (-22). <3>[ 2.129181] radeon 0000:01:00.0: failled initializing CP (-22). <3>[ 2.129192] radeon 0000:01:00.0: Disabling GPU acceleration <3>[ 2.345155] [drm:r100_cp_fini] *ERROR* Wait for CP idle timeout, shutting down CP. <4>[ 2.560986] Failed to wait GUI idle while programming pipes. Bad things might happen. <6>[ 2.561169] [drm] radeon: cp finalized
Well, there is it "Failed to wait GUI idle while programming pipes. Bad things might happen"
Yes, bad things do happen. A variety of them. Most common is the machine freezing, probably a GPU related panic. No panic message in the logs, just freeze. Sometimes the cursor freezes as well. Sometimes its preceded by what I can only call 'reticulation', if you know the effect on photographic emulsion -- a sort of overlay of green pixelation that renders things unreadable. Sometimes, not always.
The freeze is often total, the laptop doesn't respond to pings. But not always. And sometimes I can ssh in and do a proper shutdown.
But most of the time it takes the hardware OFF button to fix.
Software or hardware?
I'm using
Linux BigBoy 2.6.37.6-0.5-pae #1 SMP 2011-04-25 21:48:33 +0200 i686 i686 i386 GNU/Linux
lsmod tells me 'radeon' and /etc/X11/xorg.conf.d/50-device.conf says 'Driver "radeon"'
When things work they work OK.
When it freezes, do the kernel magic sysrq keys work? Alt+PrintScr+"r,s,e,i,u,b" should semi-safely reboot the machine -- best to avoid the hardware off if possible. Does it respond to a Ctrl-Alt-Backspace to restart X, or Ctrl-Alt-F* to switch VTs? If yes, maybe only X is breaking; if no, it's probably a kernel or hardware issue Regards, Tejas -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Tejas Guruswamy said the following on 07/20/2011 04:52 PM:
When it freezes, do the kernel magic sysrq keys work? Alt+PrintScr+"r,s,e,i,u,b" should semi-safely reboot the machine
Sometimes. Not always. (but then some don't do anything anyway!) (this is a laptop. It would take another hand to hit {r,s,e,i,u,b} after holding down the Alt+Fn at one end of the keyboard and the PrntScr at the other end of the keyboard.) Sometimes I'm in one application and it sort-of works. If I'm in a xterm I can issue commands, unless the whole thing is irredeemably frozen. But I can't change windows to the browser or mail reader, the menu won't pop-up and the clock in the panel bar isn't 'ticking'. That 'frozen' clock is a good sign that its locked up in some way. No 'sometimes' about that! Or perhaps only the cursor moves. Sometimes. Or perhaps the cursor doesn't move but everything is frozen. Sometimes. Sometimes Ctrl-Alt-backspace works to reset and I log in again. Then again, sometimes it doesn't. Everything is frozen and the the only way out is the hardware off. Sometimes. Sometimes it _seems_ to be triggered by a gross action such as trying to close a window or a tab, something that causes a massive screen redraw. But then again, other massive screen redraws such as flipping virtual desktops or 'present desktops' or the pop-up for composing a new mail message or sending it, or the darkening of the background whole-screen window when a foreground pops up DO NOT seems to cause the freeze. No 'sometimes' about those operations. I've become paranoid and when the laptop boots I SSH in from another machine to start with :-) If that is dead, if the laptop isn't returning pings then its hardware off time. The point I'm trying to make is that there are a variety of degrees of lock-up/freeze. I have no specific ideas about what triggers them. I suspect the GPU. Perhaps there is a specific bug, some 'optimisation' about a specific type of redraw ... that's why I posted the snippet from the log file. And no, after the event, after a reboot or restart, I can't find anything in the logs that gives me additional clues. Is there something I can do to make xorg produce ore logs, even at the cost of performance?
-- best to avoid the hardware off if possible.
+1 :-) I do when I can. Sometimes I can't. Sometimes.
