Bugzilla – Bug 548165
Harddrives not detected by kernal/installer on both DVD and Live CD
Last modified: 2009-12-02 10:08:19 UTC
Created attachment 323088 [details] Log file User-Agent: Mozilla/5.0 (compatible; Konqueror/4.3; Linux) KHTML/4.3.1 (like Gecko) SUSE I am not able install openSUSE 11.2 on my system with either the live CD or DVD so this is both an install and boot problem. My motherboard is an ASRock ALIveXFire-eSATA2. I think the hard disk controller is recognized but not the hard disks. There no problem in openSUSE 11.1. An attachment of openSUSE 11.2 boot log is included. Reproducible: Always Steps to Reproduce: 1. start boot process 2. 3. Actual Results: an examples from the boot log. 6.168025] ata4.00: qc timeout (cmd 0xec) [ 6.168037] ata4.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 6.169143] ata3.00: qc timeout (cmd 0xec) [ 6.169153] ata3.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 6.506022] ata1.00: qc timeout (cmd 0xec) [ 6.506032] ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 6.626035] ata4: SATA link up 3.0 Gbps (SStatus 123 SControl 300) [ 6.627152] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300) [ 6.832022] ata2.00: qc timeout (cmd 0xec) [ 6.832031] ata2.00: failed to IDENTIFY (I/O error, err_mask=0x4) [ 7.321033] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 300) [ 8.004033] ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 300) Expected Results: Able to use the computer but no access to hard drive.
http://en.opensuse.org/Bugs/Definitions#Bug_Priorities
Does "pci=nomsi" or "irqpoll" kernel parameter make any difference? Can you try to attach /var/log/boot.msg after failed boot? From the installation dvd, you can switch to command console by pressing ctrl-alt-f2 and then mount usb stick and copy the log over. Thanks.
Neither pci=nomsi nor irqpoll help for me. (tried with openSUSE-11.2-DVD-x86_64-iso) My hwinfo is attached to http://bugzilla.novell.com/show_bug.cgi?id=548021
So, 32bit works fine but 64bit doesn't and on 64bit "pci=nomsi", "irqpoll" or "noacpi" doesn't help, right? Can you please try "mem=3G"?
"mem=3G" does not help for me. Without noacpi I get errormessages as posted in https://bugzilla.novell.com/show_bug.cgi?id=548021, with "noacpi" I got messages as posted by Michael Wright. openSuSE 11.1 works (64 bit, I have not tried 32 bit), openSuSE 11.2 32 bit works, 64 bit does not.
Frank, can you please attach kernel log messages with and without noacpi here? It gets difficult to track things without direct attachments and also a lot of information is lost with indirect "same error messages". Thanks.
Created attachment 330100 [details] boot message with noacpi
I was only able to add log messages with noacpi. Otherwise the system hangs (no reaction to keyboard at all) for more than 5 minutes. Is there any other way to save the log messages? I can take pictures of the screen (but do see only the very last lines). If nothing helps, I can try a serial cable, which I do not have by hand now. Thus I would like to try another way.
Serial console would be the surest way but if you have another machine there, you can also try netconsole. Documentation/networking/netconsole.txt in the kernel source directory contains more details. You'll need to include the netconsole module and module for your network interface in initrd. Hmmmm.... It looks like "mkinitrd -f netconsole" should do the trick. Thanks.
Created attachment 330368 [details] boot messages without noacpi I managed to log the boot message by serial console. As the messages were logged with an old palmtop (used ptelnet) and put together later, I hope there are no errors. I was't able to log the messages at higher baud rate, it seems to be related to the slow hardware. Even 9600 baud failed... That's all I can provide at the moment.
cc'ing Thomas. The root cause of the problem seems to be oops coming from acpi code. noacpi avoids that but breaks IRQ delivery for ATA controller which isn't too surprising these days. Thomas, does the oops ring any bell? Thanks.
Can you attach acpidump in the other bug, please. Thanks. *** This bug has been marked as a duplicate of bug 559680 ***