Everything works properly even though a lot of error messages are generated. I can access all partitions on the RAID drives using device mapper. This problem shows up on disk drives that were formatted by Windows to use the Promise pseudo-hardware RAID contoller PDC20276. I have found references to similar problems with other psuedo-hardware controllers on the internet. It appears that some tables on the the first RAID drive (usually hde) contain a range of sector addresses that span both hard drives in the RAID array. When the module pdc202xx_new is loaded it attempts to read some of the addresses that are not on the hde disk. This results in a lot of meaningless error messages for sectors that are not on the disk. This problem is in Linux 2.6.11 and is still present in 2.6.16. It does not show up in 2.4.x because a different driver is used. I see the error messages when I boot if I compile the pdc202xx_new into the kernel. All the error messages come scrolling by as the driver attempts to access a lot of invalid addresses. I made the pdc202xx_new a module instead of compiling it in and rebuilt the kernel. This time there were no error messages, but device mapper didn't work. Realizing my error, I added "pdc202xx_new" to the /etc/modules.autoload.d/kernel-2.6 file. Now device manager works and I can access the files on the RAID disks. I noticed a slight pause when the modules were loaded, but no error messages came out on the screen. Is this a fix maybe? Nope - the messages are still there when I looked with dmesg. This is at least a work arround since it used to take a long time to boot with all those messages scrolling by. They also appear later in the boot process. I'm guessing that this is a bug for the maintainer of the pdc202xx_new driver. Example of a set of error messages. These are repeated over and over: hde: task_in_intr: status=0x51 { DriveReady SeekComplete Error } hde: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=237569222, sector=237569222 ide: failed opcode was: unknown hde: task_in_intr: status=0x51 { DriveReady SeekComplete Error } hde: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=237569222, sector=237569222 ide: failed opcode was: unknown hde: task_in_intr: status=0x51 { DriveReady SeekComplete Error } hde: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=237569222, sector=237569222 ide: failed opcode was: unknown hde: task_in_intr: status=0x51 { DriveReady SeekComplete Error } hde: task_in_intr: error=0x10 { SectorIdNotFound }, LBAsect=237569222, sector=237569222 ide: failed opcode was: unknown PDC202XX: Primary channel reset. ide2: reset: success
Can't see any existing reports of this. Please open a bug at http://bugzilla.kernel.org and post the new URL here.
