1GHz pbook 15", linux 2.6.14-rc2 oops on resume
Hi all,
when a dvd featuring some iso content is in the dvd-drive and the machine is put to sleep mode, it will give the following oops on resume. It is working without problems if no media is in the drive. Voluntary preemption is ON.
Find below the dmesg output when a dvd is in the drive.
eth1: Airport entering sleep mode
eth0: suspending, WakeOnLan disabled
radeonfb (0000:00:10.0): suspending to state: 2...
uninorth-agp: disabling AGP on device 0000:00:10.0
uninorth-agp: disabling AGP on bridge 0000:00:0b.0
radeonfb (0000:00:10.0): switching to D2 state...
radeonfb (0000:00:10.0): resuming from state: 2...
radeonfb (0000:00:10.0): switching to D0 state...
agpgart: Putting AGP V2 device at 0000:00:0b.0 into 4x mode
agpgart: Putting AGP V2 device at 0000:00:10.0 into 4x mode
eth0: resuming
A
eth1 (WE) : Driver using old /proc/net/wireless support, please fix driver !
eth1: get_wireless_stats() called while device not present
PHY ID: 2060e1, addr: 0
eth1 (WE) : Driver using old /proc/net/wireless support, please fix driver !
eth1: get_wireless_stats() called while device not present
eth1: Airport waking up
hda: Enabling Ultra DMA 4
eth1: New link status: Connected (0001)
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Pause is enabled (rxfifo: 10240 off: 7168 on: 5632)
BUG: soft lockup detected on CPU#0! NIP: C0006FCC LR: C02BBC5C SP: EEF6DC00 REGS: eef6db50 TRAP: 0901 Not tainted
MSR: 0200b032 EE: 1 PR: 0 FP: 1 ME: 1 IR/DR: 11
TASK = ef800740[3274] 'pbbuttonsd' THREAD: eef6c000
Last syscall: 54
GPR00: 00000080 EEF6DC00 EF800740 00079C96 000088B8 00000000 00000000 C05A8A50
GPR08: C05A8538 EEF6DCC8 00100000 00140040 22004282
NIP [c0006fcc] __delay+0xc/0x14
LR [c02bbc5c] ide_wait_not_busy+0x4c/0xc0
Call trace: [c02b9fa0] ide_do_request+0x5b0/0x990 [c02ba440] ide_do_drive_cmd+0xc0/0x190 [c02b6c00] generic_ide_resume+0x80/0xa0 [c0293b00] resume_device+0x70/0x150 [c0293db0] dpm_resume+0x100/0x1a0 [c0293e8c] device_resume+0x3c/0xa0 [c05438cc] pmac_wakeup_devices+0xbc/0xe0 [c0544adc] pmu_ioctl+0x58c/0x9b0 [c008e254] do_ioctl+0x84/0x90 [c008e2ec] vfs_ioctl+0x8c/0x450 [c008e744] sys_ioctl+0x94/0xb0 [c0004820] ret_from_syscall+0x0/0x44
hdc: Enabling MultiWord DMA 2
eth1 (WE) : Driver using old /proc/net/wireless support, please fix driver !
VFS: busy inodes on changed media.
adb: starting probe task...
adb devices: [2]: 2 c4 [3]: 3 1 [7]: 7 1f
ADB keyboard at 2, handler 1
ADB mouse at 3, handler set to 4 (trackpad)
adb: finished probe task...
eth1 (WE) : Driver using old /proc/net/wireless support, please fix driver !
agpgart: Putting AGP V2 device at 0000:00:0b.0 into 4x mode
agpgart: Putting AGP V2 device at 0000:00:10.0 into 4x mode [drm] Loading R200 Microcode
eth1 (WE) : Driver using old /proc/net/wireless support, please fix driver !
hdc: irq timeout: status=0xc0 { Busy }
ide: failed opcode was: unknown
hdc: DMA disabled
eth1 (WE) : Driver using old /proc/net/wireless support, please fix driver ! eth1 (WE) : Driver using old /proc/net/wireless support, please fix driver !
hdc: ATAPI reset complete
VFS: busy inodes on changed media.
