Re: IDE problem source ?

[Alan B. <al...@ms...>]

hi,

> Great. I just committed a change which moves the call to the proper
> location. Could you try this one too (after reverting your changes)?

I'm not sure if I'm alone here.... but your change doesnt work
(i know...long delay in supplying this info but my HD died
on the day i checked out this CVS addition :-|)


Your change was the addition of

ide_ack_intr(hwif);

to line ~358 in ide-probe.c



But, the change by Krystian Baclawski ...which was done after some
information by you... DOES allow the kernel to get beyond IDE probing
and boot on the A1200.  Line ~560 is the change...an addition of

hwif->hw.ack_intr(hwif);

as shown in the diff below


        __restore_flags(flags); /* local CPU only */
        if (hwif->hw.ack_intr && hwif->irq)
        {
                hwif->hw.ack_intr(hwif);
                enable_irq(hwif->irq);
        }
        for (unit = 0; unit < MAX_DRIVES; ++unit) {
                ide_drive_t *drive = &hwif->drives[unit];
                if (drive->present) {
                        ide_tuneproc_t *tuneproc = HWIF(drive)->tuneproc;
                        if (tuneproc != NULL && drive->autotune == 1)
                                tuneproc(drive, 255);
                }
        }


unfortunately its not all plain sailing and happiness. when long, prolonged
use of the IDE is undertaken  (eg  'updatedb' ) the kernel has a nice
Oops!   I havent got the full net access etc back up yet...so will ksymoops
this later to give more details... but suffice to say, i dont get this
Ooops with the working 2.4.13-20011030 kernel  (otherwise I wouldnt
have been able to compile the 2.4.18 )

I did note that at the end of the ooops! report it has the following lines.


Kernel panic: Aiee, killing interupt handler!
In interupt Handler - not syncing

>>>reboot in 180 seconds (etc etc)

these 2 lines dont appear to be normal.....

alan