Bug 15945
Summary: | Filling disk with data leads to [sda] Unhandled error code. [sda] Result hostbyte=DID_OK driverbyte=DRIVER_TIMEOUT | ||
---|---|---|---|
Product: | IO/Storage | Reporter: | Alexander Konovalenko (alexkon) |
Component: | Serial ATA | Assignee: | Jeff Garzik (jgarzik) |
Status: | RESOLVED INSUFFICIENT_DATA | ||
Severity: | normal | CC: | alan, ovilewade9, tj |
Priority: | P1 | ||
Hardware: | All | ||
OS: | Linux | ||
Kernel Version: | 2.6.32-21-generic #32-Ubuntu SMP | Subsystem: | |
Regression: | No | Bisected commit-id: | |
Attachments: |
Output of smartctl -a /dev/sda
Output of hdparm -I /dev/sda Output of dmesg log-eh-race.patch |
Description
Alexander Konovalenko
2010-05-09 11:49:02 UTC
Created attachment 26297 [details]
Output of hdparm -I /dev/sda
Can you please post the output of dmesg after such problem occurred? Please turn on printk timestamp if possible. Thanks. Created attachment 26298 [details]
Output of dmesg
Here it is. Are the timestamps in square brackets at the start of each line OK? Or should I turn on some other timestamps?
Created attachment 26409 [details]
log-eh-race.patch
Sorry about the delay. Yeap, those timestamps are good. Something definitely is fishy. I can't see how you can get READ failure without libata complaining first. The read can be from dm-crypt if you're writing blocks which are smaller than encryption unit size.
Can you please apply the attached patch, boot with ignore_loglevel kernel parameter, trigger the problem and post the dmesg output?
Thank you.
|