Soeren
Signature · 2 lines
Sometimes, there's a moment as you're waking, when you become aware of the real world around you, but you're still dreaming.
ide_wait_not_busy oops still with 2.6.14-rc3 (Re: 1GHz pbook 15", linux 2.6.14-rc2 oops on resume)
Quoted 7 lines · Soeren Sonnenburg
Hi all,
when a dvd featuring some iso content is in the dvd-drive and the machine is put to sleep mode, it will give the following oops on resume. It is working without problems if no media is in the drive. Voluntary preemption is ON.
Find below the dmesg output when a dvd is in the drive.
now it is:
GPR00: 00000080 EDEF5C00 EF894780 00079C96 000088B8 00000000 00000000 C05A8A50
GPR08: C05A8538 EDEF5CC8 00100000 00140040 22004222
NIP [c0006fcc] __delay+0xc/0x14
LR [c02bc32c] ide_wait_not_busy+0x4c/0xc0
Call trace: [c02ba670] ide_do_request+0x5b0/0x990 [c02bab10] ide_do_drive_cmd+0xc0/0x190 [c02b72d0] generic_ide_resume+0x80/0xa0 [c0294260] resume_device+0x70/0x150 [c0294510] dpm_resume+0x100/0x1a0 [c02945ec] device_resume+0x3c/0xa0 [c05438cc] pmac_wakeup_devices+0xbc/0xe0 [c0544adc] pmu_ioctl+0x58c/0x9b0 [c008e344] do_ioctl+0x84/0x90 [c008e3dc] vfs_ioctl+0x8c/0x450 [c008e834] sys_ioctl+0x94/0xb0 [c0004820] ret_from_syscall+0x0/0x44
Soeren
Re: ide_wait_not_busy oops still with 2.6.14-rc3 (Re: 1GHz pbook 15", linux 2.6.14-rc2 oops on resume)
Quoted 16 lines · Soeren Sonnenburg
Quoted 7 lines · Soeren Sonnenburg
Hi all,
when a dvd featuring some iso content is in the dvd-drive and the machine is put to sleep mode, it will give the following oops on resume. It is working without problems if no media is in the drive. Voluntary preemption is ON.
Find below the dmesg output when a dvd is in the drive.now it is:
GPR00: 00000080 EDEF5C00 EF894780 00079C96 000088B8 00000000 00000000 C05A8A50
GPR08: C05A8538 EDEF5CC8 00100000 00140040 22004222
NIP [c0006fcc]__delay+0xc/0x14
LR [c02bc32c]ide_wait_not_busy+0x4c/0xc0
Quoted 13 lines
Call trace: [c02ba670] ide_do_request+0x5b0/0x990 [c02bab10] ide_do_drive_cmd+0xc0/0x190 [c02b72d0] generic_ide_resume+0x80/0xa0 [c0294260] resume_device+0x70/0x150 [c0294510] dpm_resume+0x100/0x1a0 [c02945ec] device_resume+0x3c/0xa0 [c05438cc] pmac_wakeup_devices+0xbc/0xe0 [c0544adc] pmu_ioctl+0x58c/0x9b0 [c008e344] do_ioctl+0x84/0x90 [c008e3dc] vfs_ioctl+0x8c/0x450 [c008e834] sys_ioctl+0x94/0xb0 [c0004820] ret_from_syscall+0x0/0x44
You haven't put the complete oops, what is the trap number ? Does it help adding a delay in the wakeup code in drivers/ide/ppc/pmac.c ? Also, is the problem present without preempt ?
Ben.
