Kernel sometimes takes four or five resets to boot successfully
Hello, Running CRUX dual-booted with Windows 7 on two machines. On one everything is fine. On the other, when I turn the machine on for the first time in a day OR when I’ve just been in Windows and am switching back OR a few other random times, it takes me four or five pushes of the reset button to get it to boot. I get part of the startup output and then just after the lines: umount: /sys: device is busy. (In some cases useful info about processes that use the device is found by lsof(8) or fuser(1)) mount: sysfs already mounted or /sys busy … I get a massive spam of output, here is the part that doesn’t scroll off: [ 3.614147] [<c1273b6a>] ? ata_pio_sector+0x10a/0x140 [ 3.614147] [<c1273bff>] ? ata_pio_sectors+0x5f/0x90 [ 3.614147] [<c1273eba>] ? ata_sff_hsm_move+0x18a/0x790 [ 3.614147] [<c11babbd>] ? elv_queue_empty+0x1d/0x30 [ 3.614147] [<c11bd868>] ? __blk_run_queue+0x18/0x130 [ 3.614147] [<c11bda30>] ? blk_run_queue+0x20/0x40 [ 3.614147] [<c124ea92>] ? scsi_run_queue+0xv2/0x310 [ 3.614147] [<c1274696>] ? __ata_sff_port_intr+0xa6/0x100 [ 3.614147] [<c12734e0>] ? ata_bmdma_error_handler+0x0/0x110 [ 3.614147] [<c127471d>] ? ata_bmdma_port_intr_0x2d/0x110 [ 3.614147] [<c124e014>] ? scsi_decide_disposition+0x194/0x1a0 [ 3.614147] [<c12734e0>] ? ata_bmdma_error_handler+0x0/0x110 [ 3.614147] [<c12749b5>] ? ata_bmdma_interrupt+0x1b5/0x200 [ 3.614147] [<c107c8bd>] ? handle_IRQ_event+0x2d/0xc0 [ 3.614147] [<c107f6f1>] ? move_native_irq+0x11/0x50 [ 3.614147] [<c107ea03>] ? handle_edge_irq+0xa3/0x130 [ 3.614147] [<c1028ba5>] ? handle_irq+0x15/0x20 [ 3.614147] [<c1028887>] ? do_IRQ+0x47/0xc0 [ 3.614147] [<c103bd63>] ? smp_apic_timer_interrupt+0x53/0x90 [ 3.614147] [<c1026cb0>] ? common_interrupt+0x30/0x38 [ 3.614147] [<c102ca92>] ? mwait_idle+0x42/0x60 [ 3.614147] [<c1025225>] ? cpu_idle+0x85/0xb0 [ 3.614147] [<c15f06ed>] ? start_kernel+0x2c2/0x2c8 [ 3.614147] [<c15f019b>] ? unknown_bootoption+0x0/0x18f If anyone knows the way in which I have mis-compiled my kernel, I would love to know. Thanks. S
On 04.12.2011 18:04, Sean Whitton wrote:
Hello,
Running CRUX dual-booted with Windows 7 on two machines. On one everything is fine. On the other, when I turn the machine on for the first time in a day OR when I’ve just been in Windows and am switching back OR a few other random times, it takes me four or five pushes of the reset button to get it to boot.
I get part of the startup output and then just after the lines: umount: /sys: device is busy. (In some cases useful info about processes that use the device is found by lsof(8) or fuser(1)) mount: sysfs already mounted or /sys busy
I've been getting this message for quite a long time (don't remember since which kernel vesion), but all my Crux systems boot and run perfectly, so that's most probably not the reason for the kernel panic. Anyway, I just looked into /etc/rc again, and got the impression that the eroor is caused by the fact that /sys is already mounted when /etc/rc starts. So I made the following changes in /etc/rc: Comment out this line (ca. 13): /bin/mount -n -t sysfs none /sys and replace this (ca. 55): /bin/umount /sys /proc by /bin/umount /proc And indeed, the error message disappears, while /sys is still correctly mounted. Give it a try, but keep in mind that I'm not a kernel specialist! No guarantee! As to the real boot problem: Could it be a problem with your harddisk? Does the system boot reliably from CD-ROM? Bernd -- Bernd Eggink http://sudrala.de
Dear Bernd, On 4 Dec 2011 at 20:52Z, Bernd Eggink wrote:
I've been getting this message for quite a long time (don't remember since which kernel vesion), but all my Crux systems boot and run perfectly, so that's most probably not the reason for the kernel panic.
[…] Yes, it’s a bug in the /etc/rc script, and it’s never stopped my boot before—I was just writing it in my message to indicate what stage the script it at before the boot problem hits—thanks for providing a solution to something I’ve been meaning to fix for ages though!
As to the real boot problem: Could it be a problem with your harddisk? Does the system boot reliably from CD-ROM?
