Figuring Out SATA SErr codes

Today I had to cast my eye over a SATA error message:

ata1: exception Emask 0x10 SAct 0x0 SErr 0xd0000 action 0xe frozen

ata1: irq_stat 0x00400000, PHY RDY changed

ata1: SError: { PHYRdyChg CommWake 10B8B }

ata1: hard resetting link

So how does one interpret the cryptic SErr magic? Pages 269-270 of the SATA ATA spec serialata10a.zip explains the SError register fields in fine detail.

Well, "SErr 0xd0000" relates to the SError bits 16, 18 and 19 which are the SERR_PHYRDY_CHG, SERR_COMM_WAKE and SERR_10B_8B_ERR bits as defined in include/linux/ata.h. The kernel decodes these and dumps out the error state in the "ata1: SError: { PHYRdyChg CommWake 10B8B }" line above - so that's helpful.

Anyhow, the spec describes these error codes in the DIAG field (page 270) - they are the top 16 bits of the SError register. Armed with the spec one can then decode these error bits. It's not rocket science, one just needs to know where to look this information up.