What does system do after "Uptime: "?

[Gavin Atkinson]

On Wed, 2008-04-30 at 15:44 -0400, John Baldwin wrote:
> On Wednesday 30 April 2008 02:49:11 pm Xin LI wrote:
> > Gavin Atkinson wrote:
> > | On Sun, 2008-04-27 at 22:29 -0700, Xin LI wrote:
> > |> Hi,
> > |>
> > |> Just wonder what the system would do after Uptime: during reboot?  Will
> > |> there be any callbacks be called?  I have received some report regarding
> > |> hangs after this, and IIRC the only operation is to IPI BSP to do reboot?
> > |
> > | Several callbacks are called after that message.  Anything that
> > | registers with the shutdown_post_sync or shutdown_final event handler,
> > | for example, and the dump routine if you are using "reboot -d".
> > |
> > | Before trying to diagnose this, however, make sure you are running with
> > | sys/dev/usb/ehci_pci.c 1.30 (which has also been MFC'd to 6/7), as that
> > | seems to have been the cause of most of the recently reported "hangs
> > | after printing uptime" PRs.
> > 
> > I see...  It seems that this is the culprit.
> > 
> > To John: Do you think it's worthy to make this an errata to the
> > supported RELENG_6_X and RELENG_7_0 branches?
> 
> If someone wants to push re@ to do that they can, but my experience with 
> getting re@ to actually get errata fixes has me skeptical (I think I had 
> patches for about 6 or 7 panics or kernel deadlocks for 6.2 that never got 
> pushed as errata patches).

FWIW, there have been probably around 10 PRs in in the last few months
about this behaviour.  I'd vote for it as an errata candidate.

Gavin