Kernel 2.6.15-7-686 won't boot with root on SATA disk

[Ondřej Surý]

With following hardware:
0000:00:1f.1 IDE interface: Intel Corporation 82801FB/FBM/FR/FW/FRW (ICH6
Family) IDE Controller (rev 03)
0000:00:1f.2 IDE interface: Intel Corporation 82801FBM (ICH6M) SATA Controller
(rev 03)
kernel won't boot, but instead fall to shell after several minutes with message
that it cannot find root. After some time and trying to 'modprobe libata' SATA
controller is detected and after quiting shell kernel is loaded.

My root is /dev/sda1 and thus detected as SCSI according to /scripts/init-premount/

[Ondřej Surý (ondrej)]

Oh, and I forget to specify kernel version, it's 2.6.15-7.9.

[Chuck Short (zulcss)]

Please try the newer version in dapper.

Thanks
chuck

[Chuck Short (zulcss)]

No response, closing.

[Ondřej Surý (ondrej)]

With 2.6.15-25-686 it takes incredibly long time in "Waiting for root filesystem".
I am trying to find a way ho

[Ondřej Surý (ondrej)]

With 2.6.15-25-686 it takes incredibly long time in "Waiting for root filesystem".
I am trying to find a way how to attach bootchart picture here.... ok, it's at http://www.rfc1925.org/dapper-20060624-1.png.bz2

[Ondřej Surý (ondrej)]

Uncompressed picture can be found at http://www.rfc1925.org/dapper-20060624-1.png (size difference is neglible on normal internet connection).

[Ben Collins (ben-collins)]

Can you provide /etc/fstab, and /proc/mounts?

I suspect you are just having a problem with sda vs. hda.

[Ondřej Surý (ondrej)]

root@ookami:~# cat /proc/mounts
rootfs / rootfs rw 0 0
none /sys sysfs rw 0 0
none /proc proc rw,nodiratime 0 0
udev /dev tmpfs rw 0 0
/dev/sda1 / ext3 rw,noatime,data=ordered 0 0
/dev/sda1 /dev/.static/dev ext3 rw,noatime,data=ordered 0 0
tmpfs /var/run tmpfs rw 0 0
tmpfs /var/lock tmpfs rw 0 0
usbfs /proc/bus/usb usbfs rw 0 0
tmpfs /lib/modules/2.6.15-25-686/volatile tmpfs rw 0 0
devpts /dev/pts devpts rw 0 0
tmpfs /dev/shm tmpfs rw 0 0
tmpfs /var/run tmpfs rw 0 0
tmpfs /var/lock tmpfs rw 0 0
/dev/sda5 /home xfs rw,noatime,nosuid,nodev 0 0
/dev/sda2 /windows ntfs ro,noatime,nodiratime,uid=1000,gid=1000,fmask=0333,dmask=0222,nls=utf8,errors=continue,mft_zone_multiplier=1 0 0
root@ookami:~# cat /etc/fstab
# /etc/fstab: static file system information.
#
# <file system> <mount point> <type> <options> <dump> <pass>
proc /proc proc defaults 0 0
/dev/sda1 / ext3 noatime,errors=remount-ro 0 1
/dev/sda5 /home xfs noatime,nodev,nosuid 0 2
/dev/sda2 /windows ntfs defaults,ro,nls=utf8,uid=1000,gid=1000,dmask=0222,fmask=0333 0 0
/dev/hda /media/cdrom0 udf,iso9660 user,noauto 0 0
//pagan/downloads /mnt/downloads smbfs noauto,username=ondrej,password=Iemo2Uwi,workgroup=RFC1925,uid=1000,gid=1000,iocharset=utf-8,ro,ip=87.236.198.133,port=10445 0 0

[Ondřej Surý (ondrej)]

BTW adding ahci to /etc/mkinitramfs/modules and regenerating /boot/initrd-whatever-kernel-version-is does help. See any of:
<<EOF xargs -i wget www.rfc1925.org/{}
dapper-20060625-1.png
dapper-20060626-1.png
dapper-20060627-1.png
dapper-20060627-2.png
EOF

[Ben Collins (ben-collins)]

Likely fixed in edgy, please test.

[Ondřej Surý (ondrej)]

Nope, it's not. See the very long delay between 4th and 5th line:

[17179579.088000] SCSI subsystem initialized
[17179579.088000] libata version 1.20 loaded.
[17179579.088000] ahci 0000:00:1f.2: version 1.2
[17179579.088000] ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 18 (level, low) -> IRQ 201
[17179652.820000] PCI: Setting latency timer of device 0000:00:1f.2 to 64
[17179652.820000] ahci 0000:00:1f.2: AHCI 0001.0000 32 slots 4 ports 1.5 Gbps 0x5 impl IDE mode
[17179652.820000] ahci 0000:00:1f.2: flags: 64bit ncq pm led pmp slum part

[Ondřej Surý (ondrej)]

Fixed in 2.6.17-6.18, there is still very short delay, but I guess it's normal.

[Ondřej Surý (ondrej)]

Now it behaves just strange on lastest edgy kernel, sometimes it does boot fast, but more often there is long delay before SATA disks are brought up, see todays boot dmesg snippet:

[17179572.852000] SCSI subsystem initialized
[17179572.852000] libata version 1.20 loaded.
[17179572.852000] ahci 0000:00:1f.2: version 1.2
[17179572.852000] ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 18 (level, low) -> IRQ 201
[17179654.724000] PCI: Setting latency timer of device 0000:00:1f.2 to 64
[17179654.724000] ahci 0000:00:1f.2: AHCI 0001.0000 32 slots 4 ports 1.5 Gbps 0x5 impl IDE mode
[17179654.724000] ahci 0000:00:1f.2: flags: 64bit ncq pm led pmp slum part
[17179654.724000] ata1: SATA max UDMA/133 cmd 0xF8864500 ctl 0x0 bmdma 0x0 irq 201
[17179654.724000] ata2: SATA max UDMA/133 cmd 0xF8864580 ctl 0x0 bmdma 0x0 irq 201
[17179654.724000] ata3: SATA max UDMA/133 cmd 0xF8864600 ctl 0x0 bmdma 0x0 irq 201
[17179654.724000] ata4: SATA max UDMA/133 cmd 0xF8864680 ctl 0x0 bmdma 0x0 irq 201
[17179655.980000] ata1: SATA link up 1.5 Gbps (SStatus 113)

[Ben Collins (ben-collins)]

But it does boot on edgy, correct? I believe the latest changes I put in edgy (today, upload later) will fix the delay.

[Ondřej Surý (ondrej)]

Yes, it's fixed in edgy. Thanks

[Martin Pitt (pitti)]

Ben, are the changes in comment 14 sufficient for solving this? Are they appropriate for an SRU? What is the diff?

[Martin Pitt (pitti)]

We will not fix this in dapper. Fixing this properly is too intrusive for a stable release according to Ben Collins, and there are no duplicates or other subscribers.