ATA timeout

Matthew Dillon dillon at apollo.backplane.com
Fri Nov 5 15:17:30 PST 2004


:Matt,
:
:I was looking at the ATA code to get more familiar with it, and I decided to
:check out the logs.
:
:Revisions 1.8 and 1.9 of ata-disk.c both seem to be vaguely related to my
:problem.  Maybe.  The first change adds MPIPE and does not back off into PIO
:mode if resources are lacking.  I'm not sure about this one.  But the second
:one changes the ATA reset timing.  Could this have anything to do with my
:issue?
:
:jm

    It seems unlikely to be related to the MPIPE code, but it could certainly
    be related to the DELAY()'s.  Most of the other changes, e.g. like the
    ATA_S_READY flags checks, have kernel printf's associated with them so
    if any of those were the cause there ought to be additional error or
    warning lines printed to the console when the timeout occurs.

    I looked back at the archives trying to find a post with the actual
    kernel messages related to the timeout, or a better description of what
    is happening, but couldn't find one.  If you have that information it 
    would be helpful.  i.e. (1) The exact error and/or warning lines that
    are being printed on the console and (2) A description of the circumstances
    where the timeout occurs (e.g. during boot, after boot, during heavy disk
    I/O only, etc).

    We've suspected that there is some sort of sequencing issue in the ATA
    code for a while now but so far nobody has been able to put their finger
    on it.
    
					-Matt
					Matthew Dillon 
					<dillon at xxxxxxxxxxxxx>





More information about the Kernel mailing list