diff options
| author | David Gwynne <dlg@cvs.openbsd.org> | 2011-03-01 23:48:34 +0000 |
|---|---|---|
| committer | David Gwynne <dlg@cvs.openbsd.org> | 2011-03-01 23:48:34 +0000 |
| commit | cae9210ee31c58a7ed6cabd40c84011c6893b1ce (patch) | |
| tree | 977e664b8f3c379222ce0668035848cc8a11b82e /sys/dev | |
| parent | 65c371d965b13cb128b88d417edee06d7a8c2cfd (diff) | |
back out r1.162, the one that bumps openings up on sas and fc
devices.
my theory is that some devices report queue full conditions in ways
the firmware doesnt understand, or some firmwares default to NOT
doing the queue full handling internally. either way it reports
queue full conditions as faulted io which gets passed up to the
block layer as errors.
this makes us conservative again and safe.
this fixes panics from ajacout
ok sthen@ deraadt@
Diffstat (limited to 'sys/dev')
| -rw-r--r-- | sys/dev/ic/mpi.c | 7 |
1 files changed, 2 insertions, 5 deletions
diff --git a/sys/dev/ic/mpi.c b/sys/dev/ic/mpi.c index 7490178bd48..7b1715a9b0e 100644 --- a/sys/dev/ic/mpi.c +++ b/sys/dev/ic/mpi.c @@ -1,4 +1,4 @@ -/* $OpenBSD: mpi.c,v 1.165 2010/09/24 01:27:11 dlg Exp $ */ +/* $OpenBSD: mpi.c,v 1.166 2011/03/01 23:48:33 dlg Exp $ */ /* * Copyright (c) 2005, 2006, 2009 David Gwynne <dlg@openbsd.org> @@ -344,10 +344,7 @@ mpi_attach(struct mpi_softc *sc) sc->sc_link.adapter_softc = sc; sc->sc_link.adapter_target = sc->sc_target; sc->sc_link.adapter_buswidth = sc->sc_buswidth; - if (sc->sc_porttype == MPI_PORTFACTS_PORTTYPE_SCSI) - sc->sc_link.openings = sc->sc_maxcmds / sc->sc_buswidth; - else - sc->sc_link.openings = sc->sc_maxcmds; + sc->sc_link.openings = sc->sc_maxcmds / sc->sc_buswidth; sc->sc_link.pool = &sc->sc_iopool; bzero(&saa, sizeof(saa)); |