> Testing dev->mthca_flags & MTHCA_FLAG_FMR would be cleaner I think. No?
Yes, you're right. If Pradeep confirms this fixes things for him,
I'll queue the following:
diff --git a/drivers/infiniband/hw/mthca/mthca_mr.c
b/drivers/infiniband/hw/mthca/mthca_mr.c
index fdb576d..2ebebab 100644
--- a/drivers/infiniband/hw/mthca/mthca_mr.c
+++ b/drivers/infiniband/hw/mthca/mthca_mr.c
@@ -355,7 +355,8 @@ int mthca_write_mtt(struct mthca_dev *dev, struct mthca_mtt
*mtt,
int size = mthca_write_mtt_size(dev);
int chunk;
- if (dev->mr_table.fmr_mtt_buddy != &dev->mr_table.mtt_buddy)
+ if (dev->mr_table.fmr_mtt_buddy != &dev->mr_table.mtt_buddy ||
+ !(dev->mthca_flags & MTHCA_FLAG_FMR))
return __mthca_write_mtt(dev, mtt, start_index, buffer_list,
list_len);
while (list_len > 0) {
_______________________________________________
general mailing list
[EMAIL PROTECTED]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
