targclose doesn't return

[Matthew Jacob]

Neither am I. I had to fool around with this at one point to get it to
work and it's probably gotten broken again as other parts of the
kernel get more lock clean. Sigh.

On 4/3/07, Scott Long <scottl at samsco.org> wrote:
> My guess is that the kninit() in targopen isn't being cleaned up from in
> targclose.  I'm not terribly familiar in how the knote API works,
> though.
>
> Scott
>
>
> Matthew Jacob wrote:
> > Yeah- I've seen this too.
> >
> > On 4/3/07, Alex Dupre <ale at freebsd.org> wrote:
> >> Scott Long wrote:
> >> > Are there any other console messages from the targ driver?  Can you
> >> > turn on CAMDEBUG and send us the trace of what is going on?
> >>
> >> CAMDEBUG is already on, but simply opening/closing the targ device,
> >> without sending any ioctl to enable it, shouldn't produce any CAM
> >> message. I traced the kernel thread and I found that it doesn't return
> >> from the destroy_devl() function: csw->d_purge is NULL and
> >> dev->si_threadcount is '1'. The thread enters the following block
> >> (kern_conf.c, row 690) and never exits.
> >>
> >>
> >>        while (dev->si_threadcount != 0) {
> >>                /* Use unique dummy wait ident */
> >>                msleep(&csw, &devmtx, PRIBIO, "devdrn", hz / 10);
> >>        }
> >>
> >>
> >> --
> >> Alex Dupre
> >> _______________________________________________
> >> freebsd-scsi at freebsd.org mailing list
> >> http://lists.freebsd.org/mailman/listinfo/freebsd-scsi
> >> To unsubscribe, send any mail to "freebsd-scsi-unsubscribe at freebsd.org"
> >>
>
>