Skip to contentSkip to conversation

rdma: Add support for the netlink extack

PATCH In review Leon Romanovsky 2 participants 3 messages Jan 3, 2021 – Jan 11, 2021

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 rdma: Add support for the netlink extack

From: Patrisious Haddad <[email protected]>

Add support in rdma for extack errors to be received in userspace when sent from kernel, so now netlink extack error messages sent from kernel would be printed for the user.

---
David,

Just as a note, rdmatool is heavily influenced by the devlink and general code should probably be applicable for both tools. Most likely that any core refactoring/fix in the devlink is needed for rdmatool too.

Thanks
----
Changelog:
v2: Reused already existing function to set extack.
v1: https://lore.kernel.org/linux-rdma/[email protected]

---
 rdma/rdma.h  |  1 +
 rdma/utils.c | 24 ++++--------------------
 2 files changed, 5 insertions(+), 20 deletions(-)
rdma/rdma.h +1 -0
index fc8bcf09..470e11c8 100644
@@ -19,6 +19,7 @@
 
 #include "list.h"
 #include "utils.h"
+#include "mnl_utils.h"
 #include "json_print.h"
 
 #define pr_err(args...) fprintf(stderr, ##args)
rdma/utils.c +4 -20
index 2a201aa4..903a544c 100644
@@ -666,18 +666,12 @@ int rd_send_msg(struct rd *rd)
 {
 	int ret;
 
-	rd->nl = mnl_socket_open(NETLINK_RDMA);
+	rd->nl = mnlu_socket_open(NETLINK_RDMA);
 	if (!rd->nl) {
 		pr_err("Failed to open NETLINK_RDMA socket\n");
 		return -ENODEV;
 	}
 
-	ret = mnl_socket_bind(rd->nl, 0, MNL_SOCKET_AUTOPID);
-	if (ret < 0) {
-		pr_err("Failed to bind socket with err %d\n", ret);
-		goto err;
-	}
-
 	ret = mnl_socket_sendto(rd->nl, rd->nlh, rd->nlh->nlmsg_len);
 	if (ret < 0) {
 		pr_err("Failed to send to socket with err %d\n", ret);
@@ -692,23 +686,13 @@ err:
 
 int rd_recv_msg(struct rd *rd, mnl_cb_t callback, void *data, unsigned int seq)
 {
-	int ret;
-	unsigned int portid;
 	char buf[MNL_SOCKET_BUFFER_SIZE];
+	int ret;
 
-	portid = mnl_socket_get_portid(rd->nl);
-	do {
-		ret = mnl_socket_recvfrom(rd->nl, buf, sizeof(buf));
-		if (ret <= 0)
-			break;
-
-		ret = mnl_cb_run(buf, ret, seq, portid, callback, data);
-	} while (ret > 0);
-
+	ret = mnlu_socket_recv_run(rd->nl, seq, buf, MNL_SOCKET_BUFFER_SIZE,
+				   callback, data);
 	if (ret < 0 && !rd->suppress_errors)
 		perror("error");
-
-	mnl_socket_close(rd->nl);
 	return ret;
 }
 

--
2.29.2

#1 lore raw
Quoted 7 lines · Leon Romanovsky

From: Patrisious Haddad <[email protected]>

Add support in rdma for extack errors to be received in userspace when sent from kernel, so now netlink extack error messages sent from kernel would be printed for the user.

Quoted 9 lines

---
David,

Just as a note, rdmatool is heavily influenced by the devlink and general code should probably be applicable for both tools. Most likely that any core refactoring/fix in the devlink is needed for rdmatool too.

understood and it was not the best model to start with but here we are.

Petr did a good job of refactoring when he added dcb, but rdma was slightly different so the refactoring did not update it.

#2 lore raw
Quoted 18 lines · David Ahern
Quoted 7 lines · Leon Romanovsky

From: Patrisious Haddad <[email protected]>

Add support in rdma for extack errors to be received in userspace when sent from kernel, so now netlink extack error messages sent from kernel would be printed for the user.

Quoted 9 lines

---
David,

Just as a note, rdmatool is heavily influenced by the devlink and general code should probably be applicable for both tools. Most likely that any core refactoring/fix in the devlink is needed for rdmatool too.

Quoted 4 lines

understood and it was not the best model to start with but here we are.

Petr did a good job of refactoring when he added dcb, but rdma was slightly different so the refactoring did not update it.

I can work to reduce the gaps, just need to know them.

Thanks