From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.9) Gecko/20020408 Description of problem: Using the 7.3 kernel, I get IDE errors after enabling DMA (see below). Using vanilla 2.4.19-preX kernels, older vanilla kernels, or older RH packaged kernels this does not occur. /etc/sysconfig/harddisk* settings are the same. Intel ICH 82801AA controller Maxtor 98196H8 drive USE_DMA=1 MULTIPLE_IO=16 EIDE_32BIT=3 EXTRA_PARAMS="-u1" Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: See above combination of kernel + controller + drive + hdparm settings Actual Results: May 7 17:29:48 thor kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error } May 7 17:29:48 thor kernel: hda: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=156249063, sector=138579768 May 7 17:29:48 thor kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error } May 7 17:29:48 thor kernel: hda: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=156249063, sector=138579768 May 7 17:29:48 thor kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error } May 7 17:29:48 thor kernel: hda: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=156249063, sector=138579768 May 7 17:29:48 thor kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error } May 7 17:29:48 thor kernel: hda: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=156249063, sector=138579768 May 7 17:29:48 thor kernel: hda: DMA disabled May 7 17:29:48 thor kernel: ide0: reset: success Expected Results: DMA to not give errors. Additional info: (I really dislike the flakiness of IDE... sigh. if all the world was SCSI...)
Also - just in case this is useful: Award BIOS PCCHIPS motherboard that I don't know the number of in a small form-factor "Book PC" 80GB Maxtor drive has the 33GB cylinder limit clip on it so I can boot off of it. I use "setmax -d0 /dev/hda in rc.modules to enable the full size before it gets to the hdparm or fsck+mounting non-root filesystems part of the boot. (This part works ok even with the shipped kernel) Let me know if you need additional info like /proc/ide contents or whatnot.
Just released 2.4.18-4 has the same behavior
this bug was never updated... anyway, stock RH8.0 or RH9 kernel didn't do this on that particular setup.