Date: Sun, 6 Nov 2016 21:57:02 +0100 From: Michael Tuexen <tuexen@freebsd.org> To: Sean Bruno <sbruno@freebsd.org> Cc: freebsd-scsi@freebsd.org, scottl@freebsd.org Subject: Re: mpt driver issue? Message-ID: <CA6234F2-C0BF-472E-B904-FFBAE0ED29F9@freebsd.org> In-Reply-To: <e9698b93-11e6-530a-1d27-ef45706eae66@freebsd.org> References: <CA9B9B19-5311-4A4E-87ED-8771C1000BEC@freebsd.org> <4398FA07-AABF-441E-A412-42BABCD59444@freebsd.org> <e9698b93-11e6-530a-1d27-ef45706eae66@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> On 6 Nov 2016, at 20:41, Sean Bruno <sbruno@freebsd.org> wrote: > > > > On 11/06/16 10:54, Michael Tuexen wrote: >> Dear all, >> >> it was suggested on current@ to forward this to freebsd-scsi@ >> >> The problem shows up since r308268 and it looks like an issue >> in the mpt driver... >> >> Best regards >> Michael > > I don't see any *obvious* reason that we should hold the mpt lock here. > It looks ok to just drop it. I tried that and the system (with WITNESS) panic'ed... Can't write a core (not a surprise). Best regards Michael > > sean > > Index: sys/dev/mpt/mpt.c > =================================================================== > --- sys/dev/mpt/mpt.c (revision 308345) > +++ sys/dev/mpt/mpt.c (working copy) > @@ -2284,9 +2284,7 @@ > mpt_lprt(mpt, MPT_PRT_DEBUG, "doorbell req = %s\n", > mpt_ioc_diag(mpt_read(mpt, MPT_OFFSET_DOORBELL))); > > - MPT_LOCK(mpt); > error = mpt_configure_ioc(mpt, 0, 0); > - MPT_UNLOCK(mpt); > > return (error); > } > > > > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CA6234F2-C0BF-472E-B904-FFBAE0ED29F9>