| Message ID | 20150727195631.26840.93266.stgit@ahduyck-vm-fedora22 |
|---|---|
| State | Accepted, archived |
| Delegated to: | David Miller |
| Headers | show |
On 7/27/15 2:08 PM, Alexander Duyck wrote: > It was reported that update_suffix was taking a long time on systems where > a large number of leaves were attached to a single node. As it turns out > fib_table_flush was calling update_suffix for each leaf that didn't have all > of the aliases stripped from it. As a result, on this large node removing > one leaf would result in us calling update_suffix for every other leaf on > the node. > > The fix is to just remove the calls to leaf_pull_suffix since they are > redundant as we already have a call in resize that will go through and > update the suffix length for the node before we exit out of > fib_table_flush or fib_table_flush_external. > > Reported-by: David Ahern<dsa@cumulusnetworks.com> > Signed-off-by: Alexander Duyck<alexander.h.duyck@redhat.com> > --- > > This patch should apply to linux-4.1.y and newer kernels. > > I've done a bit of testing on my system and I no longer see update_suffix > dominating the performance traces. David if you can test with this patch > to see if you still see the issue I would appreciate it. > Works for me. Thanks. Tested-by: David Ahern <dsa@cumulusnetworks.com> -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
From: Alexander Duyck <alexander.h.duyck@redhat.com> Date: Mon, 27 Jul 2015 13:08:06 -0700 > It was reported that update_suffix was taking a long time on systems where > a large number of leaves were attached to a single node. As it turns out > fib_table_flush was calling update_suffix for each leaf that didn't have all > of the aliases stripped from it. As a result, on this large node removing > one leaf would result in us calling update_suffix for every other leaf on > the node. > > The fix is to just remove the calls to leaf_pull_suffix since they are > redundant as we already have a call in resize that will go through and > update the suffix length for the node before we exit out of > fib_table_flush or fib_table_flush_external. > > Reported-by: David Ahern <dsa@cumulusnetworks.com> > Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com> Applied and queued up for -stable, thanks. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c index ef90d73911de..70168ca4716b 100644 --- a/net/ipv4/fib_trie.c +++ b/net/ipv4/fib_trie.c @@ -1791,8 +1791,6 @@ void fib_table_flush_external(struct fib_table *tb) if (hlist_empty(&n->leaf)) { put_child_root(pn, n->key, NULL); node_free(n); - } else { - leaf_pull_suffix(pn, n); } } } @@ -1862,8 +1860,6 @@ int fib_table_flush(struct fib_table *tb) if (hlist_empty(&n->leaf)) { put_child_root(pn, n->key, NULL); node_free(n); - } else { - leaf_pull_suffix(pn, n); } }
It was reported that update_suffix was taking a long time on systems where a large number of leaves were attached to a single node. As it turns out fib_table_flush was calling update_suffix for each leaf that didn't have all of the aliases stripped from it. As a result, on this large node removing one leaf would result in us calling update_suffix for every other leaf on the node. The fix is to just remove the calls to leaf_pull_suffix since they are redundant as we already have a call in resize that will go through and update the suffix length for the node before we exit out of fib_table_flush or fib_table_flush_external. Reported-by: David Ahern <dsa@cumulusnetworks.com> Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com> --- This patch should apply to linux-4.1.y and newer kernels. I've done a bit of testing on my system and I no longer see update_suffix dominating the performance traces. David if you can test with this patch to see if you still see the issue I would appreciate it. net/ipv4/fib_trie.c | 4 ---- 1 file changed, 4 deletions(-) -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html