The aic94xx still isn't stable in complex topologies. This patch fixes
a panic in a two expander two device wide port configuration. However,
there's still a fairly nasty soft lockup on module insertion while it's
scanning the ports.
James
Index: BUILD-2.6/drivers/scsi/sas/sas_expander.c
===================================================================
--- BUILD-2.6.orig/drivers/scsi/sas/sas_expander.c 2006-07-03 04:42:18.000000000 -0500
+++ BUILD-2.6/drivers/scsi/sas/sas_expander.c 2006-07-03 04:42:37.000000000 -0500
@@ -683,15 +683,12 @@
BUG();
}
port = parent->port;
- port->phy = phy->phy;
- port->port = phy->port;
child->rphy = rphy;
edev = rphy_to_expander_device(rphy);
child->dev_type = phy->attached_dev_type;
child->parent = parent;
port->port_dev = child;
child->port = port;
- child->port->port = phy->port;
child->iproto = phy->attached_iproto;
child->tproto = phy->attached_tproto;
memcpy(child->sas_addr, phy->attached_sas_addr, SAS_ADDR_SIZE);
@@ -702,6 +699,11 @@
edev->level);
sas_init_dev(child);
sas_rphy_add(child->rphy);
+
+ spin_lock(&parent->port->dev_list_lock);
+ list_add_tail(&child->dev_list_node, &parent->port->dev_list);
+ spin_unlock(&parent->port->dev_list_lock);
+
res = sas_discover_expander(child);
if (res) {
kfree(child);
-
: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
