Re: Failed to detect hard disk on TP X31

From: Bruce Evans <bde_at_zeta.org.au>
Date: Wed, 21 Apr 2004 22:25:38 +1000 (EST)
On Wed, 21 Apr 2004, Alex Dupre wrote:

> Dennis Kristensen wrote:
>
> > Just a "me too". With latest source from a few hours ago my ATA devices
> > are not probed again but my floppy suddenly is. Reverting the recent
> > ata-pci.c/pci.c makes them re-appear, but the floppy to disappear.
>
> Try to revert only pci.c. This makes my system work againg, and floppy too.

Here it did the same, except with pci.c not reverted attempting to use
DMA failures on ad0 and a panic in the turnstile code soon after.  This
was primarily because pci.c broke the ata DMA BAR.  Without pci.c reverted,
I get the following boot messages:

%%%
atapci0: <nVidia nForce2 UDMA133 controller> port 0xf000-0xf00f at device 9.0 on pci0
atapci0: Reserved 0x10 bytes for rid 0x20 type 4 at 0xf000
atapci0: Lazy allocation of 0x8 bytes rid 0x10 type 4 at 0x1f0
atapci0: Lazy allocation of 0x1 bytes rid 0x14 type 4 at 0x3f6
%%%

With pci.c unreverted, rid 0x20 is at 0 instead of at 0xf000.  This does
not work.

Bogus allocations for memory-mapped devices fail safer than bogus
allocations for i/o-mapped devices because memory accesses to address 0
cause null pointer panics while port accesses to address 0 just access
a wrong port.

Bruce
Received on Wed Apr 21 2004 - 03:25:51 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:51 UTC