Comment 25 for bug 286672

Revision history for this message
TJ (tj) wrote :

Forgot to mention my primary concern here!

Although the change in the arbitrary time-out will remove the 'link is slow to respond' message it won't of itself remove the kernel oops.

The ideal solution would be parallel processing for multiple devices maybe combined with some intelligence in the kernel's slow-path detection to account for situations that are mandated in specifications (in this case ATA-6 allows such delays) or imposed by hardware constraints.