Re: ide_wait_not_busy oops still with 2.6.14-rc3 (Re: 1GHz pbook 15", linux 2.6.14-rc2 oops on resume)
Quoted 5 lines · Benjamin Herrenschmidt
Quoted 4 lines · Soeren Sonnenburg
Quoted 2 lines · Soeren Sonnenburg
Hi all,
Quoted 7 lines
Quoted 7 lines
Quoted 5 lines
when a dvd featuring some iso content is in the dvd-drive and the machine is put to sleep mode, it will give the following oops on resume. It is working without problems if no media is in the drive. Voluntary preemption is ON.
Find below the dmesg output when a dvd is in the drive.now it is:
[incomplet oops]
ok, here is the complete one:
BUG: soft lockup detected on CPU#0! NIP: C0006FCC LR: C02BC32C SP: EDEF5C00 REGS: edef5b50 TRAP: 0901 Not tainted
MSR: 0200b032 EE: 1 PR: 0 FP: 1 ME: 1 IR/DR: 11
TASK = ef894780[3425] 'pbbuttonsd' THREAD: edef4000
Last syscall: 54
GPR00: 00000080 EDEF5C00 EF894780 00079C96 000088B8 00000000 00000000 C05A8A50
GPR08: C05A8538 EDEF5CC8 00100000 00140040 22004222
NIP [c0006fcc] __delay+0xc/0x14
LR [c02bc32c] ide_wait_not_busy+0x4c/0xc0
Call trace: [c02ba670] ide_do_request+0x5b0/0x990 [c02bab10] ide_do_drive_cmd+0xc0/0x190 [c02b72d0] generic_ide_resume+0x80/0xa0 [c0294260] resume_device+0x70/0x150 [c0294510] dpm_resume+0x100/0x1a0 [c02945ec] device_resume+0x3c/0xa0 [c05438cc] pmac_wakeup_devices+0xbc/0xe0 [c0544adc] pmu_ioctl+0x58c/0x9b0 [c008e344] do_ioctl+0x84/0x90 [c008e3dc] vfs_ioctl+0x8c/0x450 [c008e834] sys_ioctl+0x94/0xb0 [c0004820] ret_from_syscall+0x0/0x44
hdc: Enabling MultiWord DMA 2
adb: starting probe task...
adb devices: [2]: 2 c4 [3]: 3 1 [7]: 7 1f
ADB keyboard at 2, handler 1
ADB mouse at 3, handler set to 4 (trackpad)
adb: finished probe task...
agpgart: Putting AGP V2 device at 0000:00:0b.0 into 4x mode
agpgart: Putting AGP V2 device at 0000:00:10.0 into 4x mode [drm] Loading R200 Microcode
hdc: irq timeout: status=0xc0 { Busy }
ide: failed opcode was: unknown
hdc: DMA disabled
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Pause is enabled (rxfifo: 10240 off: 7168 on: 5632)
hdc: ATAPI reset complete
Quoted 2 lines
You haven't put the complete oops, what is the trap number ? Does it help adding a delay in the wakeup code in
drivers/ide/ppc/pmac.c? Also,
are you talking about increasing this delay
#define IDE_WAKEUP_DELAY (1*HZ) or sth. to pmac_ide_do_resume() ?
Quoted 1 line
is the problem present without preempt ?
Currently it is:
Preemption Model (Voluntary Kernel Preemption (Desktop))
I will try.
Soeren
Signature · 2 lines
Sometimes, there's a moment as you're waking, when you become aware of the real world around you, but you're still dreaming.
Re: ide_wait_not_busy oops still with 2.6.14-rc3 (Re: 1GHz pbook 15", linux 2.6.14-rc2 oops on resume)
Quoted 4 lines
ok, here is the complete one:
BUG: soft lockup detected on CPU#0!