Does it respond to a Ctrl-Alt-Backspace to restart X, or Ctrl-Alt-F* to switch VTs? If yes, maybe only X is breaking; if no, it's probably a kernel or hardware issue
Sometimes. Oh, ctl-alt-F* hasn't seemed to worked for me at all on 11.4 -- Misfortune shows those who are not really friends. Aristotle (384 BC - 322 BC), Eudemian Ethics -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 07/21/2011 08:43 AM, Anton Aylward pecked at the keyboard and wrote:
Tejas Guruswamy said the following on 07/20/2011 04:52 PM:
When it freezes, do the kernel magic sysrq keys work? Alt+PrintScr+"r,s,e,i,u,b" should semi-safely reboot the machine
Sometimes. Not always. (but then some don't do anything anyway!) (this is a laptop. It would take another hand to hit {r,s,e,i,u,b} after holding down the Alt+Fn at one end of the keyboard and the PrntScr at the other end of the keyboard.)
You only hit _one_ of the {r,s,e,i,u,b} keys, not all of them together. -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Ken Schneider - openSUSE said the following on 07/21/2011 05:04 PM:
On 07/21/2011 08:43 AM, Anton Aylward pecked at the keyboard and wrote:
Tejas Guruswamy said the following on 07/20/2011 04:52 PM:
When it freezes, do the kernel magic sysrq keys work? Alt+PrintScr+"r,s,e,i,u,b" should semi-safely reboot the machine
Sometimes. Not always. (but then some don't do anything anyway!) (this is a laptop. It would take another hand to hit {r,s,e,i,u,b} after holding down the Alt+Fn at one end of the keyboard and the PrntScr at the other end of the keyboard.)
You only hit _one_ of the {r,s,e,i,u,b} keys, not all of them together.
While that is only 9 keys I'd need to press I would have to be contortionist or an octopus. -- Parents complain that "kids don't do anything for themselves any more." Then they write letters to the board demanding that the schools do something about it. -- Arnold Lapiner -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 07/21/2011 08:19 PM, Anton Aylward pecked at the keyboard and wrote:
Ken Schneider - openSUSE said the following on 07/21/2011 05:04 PM:
On 07/21/2011 08:43 AM, Anton Aylward pecked at the keyboard and wrote:
Tejas Guruswamy said the following on 07/20/2011 04:52 PM:
When it freezes, do the kernel magic sysrq keys work? Alt+PrintScr+"r,s,e,i,u,b" should semi-safely reboot the machine
Sometimes. Not always. (but then some don't do anything anyway!) (this is a laptop. It would take another hand to hit {r,s,e,i,u,b} after holding down the Alt+Fn at one end of the keyboard and the PrntScr at the other end of the keyboard.)
You only hit _one_ of the {r,s,e,i,u,b} keys, not all of them together.
While that is only 9 keys I'd need to press I would have to be contortionist or an octopus.
How does hitting <alt><PrtScr>r (or s, or e, or i, or u, or b) require you to be a contortionist? It is only three keys! -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
While that is only 9 keys I'd need to press I would have to be contortionist or an octopus.
How does hitting <alt><PrtScr>r (or s, or e, or i, or u, or b) require you to be a contortionist? It is only three keys!
as a laptop keyboard, prtscr isn't a stand alone key, so it'd be 4: <alt><fn><prtscr><r|s|e|i|u|b> it seems feasible (at least on my machine in front of me) to hit fn, alt and s all with left hand, leaving right free to get to prtscr. it's not quite comfortable, but it should be something that can happen on rare occasion. -- Even the Magic 8 ball has an opinion on email clients: Outlook not so good. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Can we get past the keyboard issue and get back to the messages from the GPU and the problem with the machine freezing _PLEASE_ -- Relying on the government to protect your privacy is like asking a peeping tom to install your window blinds. -- John Perry Barlow -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (4)
-
Anton Aylward -
Ken Schneider - openSUSE -
Tejas Guruswamy -
zGreenfelder