device reported invalid CHS sector 0 / NCQ disabled due to excessive errors
in [Kernel]
|
Prev: 2.6.34-rc3: simple du (on a big xfs tree) triggers oom killer [bisected: 57817c68229984818fea9e614d6f95249c3fb098]
Next: device reported invalid CHS sector 0 / NCQ disabled due to excessive errors
From: Robert Hancock on 25 Apr 2010 12:50 On 04/25/2010 09:16 AM, Richard Genoud wrote: > Hi ! > > I've bought a new motherboard (asus P7P55D) and I'm seeing annoying > errors about NCQ. > To reproduce the problem, I compile a kernel with -j4. > At some point, the compilation freezes, it seems that nothing is > written to the hdds for something like 1 or 2 minutes. > Then, all starts again and I can see in the logs a lot of "device > reported invalid CHS sector 0". > Sometimes, after that, NCQ is disabled. > > It happens on disks on my RAID 1 array (sda/sdb, ext4 formatted), but > not on my third disk (sdc, ext3 formatted) > > kernel is from ubuntu (with the commit > 8da854cb02156c90028233ae1e85ce46a1d3f82c (hpet erratum for intel P55)) > I tried with 2.6.34-rc5, and the problem is still there. > > Does this ring a bell ? (maybe it's not kernel related, and it's a > problem with my HDDs...) Usually this sort of thing is some kind of hardware problem - either a firmware bug in the drive, or else a SATA link communication problem like a bad cable, or the drives not getting clean power under load. > > Apr 21 19:20:03 desktop kernel: [ 119.920702] ata2: hard resetting link > Apr 21 19:20:03 desktop kernel: [ 120.449693] ata2: SATA link up 3.0 > Gbps (SStatus 123 SControl 300) > Apr 21 19:20:03 desktop kernel: [ 120.451774] ata2.00: configured for UDMA/133 > Apr 21 19:20:03 desktop kernel: [ 120.451782] ata2.00: device > reported invalid CHS sector 0 > Apr 21 19:20:03 desktop kernel: [ 120.451786] ata2.00: device > reported invalid CHS sector 0 > Apr 21 19:20:03 desktop kernel: [ 120.451790] ata2.00: device > reported invalid CHS sector 0 > Apr 21 19:20:03 desktop kernel: [ 120.451801] ata2: EH complete The log is missing the detail messages indicating what actually happened - can you post the output from dmesg instead? -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ |