On Wednesday 08 of September 2004 16:47, Rafael J. Wysocki wrote: > On Wednesday 08 of September 2004 14:55, Zwane Mwaikambo wrote: > > On Wed, 8 Sep 2004, Rafael J. Wysocki wrote: [- snip -] > > Try nmi_watchdog=0 > > It works, thanks. > > Now, I'm able to suspend the box, seemingly, but there are some problems with > resuming. There is a crash related to USB (gets "fixed" if I unload usb_ohci > etc. before suspending) and a preepmtion problem. I'll send some traces as > soon as I can get output from the serial console. In the meantime, I'll try > to compile out preemption. I've got the traces. First, I start the kernel with the following command line: root=/dev/hdc6 vga=792 resume=/dev/hdc3 nmi_watchdog=0 console=ttyS0,57600 Then, if I log in as root and do "echo 4 > /proc/acpi/sleep", it seems to suspend successfully. However, when I press the power button and start the kernel with the above command line, I get either this: Relocating pagedir ....::::| Reading image data (11699 pages): .......................................................................................... Stopping tasks: ==| Freeing memory: | PM: Restoring saved image. <7>PM: Image restored successfully. Losing some ticks... checking if CPU frequency changed. Warning: CPU frequency out of sync: cpufreq and timing core thinks of 800000, is 1800000 kHz. PCI: Setting latency timer of device 0000:00:02.0 to 64 ohci_hcd 0000:00:02.0: HC died; cleaning up usb 1-2: USB disconnect, address 2 Badness in hcd_endpoint_disable at drivers/usb/core/hcd.c:1310 Call Trace:{:usbcore:hcd_endpoint_disable+107} {:usbcore:usb_disable_endpoint+41}<7>PCI: Setting latency timer of device 0000:00:02.1 to 64 {:usbcore:usb_disable_device+26} {:usbcore:usb_disconnect+188} <3>ohci_hcd 0000:00:02.1: HC died; cleaning up {:usbcore:usb_disconnect+2} {:usbcore:hcd_panic+0} {:usbcore:hcd_panic+74} {worker_thread+733} {default_wake_function+0} {__wake_up_common+67} for the kernel with preemption compiled in (that's all I get), or this: Relocating pagedir ....::::| Reading image data (11543 pages): .......................................................................................... Stopping tasks: ==| Freeing memory: | PM: Restoring saved image. <7>PM: Image restored successfully. Losing some ticks... checking if CPU frequency changed. warning: many lost ticks. Your time source seems to be instable or some driver is hogging interupts rip swsusp_free+0xfe/0x1a0 Warning: CPU frequency out of sync: cpufreq and timing core thinks of 800000, is 1800000 kHz. PCI: Setting latency timer of device 0000:00:02.0 to 64 ohci_hcd 0000:00:02.0: HC died; cleaning up usb 1-2: USB disconnect, address 2 Badness in hcd_endpoint_disable at drivers/usb/core/hcd.c:1310 Call Trace:{:usbcore:hcd_endpoint_disable+107} {:usbcore:usb_disable_endpoint+41} {:usbcore:usb_disable_device+26} {:usbcore:usb_disconnect+188} {:usbcore:usb_disconnect+2} {:usbcore:hcd_panic+0} {:usbcore:hcd_panic+74} {worker_thread+674} {default_wake_function+0} {__wake_up_common+67} {default_wake_function+0} {worker_thread+0} {kthread+205} {child_rip+8} {kthread+0} {child_rip+0} Badness in hcd_endpoint_disable at drivers/usb/core/hcd.c:1310 for the kernel without preemption. Hence, there's a problem with USB drivers, so I've repeated this experiment with the USB drivers unloaded (except for usbcore which is always in use, apparently), but the results haven't been much better. Specifically, if I unload USB drivers (ie "rmmod ohci_hcd ehci_hcd usbserial usbhid") before issuing "echo 4 > /proc/acpi/sleep", I get either this: Relocating pagedir not necessary Reading image data (11729 pages): ...................................................................................