Still no Radeon 9k support ? :-)
Still no -march=pentium(x) support ? :-)
GCC 3.3 inline ? hint -finline-limit and/or compiler.h :-)
Margit
> Mainly due to a IDE DMA problem which would happen on boxes with lots of
> RAM, here is -rc5.
Hmm, should I say "Thanks, Santa!", or is this a completely different problem
from the one I reported on Saturday 24th May:
"[BUG] IDE DMA timeout, then crash on reenable (2.4.20-SMP)"?
If this fix *is* related to my bug report then is there any way that I can test
this, now that we know what might be causing it? My usual way of reproducing it
is "boot, then wait for 28+ days of uptime", which doesn't sound like a part of
a plan to get 2.4.21 out the door...
Cheers,
Chris
> Mainly due to a IDE DMA problem which would happen on boxes with lots of
> RAM, here is -rc5.
Neither -rc5, nor a patch by Ivan, nor both, can get UDMA to turn on
automatically at boot time on my machine. This is a 466MHz DS10 Alphaserver,
2GB RAM, ALI 1543 IDE controller that is capable of UDMA2. hdparm can turn
on multiword DMA but not UDMA.
Hope this helps,
jasonp