Re: IDE Hard disk geometry problem in 2.4.19 / 2.4.20pre7

[Jeff Garzik]

James Finnie wrote:
> Hi,
> 
> I have a large number of IDE hard disks here where I work.  Since moving to
> 2.4.19 from 2.4.17, several of these disks have stopped working, resulting
> in a kernel panic after the drive has declared itself to have 0 cylinders!
> All the disks that have broken are the following:
> 
> Seagate 80GB U6 ST380020ACE with Firmware version 4.65
> 
> We have lots of these same drives, with FW v 3.34, that all work fine.  I do
> not have a single drive with the 4.65 firmware working.  
> 
> My problem is that these drives used to work fine, with 2.4.17.  They are
> not obsolete hardware, I think they are all less than 6 months old.
> 
> I have seen this on CS5530 with the standard kernel PCI IDE, and on SIS630
> with the SIS Kernel IDE driver.  
> 
> Setting ide=nodma makes no difference.
> 
> Here is an excerpt from the kernel console booting:
> 
> 
> ....
> hda: ST380020ACE, ATA DISK drive
> ....
> hda: task_no_data_intr: status=0x51 { DriveReady SeekComplete Error }
> hda: task_no_data_intr: error=0x04  { DriveStatusError }
> hda: setmax_ext LBA 1, native 0
> hda: 0 sectors (0 MB) w/2048KiB Cache, CHS 0/255/63, (U)DMA
> ....
> hda2: bad access: block=2, count=2
> end_request: I/O error, dev 03:02 (hda), sector 2
> EXT3-fs: unable to read superblock
> hda2: bad access: block=2, count=2
> end_request: I/O error, dev 03:02 (hda), sector 2
> EXT3-fs: unable to read superblock
> Kernel panic: VFS: Unable to mount root fs on 03:02
> 
> 
> Is this a result of all the new IDE stuff that went in in 2.4.19???

"All the new IDE stuff" hasn't really hit 2.4.x yet...

Do you have CONFIG_IDEDISK_STROKE set?  If yes, do things start working 
if you disable it?

Can you post the other IDE options you have set in your kernel config?

	Jeff