Re: 2.2.19 + ide 2.2.19 03252001 patch problem

[Willy Tarreau]

Quoting "Robert A. Morris" <[e-mail]>:
[snip]
> Apr  5 18:15:14 ryoko kernel: hdb: task_no_data_intr: status=0x51 {
> DriveReady SeekComplete Error }
> Apr  5 18:15:14 ryoko kernel: hdb: task_no_data_intr: error=0x04 {
> DriveStatusError }
> Apr  5 18:15:14 ryoko kernel: hdb: Write Cache FAILED Flushing!
[snip]
> This did NOT happen with 2.2.18 and the corresponding 
> ide.2.2.18.1209.patch.  It does NOT seem to happen on
> /dev/hda or /dev/hdc, which is lucky, since /dev/hdb
> is unused.  I'm using lilo.conf to specify idebus=33.
[snip] 
> The controller is a VIA 82C686A (Asus K7V mainboard).
> hda: WDC WD307AA, 29333MB w/2048kB Cache, CHS=3739/255/63, UDMA(66)
> hdb: WDC AC28400R, 8063MB w/512kB Cache, CHS=1027/255/63, (U)DMA
> hdc: WDC WD307AA-00BAA0, 29333MB w/2048kB Cache, CHS=59598/16/63,
> UDMA(66)

same problem observed here on same motherboard. The hard disk is a WDC AC23200L
configured as hda. I have tested several ide/kernel combinations and all I can
say is that 2.2.18 and 2.2.19 behave the same, but it worked till
ide.2.2.18.1221 included, and the bug appeared since ide.2.2.18.02122001.
I tried with and without vojtech's via patches (3.2, 4.2 and 4.3), but this
didn't change anything (to be honest, some combinations were obviously not made
to live together, and I had so many problems fitting all patches in one kernel
that it sometimes even didn't boot).

I can also say that this problem didn't show up on other chipsets (ali and
intel) with the same kernel+ide patch.

finally, I made my kernel with ide.2.2.18.1221 and all seems to be OK (one week
now). The diffs between the 2 versions were too important and I have not
investigated further into this, but I'm ready to make some tests if needed.

Regards,
Willy

PS: BTW Andre, could you please name your patches ide-2.2.19-YYYYMMDD so that a
    directory listing show the chronological order ?