I’ve never seen the problem booting from a CD nor booting to Windows, so I don’t think it’s a HDD problem. Thanks for your reply. S
On Sun, 04 Dec 2011 17:04:06 +0000 Sean Whitton <sean@silentflame.com> wrote:
Hello,
Running CRUX dual-booted with Windows 7 on two machines. On one everything is fine. On the other, when I turn the machine on for the first time in a day OR when I’ve just been in Windows and am switching back OR a few other random times, it takes me four or five pushes of the reset button to get it to boot.
I get part of the startup output and then just after the lines: umount: /sys: device is busy. (In some cases useful info about processes that use the device is found by lsof(8) or fuser(1)) mount: sysfs already mounted or /sys busy
… I get a massive spam of output, here is the part that doesn’t scroll off:
[ 3.614147] [<c1273b6a>] ? ata_pio_sector+0x10a/0x140 [ 3.614147] [<c1273bff>] ? ata_pio_sectors+0x5f/0x90 [ 3.614147] [<c1273eba>] ? ata_sff_hsm_move+0x18a/0x790 [ 3.614147] [<c11babbd>] ? elv_queue_empty+0x1d/0x30 [ 3.614147] [<c11bd868>] ? __blk_run_queue+0x18/0x130 [ 3.614147] [<c11bda30>] ? blk_run_queue+0x20/0x40 [ 3.614147] [<c124ea92>] ? scsi_run_queue+0xv2/0x310 [ 3.614147] [<c1274696>] ? __ata_sff_port_intr+0xa6/0x100 [ 3.614147] [<c12734e0>] ? ata_bmdma_error_handler+0x0/0x110 [ 3.614147] [<c127471d>] ? ata_bmdma_port_intr_0x2d/0x110 [ 3.614147] [<c124e014>] ? scsi_decide_disposition+0x194/0x1a0 [ 3.614147] [<c12734e0>] ? ata_bmdma_error_handler+0x0/0x110 [ 3.614147] [<c12749b5>] ? ata_bmdma_interrupt+0x1b5/0x200 [ 3.614147] [<c107c8bd>] ? handle_IRQ_event+0x2d/0xc0 [ 3.614147] [<c107f6f1>] ? move_native_irq+0x11/0x50 [ 3.614147] [<c107ea03>] ? handle_edge_irq+0xa3/0x130 [ 3.614147] [<c1028ba5>] ? handle_irq+0x15/0x20 [ 3.614147] [<c1028887>] ? do_IRQ+0x47/0xc0 [ 3.614147] [<c103bd63>] ? smp_apic_timer_interrupt+0x53/0x90 [ 3.614147] [<c1026cb0>] ? common_interrupt+0x30/0x38 [ 3.614147] [<c102ca92>] ? mwait_idle+0x42/0x60 [ 3.614147] [<c1025225>] ? cpu_idle+0x85/0xb0 [ 3.614147] [<c15f06ed>] ? start_kernel+0x2c2/0x2c8 [ 3.614147] [<c15f019b>] ? unknown_bootoption+0x0/0x18f
If anyone knows the way in which I have mis-compiled my kernel, I would love to know. Thanks.
S
Had the kernel panic too when configuring my kernel. For me it was that the driver for my SATA-controller in "Device Drivers->Serial ATA and Parallel ATA drivers" didn't work correctly (or was the wrong one). With the driver in "ATA/ATAPI/MFM/RLL support (DEPRECATED)" it worked for me. So check if you really have the correct driver for your SATA-controller compiled in or try the deprecated driver. -- martu <martu_xd@gmx.de>
On 4 Dec 2011 at 23:04Z, martu wrote:
Had the kernel panic too when configuring my kernel. For me it was that the driver for my SATA-controller in "Device Drivers->Serial ATA and Parallel ATA drivers" didn't work correctly (or was the wrong one). With the driver in "ATA/ATAPI/MFM/RLL support (DEPRECATED)" it worked for me. So check if you really have the correct driver for your SATA-controller compiled in or try the deprecated driver.
I have an nvidia motherboard and disabling the nvidia driver under “Serial ATA and Parallel ATA drivers” stopped machine from booting with a kernel panic, good, I am pretty sure therefore that it’s the correct driver. I tried replacing it with the deprecated drivers, but wasn’t sure which one to enable—I enabled all the generic looking ones, but couldn’t see any kind of nvidia driver. What should I be looking for? Thanks. S
On Mon, Dec 05, 2011 at 01:23:08PM +0000, Sean Whitton wrote:
I have an nvidia motherboard and disabling the nvidia driver under “Serial ATA and Parallel ATA drivers” stopped machine from booting with a kernel panic, good, I am pretty sure therefore that it’s the correct driver.
I tried replacing it with the deprecated drivers, but wasn’t sure which one to enable—I enabled all the generic looking ones, but couldn’t see any kind of nvidia driver. What should I be looking for?
Thanks.
S
Could you perhaps post your output from "lspci -v" from the crux-install cd? Should help us figure out what driver you need. -- Fredrik Rinnestam
participants (4)
-
Bernd Eggink -
Fredrik Rinnestam -
martu -
Sean Whitton