Gack, the soft lockup thing. Can you disable that ? If you do so, does it crashes instead of oopsing or just "pauses" for a little while on wakeup ? The problem is that ide_do_request does a synchronous wait for the drive to get out of busy state which can take a while with some optical drives on wakeup. It might be possible to allow scheduling there, I have to look at it. In the meantime, disable the lockup detector (CONFIG_DETECT_SOFTLOCKUP) and tell me if that's enough.
Ben.
Quoted 50 lines
NIP: C0006FCC LR: C02BC32C SP: EDEF5C00 REGS:
edef5b50TRAP: 0901 Not tainted
MSR:0200b032EE: 1 PR: 0 FP: 1 ME: 1 IR/DR: 11
TASK = ef894780[3425] 'pbbuttonsd' THREAD:edef4000
Last syscall: 54
GPR00: 00000080 EDEF5C00 EF894780 00079C96 000088B8 00000000 00000000 C05A8A50
GPR08: C05A8538 EDEF5CC8 00100000 00140040 22004222
NIP [c0006fcc]__delay+0xc/0x14
LR [c02bc32c]ide_wait_not_busy+0x4c/0xc0Call trace: [c02ba670] ide_do_request+0x5b0/0x990 [c02bab10] ide_do_drive_cmd+0xc0/0x190 [c02b72d0] generic_ide_resume+0x80/0xa0 [c0294260] resume_device+0x70/0x150 [c0294510] dpm_resume+0x100/0x1a0 [c02945ec] device_resume+0x3c/0xa0 [c05438cc] pmac_wakeup_devices+0xbc/0xe0 [c0544adc] pmu_ioctl+0x58c/0x9b0 [c008e344] do_ioctl+0x84/0x90 [c008e3dc] vfs_ioctl+0x8c/0x450 [c008e834] sys_ioctl+0x94/0xb0 [c0004820] ret_from_syscall+0x0/0x44hdc: Enabling MultiWord DMA 2
adb: starting probe task...
adb devices: [2]: 2 c4 [3]: 3 1 [7]: 7 1f
ADB keyboard at 2, handler 1
ADB mouse at 3, handler set to 4 (trackpad)
adb: finished probe task...
agpgart: Putting AGP V2 device at 0000:00:0b.0 into 4x mode
agpgart: Putting AGP V2 device at 0000:00:10.0 into 4x mode [drm] Loading R200 Microcode
hdc: irq timeout: status=0xc0 { Busy }
ide: failed opcode was: unknown
hdc: DMA disabled
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Pause is enabled (rxfifo: 10240 off: 7168 on: 5632)
hdc: ATAPI reset completeQuoted 2 lines
You haven't put the complete oops, what is the trap number ? Does it help adding a delay in the wakeup code in
drivers/ide/ppc/pmac.c? Also,are you talking about increasing this delay
#defineIDE_WAKEUP_DELAY(1*HZ) or sth. topmac_ide_do_resume()?Quoted 1 line
is the problem present without preempt ?
Currently it is:
Preemption Model (Voluntary Kernel Preemption (Desktop))I will try.
Soeren
Re: ide_wait_not_busy oops still with 2.6.14-rc3 (Re: 1GHz pbook 15", linux 2.6.14-rc2 oops on resume)
Quoted 11 lines · Benjamin Herrenschmidt
Quoted 3 lines
ok, here is the complete one:
BUG: soft lockup detected on CPU#0!Gack, the soft lockup thing. Can you disable that ? If you do so, does it crashes instead of oopsing or just "pauses" for a little while on wakeup ? The problem is that
ide_do_requestdoes a synchronous wait for the drive to get out of busy state which can take a while with some optical drives on wakeup. It might be possible to allow scheduling there, I have to look at it. In the meantime, disable the lockup detector (CONFIG_DETECT_SOFTLOCKUP) and tell me if that's enough.
