crda: rename nl_handle to nl_sock for libnl-2.0

Upstream has renamed nl_handle to nl_sock. Update crda to the new names and add #define for libnl-1.1.

---

I'm not too proud to admit that I did proofread this, twice, and still missed changing iw to crda in the commit message.

In the future, should I keep pushing these sorts of patches, or should we back them out and wait for a 'release' of libnl-2.0 and push them all in then?

index afc5df1..f2c471f 100644
@@ -43,10 +43,11 @@ static inline int __genl_ctrl_alloc_cache(struct

nl_handle *h, struct nl_cache *
}

#define genl_ctrl_alloc_cache __genl_ctrl_alloc_cache
+#define nl_sock nl_handle
#endif /* CONFIG_LIBNL20 */

struct nl80211_state {
- struct nl_handle *nl_handle;
+ struct nl_sock *nl_sock;
struct nl_cache *nl_cache;
struct genl_family *nl80211;
};
@@ -55,22 +56,22 @@ static int nl80211_init(struct nl80211_state *state)
{
int err;

- state->nl_handle = nl_socket_alloc();
- if (!state->nl_handle) {
- fprintf(stderr, "Failed to allocate netlink handle.\n");
+ state->nl_sock = nl_socket_alloc();
+ if (!state->nl_sock) {
+ fprintf(stderr, "Failed to allocate netlink sock.\n");
return -ENOMEM;
}

- if (genl_connect(state->nl_handle)) {
+ if (genl_connect(state->nl_sock)) {
fprintf(stderr, "Failed to connect to generic netlink.\n"); err = -ENOLINK;
- goto out_handle_destroy;
+ goto out_sock_destroy;
}

- if (genl_ctrl_alloc_cache(state->nl_handle, &state->nl_cache)) {
+ if (genl_ctrl_alloc_cache(state->nl_sock, &state->nl_cache)) {
fprintf(stderr, "Failed to allocate generic netlink cache.\n"); err = -ENOMEM;
- goto out_handle_destroy;
+ goto out_sock_destroy;
}

state->nl80211 = genl_ctrl_search_by_name(state->nl_cache, "nl80211");
@@ -84,8 +85,8 @@ static int nl80211_init(struct nl80211_state *state)

out_cache_free:
nl_cache_free(state->nl_cache);
- out_handle_destroy:
- nl_socket_free(state->nl_handle);
+ out_sock_destroy:
+ nl_socket_free(state->nl_sock);
return err;
}

@@ -93,7 +94,7 @@ static void nl80211_cleanup(struct nl80211_state *state)
{
genl_family_put(state->nl80211); nl_cache_free(state->nl_cache);
- nl_socket_free(state->nl_handle);
+ nl_socket_free(state->nl_sock);
}

static int reg_handler(struct nl_msg __attribute__((unused)) *msg,
@@ -294,7 +295,7 @@ int main(int argc, char **argv)
if (!cb)
goto cb_out;

- r = nl_send_auto_complete(nlstate.nl_handle, msg);
+ r = nl_send_auto_complete(nlstate.nl_sock, msg);

if (r < 0) {
fprintf(stderr, "Failed to send regulatory request: %d\n", r);
@@ -306,7 +307,7 @@ int main(int argc, char **argv)
nl_cb_err(cb, NL_CB_CUSTOM, error_handler, NULL);

if (!finished) {
- r = nl_wait_for_ack(nlstate.nl_handle);
+ r = nl_wait_for_ack(nlstate.nl_sock);
if (r < 0) {
fprintf(stderr, "Failed to set regulatory domain: "
"%d\n", r);

But libnl-2.0 comes with a define too:
#define nl_handle nl_sock

you just need to include the right headers, no?

johannes

dunno.... but I can't find it:

libnl $ egrep "#define nl_handle" * -r
libnl $ git pull
Already up-to-date.

Pat

Ah. Grr. Just recently removed.

johannes

Which leads back around to my question, should I continue pushing patches like this your way (after making sure the right patch is on the right e-mail with the right description), or do we want to wait and I'll just maintain the patches outside of git, and push them once libnl-2.0 hits? On one hand, it makes testing both versions easier, on the other, it makes a lot of little 'fix' commits like this. I have no issues maintaining these sorts of patches, but I don't want to do it if you don't want them.

Pat

Heh :) Also make sure your patch isn't line-wrapped next time, please. The iw resend was line-wrapped.

No, it's fine, I've committed this now, thanks. I had just thought it ought to work, based on testing with git snapshot a couple of weeks ago...

Not that libnl-2.0 compiles cleanly on my system...

johannes

I'll gladly take them, sure, actually thanks!

Luis