This patch does apear to prevent the panic as described in PR kern/57369.
However the system still falls back to PIO4 when it should be capable of
UDMA33 (at least)
On October 30, 2003 09:32 pm, Juan Manuel Sanchez wrote:> The ata raid code in STABLE allows 15 seconds for reading the disk
> headers. If during probe the ata channels need several resets to find a
> working configuration that time might be exceeded. When the request
> finally succeds the buffer was already freed by ar_rw(), and kernel
> panics at any access to corrupt pointers.
>
> Try this simple patch:
>
> -----------------------------------------------------------------------
> Index: ata-raid.c
> ==================================================================> RCS
file: /usr/home/ncvs/src/sys/dev/ata/ata-raid.c,v
> retrieving revision 1.3.2.19
> diff -w -u -r1.3.2.19 ata-raid.c
> --- ata-raid.c 30 Jan 2003 07:19:59 -0000 1.3.2.19
> +++ ata-raid.c 30 Oct 2003 22:28:30 -0000
> @@ -1427,8 +1427,8 @@
> AR_STRATEGY((struct buf *)bp);
>
> if (flags & AR_WAIT) {
> - while ((retry++ < (15*hz/10)) && (error = !(bp->b_flags
& B_DONE)))
> - error = tsleep(bp, PRIBIO, "arrw", 10);
> + /* AR_STRATEGY success or timeout will wake us up */
> + error = tsleep(bp, PRIBIO, "arrw", 0);
> if (!error && (bp->b_flags & B_ERROR))
> error = bp->b_error;
> free(bp, M_AR);
> -----------------------------------------------------------------------
>
> _______________________________________________
> freebsd-stable@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-stable
> To unsubscribe, send any mail to
"freebsd-stable-unsubscribe@freebsd.org"
--
Darcy Buskermolen
Wavefire Technologies Corp.
ph: 250.717.0200
fx: 250.763.1759
http://www.wavefire.com