Hmmhh, I already compiled 2.6.14-rc4 but did not disable soft-lockup-ing, should I still do it - the oops looks better as it is not followed by a ATAPI reset anymore:
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Pause is enabled (rxfifo: 10240 off: 7168 on: 5632)
BUG: soft lockup detected on CPU#0! NIP: C0006FCC LR: C02BAD1C SP: EE60DBF0 REGS: ee60db40 TRAP: 0901 Not tainted
MSR: 0200b032 EE: 1 PR: 0 FP: 1 ME: 1 IR/DR: 11
TASK = c128b870[3312] 'pbbuttonsd' THREAD: ee60c000
Last syscall: 54
GPR00: 00000080 EE60DBF0 C128B870 00079C96 000088B8 00000000 00000000 C05A6A50
GPR08: C05A6538 EE60DCC8 00100000 00140040 22004282
NIP [c0006fcc] __delay+0xc/0x14
LR [c02bad1c] ide_wait_not_busy+0x4c/0xc0
Call trace: [c02b918c] ide_do_request+0x74c/0xa00 [c02b9500] ide_do_drive_cmd+0xc0/0x190 [c02b5c80] generic_ide_resume+0x80/0xa0 [c0292cec] resume_device+0x6c/0x140 [c0292f8c] dpm_resume+0xfc/0x1a0 [c0293068] device_resume+0x38/0xa0 [c05418cc] pmac_wakeup_devices+0xbc/0xe0 [c0542adc] pmu_ioctl+0x58c/0x9b0 [c008def4] do_ioctl+0x84/0x90 [c008df8c] vfs_ioctl+0x8c/0x450 [c008e3e4] sys_ioctl+0x94/0xb0 [c0004820] ret_from_syscall+0x0/0x44
hdc: Enabling MultiWord DMA 2
Quoted 1 line
Ben.
Soeren
Quoted 51 lines
Quoted 50 lines
NIP: C0006FCC LR: C02BC32C SP: EDEF5C00 REGS:
edef5b50TRAP: 0901 Not tainted
MSR:0200b032EE: 1 PR: 0 FP: 1 ME: 1 IR/DR: 11
TASK = ef894780[3425] 'pbbuttonsd' THREAD:edef4000
Last syscall: 54
GPR00: 00000080 EDEF5C00 EF894780 00079C96 000088B8 00000000 00000000 C05A8A50
GPR08: C05A8538 EDEF5CC8 00100000 00140040 22004222
NIP [c0006fcc]__delay+0xc/0x14
LR [c02bc32c]ide_wait_not_busy+0x4c/0xc0Call trace: [c02ba670] ide_do_request+0x5b0/0x990 [c02bab10] ide_do_drive_cmd+0xc0/0x190 [c02b72d0] generic_ide_resume+0x80/0xa0 [c0294260] resume_device+0x70/0x150 [c0294510] dpm_resume+0x100/0x1a0 [c02945ec] device_resume+0x3c/0xa0 [c05438cc] pmac_wakeup_devices+0xbc/0xe0 [c0544adc] pmu_ioctl+0x58c/0x9b0 [c008e344] do_ioctl+0x84/0x90 [c008e3dc] vfs_ioctl+0x8c/0x450 [c008e834] sys_ioctl+0x94/0xb0 [c0004820] ret_from_syscall+0x0/0x44hdc: Enabling MultiWord DMA 2
adb: starting probe task...
adb devices: [2]: 2 c4 [3]: 3 1 [7]: 7 1f
ADB keyboard at 2, handler 1
ADB mouse at 3, handler set to 4 (trackpad)
adb: finished probe task...
agpgart: Putting AGP V2 device at 0000:00:0b.0 into 4x mode
agpgart: Putting AGP V2 device at 0000:00:10.0 into 4x mode [drm] Loading R200 Microcode
hdc: irq timeout: status=0xc0 { Busy }
ide: failed opcode was: unknown
hdc: DMA disabled
eth0: Link is up at 1000 Mbps, full-duplex.
eth0: Pause is enabled (rxfifo: 10240 off: 7168 on: 5632)
hdc: ATAPI reset completeQuoted 2 lines
You haven't put the complete oops, what is the trap number ? Does it help adding a delay in the wakeup code in
drivers/ide/ppc/pmac.c? Also,are you talking about increasing this delay
#defineIDE_WAKEUP_DELAY(1*HZ) or sth. topmac_ide_do_resume()?Quoted 1 line
is the problem present without preempt ?
