diff options
| author | David Gwynne <dlg@cvs.openbsd.org> | 2014-06-11 04:28:44 +0000 |
|---|---|---|
| committer | David Gwynne <dlg@cvs.openbsd.org> | 2014-06-11 04:28:44 +0000 |
| commit | 4e23a601c5f649c0e1441318f6cc15eb27d2a914 (patch) | |
| tree | d537195a98ebee765f203dd8c5485c6b6bfe4ef0 | |
| parent | c1d92a61b0a34f143c67dfd663f697d77cde6c33 (diff) | |
em(4) receives jumbos by chaining its MCLBYTES sized descriptors
together, so to receive a full 9000 byte frame, it needs 5 descriptors.
its current mclgeti low watermark is 4.
it appears that the chip will block if it has rxed a packet into
its internal buffers waiting for descriptors. if you're at the lwm,
that means you're DoSed.
this raises the lwm so we can get at least two jumbos, ie, 10
descriptors.
ok deraadt@
| -rw-r--r-- | sys/dev/pci/if_em.c | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/sys/dev/pci/if_em.c b/sys/dev/pci/if_em.c index 5b2b4f256d5..d60a2386aaf 100644 --- a/sys/dev/pci/if_em.c +++ b/sys/dev/pci/if_em.c @@ -31,7 +31,7 @@ POSSIBILITY OF SUCH DAMAGE. ***************************************************************************/ -/* $OpenBSD: if_em.c,v 1.279 2014/03/10 04:09:53 jsg Exp $ */ +/* $OpenBSD: if_em.c,v 1.280 2014/06/11 04:28:43 dlg Exp $ */ /* $FreeBSD: if_em.c,v 1.46 2004/09/29 18:28:28 mlaier Exp $ */ #include <dev/pci/if_em.h> @@ -1859,7 +1859,8 @@ em_setup_interface(struct em_softc *sc) IFQ_SET_MAXLEN(&ifp->if_snd, sc->num_tx_desc - 1); IFQ_SET_READY(&ifp->if_snd); - m_clsetwms(ifp, MCLBYTES, 4, sc->num_rx_desc); + m_clsetwms(ifp, MCLBYTES, 2 * ((ifp->if_hardmtu / MCLBYTES) + 1), + sc->num_rx_desc); ifp->if_capabilities = IFCAP_VLAN_MTU; |