|
Bugzilla – Full Text Bug Listing |
| Summary: | No drives found when installing | ||
|---|---|---|---|
| Product: | [openSUSE] SUSE LINUX 10.0 | Reporter: | Forgotten User jq9zgB7cRO <forgotten_jq9zgB7cRO> |
| Component: | Kernel | Assignee: | Greg Kroah-Hartman <gregkh> |
| Status: | RESOLVED FIXED | QA Contact: | E-mail List <qa-bugs> |
| Severity: | Critical | ||
| Priority: | P5 - None | CC: | forgotten_jq9zgB7cRO, snwint |
| Version: | Final | ||
| Target Milestone: | --- | ||
| Hardware: | i386 | ||
| OS: | SuSE Linux 10.0 | ||
| Whiteboard: | |||
| Found By: | Customer | Services Priority: | |
| Business Priority: | Blocker: | --- | |
| Marketing QA Status: | --- | IT Deployment: | --- |
| Bug Depends on: | 117541 | ||
| Bug Blocks: | |||
|
Description
Forgotten User jq9zgB7cRO
2005-10-18 06:55:07 UTC
This looks suspiciously like an ACPI problem. I added dependency on my bug 117541 since there are several similarities. sis chipset (but AMD processor), insmod=ide-generic helps a little, lots of other boot options tried with no success, errors from hda Note that 10.0 beta 3 did work!!! Don't mix up too much things here.
> the same issue with Suse 9.3 Pro. I used the "insmod=ide-generic" in
> order for 9.3 to install.
-> I doubt that this is the problem in 10.0, we are probably seeing something else here, possibly slipped in at Beta3.
So this is the problem you are facing from dmesg:
-----------------------------------------------
linux kernel: hda: task_no_data_intr: status=0x51 { DriveReady SeekComplete Error }
linux kernel: hda: task_no_data_intr: error=0x04 { DriveStatusError }
linux kernel: hda: failed opcode was: 0xea
linux kernel: hda: wcache flush failed!
-----------------------------------------------
Looks like a disk driver problem to me or more likely a broken hard disk?
Hmm could it still be an incorrect/not assigned interrupt? -> then it could be ACPI. Please try pci=noacpi and/or noapic.
Please also try the smart mon tools, maybe they reveal that you better should buy a new disk.
Hmm, I just had another look at #117541, unlikely that there are two broken disks which work on other systems (Knoppix)? Jens, do you have an idea what these hda messages mean? BTW: This is not a blocker, it is probably something machine specific. We are seeing this all over the forums. I think it is at least critical Could you try pci=noacpi and/or noapic, please. Maybe also provide /proc/interrupts of all these boot params. If possible please also attach a complete dmesg output if booted when the error happens. If the error still apears with all, it's unlikely to be an ACPI interrupt problem and Jens has to have a deeper look into it. Could you also try the current kernel: ftp://ftp.suse.com/pub/people/mantel/kotd/i386/HEAD/kernel-default.i586.rpm (Best install with rpm -i --force and add both kernels in /boot/grub/menu.lst, so that you can boot the old one again, in case the new one fails). In my room at work, we have also such a machine (Antje's old workstation); at the moment, I don't have the time for testing. Feel free to pick it up. IIRC, Steffen know a workaround to make the disk visible. > Steffen knows a workaround to make the disk visible.
-> CC'ing Steffen.
If you are asking me for more info then that is all I have, the ide=generic, acpi and apic have all been tried, there are no messages as the people cannot even get the disks recognised to install the release. I cannot reproduce here as I am on 64bit with nforce3 all SATA and that works fine, the only machine I had any probs with was an old Pc-Chips mobo with a sis chipset but the irqpoll and pci=usepirqmask worked for me. Some of the guys who have reported the problem are seasoned guys and know a bit or two but still the disks are not recognised from the get-go. Setting needinfo to Steffen. If he knows a workaround further investigations would be a wast of time... Ahh ok, the workaround is: irqpoll and pci=usepirqmask? So this is not an ACPI, but a SIS chipset specific bug? -> Assigning to Steffen, sorry don't know who I should assign it to, but as this is not an ACPI bug, it's nothing for me... Maybe you want to assign it back to Hubert who should know who can fix that? Note that the workaround suggested in comment 11 was rejected in the initial bug description! Please, if you want me to try anything ask in my bug report #117541 (see other comments there) I was not cc:d to this until today... There were several reports about sis5513 not working. I think it's a driver problem. It should work in the kernels from both HEAD and 10.0 now (patches.drivers/sis5513-support-sis-965l) in both. Grab a KOTD rpm with that in the log. HEAD didn't get it until today, but the 10.0 kernels have it since a week or so. Reopen for 10.1 if the bug still exists. |