netfilter: don't use mutex_lock_interruptible()

Eric Dumazet reports that getsockopt() or setsockopt() sometimes
returns -EINTR instead of -ENOPROTOOPT, causing headaches to
application developers.

This patch replaces all the mutex_lock_interruptible() by mutex_lock()
in the netfilter tree, as there is no reason we should sleep for a
long time there.

Reported-by: Eric Dumazet <edumazet@google.com>
Suggested-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Acked-by: Julian Anastasov <ja@ssi.bg>
This commit is contained in:
Pablo Neira Ayuso
2014-07-31 20:38:46 +02:00
parent b88825de85
commit 7926dbfa4b
5 changed files with 22 additions and 73 deletions

View File

@@ -327,10 +327,7 @@ find_inlist_lock_noload(struct list_head *head, const char *name, int *error,
char name[EBT_FUNCTION_MAXNAMELEN];
} *e;
*error = mutex_lock_interruptible(mutex);
if (*error != 0)
return NULL;
mutex_lock(mutex);
list_for_each_entry(e, head, list) {
if (strcmp(e->name, name) == 0)
return e;
@@ -1203,10 +1200,7 @@ ebt_register_table(struct net *net, const struct ebt_table *input_table)
table->private = newinfo;
rwlock_init(&table->lock);
ret = mutex_lock_interruptible(&ebt_mutex);
if (ret != 0)
goto free_chainstack;
mutex_lock(&ebt_mutex);
list_for_each_entry(t, &net->xt.tables[NFPROTO_BRIDGE], list) {
if (strcmp(t->name, table->name) == 0) {
ret = -EEXIST;