Currently it is:
Preemption Model (Voluntary Kernel Preemption (Desktop))I will try.
Soeren
Signature · 2 lines
Sometimes, there's a moment as you're waking, when you become aware of the real world around you, but you're still dreaming.
Re: ide_wait_not_busy oops still with 2.6.14-rc3 (Re: 1GHz pbook 15", linux 2.6.14-rc2 oops on resume)
Quoted 13 lines · Soeren Sonnenburg
Quoted 11 lines · Benjamin Herrenschmidt
Quoted 3 lines
ok, here is the complete one:
BUG: soft lockup detected on CPU#0!Gack, the soft lockup thing. Can you disable that ? If you do so, does it crashes instead of oopsing or just "pauses" for a little while on wakeup ? The problem is that
ide_do_requestdoes a synchronous wait for the drive to get out of busy state which can take a while with some optical drives on wakeup. It might be possible to allow scheduling there, I have to look at it. In the meantime, disable the lockup detector (CONFIG_DETECT_SOFTLOCKUP) and tell me if that's enough.
Quoted 3 lines
Hmmhh, I already compiled 2.6.14-rc4 but did not disable soft-lockup-ing, should I still do it - the oops looks better as it is not followed by a ATAPI reset anymore:
It's still pretty annoying. I'll see what I can do but it won't be for 2.6.14 timeframe, so in the meantime, just ignore it or remove soft lockup detection.
Ben.
Re: ide_wait_not_busy oops still with 2.6.14-rc3 (Re: 1GHz pbook 15", linux 2.6.14-rc2 oops on resume)
Quoted 3 lines · Benjamin Herrenschmidt
Quoted 2 lines · Soeren Sonnenburg
Quoted 1 line · Benjamin Herrenschmidt
Quoted 1 line
ok, here is the complete one:
[...]
Quoted 7 lines
Quoted 3 lines
Hmmhh, I already compiled 2.6.14-rc4 but did not disable soft-lockup-ing, should I still do it - the oops looks better as it is not followed by a ATAPI reset anymore:
It's still pretty annoying. I'll see what I can do but it won't be for 2.6.14 timeframe, so in the meantime, just ignore it or remove soft lockup detection.
OK, it is a minor issue anyway as one needs a cd/dvd in the drive (on resume) to trigger that.
Thanks
Soeren.
Signature · 2 lines
Sometimes, there's a moment as you're waking, when you become aware of the real world around you, but you're still dreaming.
Re: ide_wait_not_busy oops still with 2.6.14-rc3 (Re: 1GHz pbook 15", linux 2.6.14-rc2 oops on resume)
In reply to Soeren SonnenburgQuoted 12 lines · Soeren Sonnenburg
Quoted 11 lines · Benjamin Herrenschmidt
Quoted 10 lines · Soeren Sonnenburg
Quoted 7 lines · Soeren Sonnenburg
Hi all,
when a dvd featuring some iso content is in the dvd-drive and the machine is put to sleep mode, it will give the following oops on resume. It is working without problems if no media is in the drive. Voluntary preemption is ON.
Find below the dmesg output when a dvd is in the drive.
Quoted 6 lines
Quoted 1 line
Quoted 1 line
now it is:
[incomplet oops]
ok, here is the complete one:
BUG: soft lockup detected on CPU#0!
This isn't an Oops. What the soft lockup detector is telling you is that one CPU detected that another CPU was running something in the kernel for a LONG time without rescheduling which is considered a bug. The fix is to make ide_wait_not_busy() preemptible.
Lee