Colleagues,
although one week ago we have made a very big merge of the
em(4) driver from HEAD to RELENG_6, during this week the driver
has experienced few important changes in HEAD:
1) Pyun YongHyeon has fixed problems in both TX and RX path
that can happen when system experiences mbuf shortage.
2) Intel's driver version 6.1.4 has been merged.
The patch for RELENG_6 is available here:
http://people.freebsd.org/~glebius/em-RELENG_6.patch
Currently there is also a known problem - a race condition,
that occurs on SMP triggering a fake watchdog event, that
leads to unwanted NIC reset.
--
Totus tuus, Glebius.
GLEBIUS-RIPN GLEB-RIPE
At 06:45 AM 14/08/2006, Gleb Smirnoff wrote:>The patch for RELENG_6 is available here: > >http://people.freebsd.org/~glebius/em-RELENG_6.patch > >Currently there is also a known problem - a race condition, >that occurs on SMP triggering a fake watchdog event, that >leads to unwanted NIC reset.Hi, Is that problem in the driver in RELENG_6, or what is in the patch ? ---Mike
On Mon, Aug 14, 2006 at 08:12:21AM -0400, Mike Tancsa wrote: M> At 06:45 AM 14/08/2006, Gleb Smirnoff wrote: M> M> >The patch for RELENG_6 is available here: M> > M> >http://people.freebsd.org/~glebius/em-RELENG_6.patch M> > M> >Currently there is also a known problem - a race condition, M> >that occurs on SMP triggering a fake watchdog event, that M> >leads to unwanted NIC reset. M> M> Is that problem in the driver in RELENG_6, or what is in the patch ? It is present both in RELENG_6 and in HEAD. -- Totus tuus, Glebius. GLEBIUS-RIPN GLEB-RIPE
...> although one week ago we have made a very big merge of the >em(4) driver from HEAD to RELENG_6, during this week the driver >has experienced few important changes in HEAD: > > 1) Pyun YongHyeon has fixed problems in both TX and RX paththat can happen when system experiences mbuf shortage. ... Does this have any relation to the post "bce0: Error mapping mbuf into TX chain"? http://lists.freebsd.org/pipermail/freebsd-stable/2006-August/027449.htm l I also have the above problem (with bce0) on a Dell PowerEdge 2950, although it doesn't seem to take much io to crash the interface. Thanks, Bucky