<1>Una {free_block+128} PML4 0 Oops: 0000 [1] PREEMPT CPU 0 Modules linked in: Pid: 3, comm: events/0 Not tainted 2.6.9-rc1-mm4 RIP: 0010:[] {free_block+128} RSP: 0000:000001001fe39db8 EFLAGS: 00010016 RAX: 000001000105c110 RBX: 000001001ff02880 RCX: 0000010001a4e088 RDX: 0000000000000000 RSI: 0000010001a4e000 RDI: 000001001ff02880 RBP: 0000000000000000 R08: 000001001ff21010 R09: 0000000000000002 R10: 0000000000000000 R11: 0000000000000000 R12: 000001001ff02890 R13: 0000000000000008 R14: 000001001ff09010 R15: 000001001ff028b0 FS: 0000000000000000(0000) GS:ffffffff805e7500(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000000 CR3: 0000000000101000 CR4: 00000000000006e0 Process events/0 (pid: 3, threadinfo 000001001fe38000, task 000001001fe10030) Stack: ffffffff8049bf60 000001001ff09010 000001001ff09000 0000000000000008 000001001ff02960 000001001fc7e9b0 ffffffff805b1520 ffffffff80179146 000001001ff02880 000001001fc7e9b0 Call Trace:{drain_array_locked+118} {cache_reap+268} {cache_reap+0} {worker_thread+733} {default_wake_function+0} {__wake_up_common+67} {default_wake_function+0} {worker_thread+0} {kthread+205} {child_rip+8} {kthread+0} {child_rip+0} Code: 48 39 32 74 0c 0f 0b 1d a9 44 80 ff ff ff ff a4 00 48 8b 06 RIP {free_block+128} RSP <000001001fe39db8> CR2: 0000000000000000 <6>note: events/0[3] exited with preempt_count 1 ................................... 11729 done. Reading resume file was successful PM: Preparing system for restore. for the preemptible kernel, or this: Relocating pagedir not necessary Reading image data (11928 pages): ...................................................................................------- Kernel BUG at slab:164 invalid operand: 0000 [1] CPU 0 Modules linked in: Pid: 3, comm: events/0 Not tainted 2.6.9-rc1-mm4 RIP: 0010:[] {free_block+133} RSP: 0000:000001001fe39db8 EFLAGS: 00010083 RAX: 000001000105c1b8 RBX: 000001001ff02080 RCX: 0000010001a51078 RDX: 00000100010e3048 RSI: 0000010001a51040 RDI: 000001001ff02080 RBP: 0000000000000000 R08: 000001001ff24010 R09: 0000000000000002 R10: 0000000000000000 R11: 0000000000000000 R12: 000001001ff02090 R13: 000000000000000b R14: 000001001ff21010 R15: 000001001ff020b0 FS: 0000000000000000(0000) GS:ffffffff805b0500(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000004000 CR3: 0000000000101000 CR4: 00000000000006e0 Process events/0 (pid: 3, threadinfo 000001001fe38000, task 000001001fe10030) Stack: ffffffff80464260 000001001ff21010 000001001ff21000 000000000000000b 000001001ff02160 000001001fc7e9b0 ffffffff8057a520 ffffffff801721a6 000001001fc7e9b0 000001001ff02080 Call Trace:{drain_array_locked+118} {cache_reap+254} {cache_reap+0} {worker_thread+674} {default_wake_function+0} {__wake_up_common+67} {default_wake_function+0} {worker_thread+0} {kthread+205} {child_rip+8} {kthread+0} {child_rip+0} Code: 0f 0b 75 2c 41 80 ff ff ff ff a4 00 48 8b 06 48 39 70 08 74 RIP {free_block+133} RSP <000001001fe39db8> ..................................... 11928 done. Reading resume file was successful PM: Preparing system for restore. for the non-preemptible kernel. So, it does not work in either case, but for different reasons. I'm attaching the messages I get when the box is suspending (isn't the "swsusp: Pagedir: 0 Pages" suspicious?) and the .config. Please let me know if you need any more info. Also, if you have any patches that may help fix this, please send them to me for testing. Greets, RJW -- - Would you tell me, please, which way I ought to go from here? - That depends a good deal on where you want to get to. -- Lewis Carroll "Alice's Adventures in Wonderland"