I think I have this figured out: if you unload ib_ipoib and
ib_sa_query in quick succession, ib_ipoib sends MCMember requests to
the SA to leave its multicast groups. Normally, because IPoIB sets a
timeout of 0, no callback is generated and so it's fine that IPoIB
passes a NULL callback. However, if ib_sa_query is unloaded right
afterwards, the send of the request doesn't get a chance to complete
and so a cancel callback is generated.
If this crash is at all reproducible for you, can you try this patch
and see if it helps?
Thanks,
Roland
--- infiniband/core/sa_query.c (revision 1781)
+++ infiniband/core/sa_query.c (working copy)
@@ -587,7 +587,7 @@
init_mad(query->sa_query.mad, agent);
- query->sa_query.callback = ib_sa_path_rec_callback;
+ query->sa_query.callback = callback ?
ib_sa_path_rec_callback : NULL;
query->sa_query.release = ib_sa_path_rec_release;
query->sa_query.port = port;
query->sa_query.mad->mad_hdr.method = IB_MGMT_METHOD_GET;
@@ -663,7 +663,7 @@
init_mad(query->sa_query.mad, agent);
- query->sa_query.callback = ib_sa_mcmember_rec_callback;
+ query->sa_query.callback = callback ?
ib_sa_mcmember_rec_callback : NULL;
query->sa_query.release = ib_sa_mcmember_rec_release;
query->sa_query.port = port;
query->sa_query.mad->mad_hdr.method = method;
@@ -698,20 +698,21 @@
if (!query)
return;
- switch (mad_send_wc->status) {
- case IB_WC_SUCCESS:
- /* No callback -- already got recv */
- break;
- case IB_WC_RESP_TIMEOUT_ERR:
- query->callback(query, -ETIMEDOUT, NULL);
- break;
- case IB_WC_WR_FLUSH_ERR:
- query->callback(query, -EINTR, NULL);
- break;
- default:
- query->callback(query, -EIO, NULL);
- break;
- }
+ if (query->callback)
+ switch (mad_send_wc->status) {
+ case IB_WC_SUCCESS:
+ /* No callback -- already got recv */
+ break;
+ case IB_WC_RESP_TIMEOUT_ERR:
+ query->callback(query, -ETIMEDOUT, NULL);
+ break;
+ case IB_WC_WR_FLUSH_ERR:
+ query->callback(query, -EINTR, NULL);
+ break;
+ default:
+ query->callback(query, -EIO, NULL);
+ break;
+ }
dma_unmap_single(agent->device->dma_device,
pci_unmap_addr(query, mapping),
@@ -736,7 +737,7 @@
query = idr_find(&query_idr, mad_recv_wc->wc->wr_id);
spin_unlock_irqrestore(&idr_lock, flags);
- if (query) {
+ if (query && query->callback) {
if (mad_recv_wc->wc->status == IB_WC_SUCCESS)
query->callback(query,
mad_recv_wc->recv_buf.mad->mad_hdr.status ?
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
