ad0: WRITE command timeout

[Hiten Pandya]

Jonathon McKitrick wrote:
> On Mon, Mar 01, 2004 at 08:39:56AM +0900, YONETANI Tomokazu wrote:
> Sorry none of these are real answers, but I replaced DFly with 4.8,
> and since it works, I had to keep it to get some work done.  But the
> replacement drive will help answer these questions, hopefully.
>
	FWIW, I just started receiving a variant of this error in my
	VMware-4 DragonFly:
	ad0: WRITE command timeout tag=0 serv=0 - resetting
	ata0: resetting devices done
	I have not tried to reproduce this scenario yet, but this
	seems to happen when I do a large transfer from my NFS share
	into /usr/src (i.e. a cvs checkout).  The machine does not
	hang, but key input through main console is out of question
	although networking connections stay established, but if you
	initiate, say a new SSH session, it will close the rest of
	them and kill networking ability too; the CVS process goes into
	sbwait and is blocking forever:
	load: 0.00  cmd: cvs 662 [sbwait] 50.72u 252.10s 0% 1296k

	My bet is on the ATA DMA code... although I need to do a
	little more testing and find out what causes this annoyance
	and pure pain in the ass.
		-Hiten
		Hiten Pandya
		hmp at xxxxxxxxxxxxx || hmp at xxxxxxxxxxx