All of lore.kernel.org
 help / color / mirror / Atom feed
From: Kuniyuki Iwashima <kuniyu@amazon.com>
To: "David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>
Cc: Kuniyuki Iwashima <kuniyu@amazon.com>,
	Kuniyuki Iwashima <kuni1840@gmail.com>, <netdev@vger.kernel.org>
Subject: [PATCH v5 net-next 10/15] af_unix: Skip GC if no cycle exists.
Date: Mon, 25 Mar 2024 13:24:20 -0700	[thread overview]
Message-ID: <20240325202425.60930-11-kuniyu@amazon.com> (raw)
In-Reply-To: <20240325202425.60930-1-kuniyu@amazon.com>

We do not need to run GC if there is no possible cyclic reference.
We use unix_graph_maybe_cyclic to decide if we should run GC.

If a fd of an AF_UNIX socket is passed to an already inflight AF_UNIX
socket, they could form a cyclic reference.  Then, we set true to
unix_graph_maybe_cyclic and later run Tarjan's algorithm to group
them into SCC.

Once we run Tarjan's algorithm, we are 100% sure whether cyclic
references exist or not.  If there is no cycle, we set false to
unix_graph_maybe_cyclic and can skip the entire garbage collection
next time.

When finalising SCC, we set true to unix_graph_maybe_cyclic if SCC
consists of multiple vertices.

Even if SCC is a single vertex, a cycle might exist as self-fd passing.
Given the corner case is rare, we detect it by checking all edges of
the vertex and set true to unix_graph_maybe_cyclic.

With this change, __unix_gc() is just a spin_lock() dance in the normal
usage.

Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
---
 net/unix/garbage.c | 48 +++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 47 insertions(+), 1 deletion(-)

diff --git a/net/unix/garbage.c b/net/unix/garbage.c
index cec6189126ba..5a1fae78d6dc 100644
--- a/net/unix/garbage.c
+++ b/net/unix/garbage.c
@@ -112,6 +112,19 @@ static struct unix_vertex *unix_edge_successor(struct unix_edge *edge)
 	return edge->successor->vertex;
 }
 
+static bool unix_graph_maybe_cyclic;
+
+static void unix_update_graph(struct unix_vertex *vertex)
+{
+	/* If the receiver socket is not inflight, no cyclic
+	 * reference could be formed.
+	 */
+	if (!vertex)
+		return;
+
+	unix_graph_maybe_cyclic = true;
+}
+
 static LIST_HEAD(unix_unvisited_vertices);
 
 enum unix_vertex_index {
@@ -138,12 +151,16 @@ static void unix_add_edge(struct scm_fp_list *fpl, struct unix_edge *edge)
 
 	vertex->out_degree++;
 	list_add_tail(&edge->vertex_entry, &vertex->edges);
+
+	unix_update_graph(unix_edge_successor(edge));
 }
 
 static void unix_del_edge(struct scm_fp_list *fpl, struct unix_edge *edge)
 {
 	struct unix_vertex *vertex = edge->predecessor->vertex;
 
+	unix_update_graph(unix_edge_successor(edge));
+
 	list_del(&edge->vertex_entry);
 	vertex->out_degree--;
 
@@ -227,6 +244,7 @@ void unix_del_edges(struct scm_fp_list *fpl)
 void unix_update_edges(struct unix_sock *receiver)
 {
 	spin_lock(&unix_gc_lock);
+	unix_update_graph(unix_sk(receiver->listener)->vertex);
 	receiver->listener = NULL;
 	spin_unlock(&unix_gc_lock);
 }
@@ -268,6 +286,26 @@ void unix_destroy_fpl(struct scm_fp_list *fpl)
 	unix_free_vertices(fpl);
 }
 
+static bool unix_scc_cyclic(struct list_head *scc)
+{
+	struct unix_vertex *vertex;
+	struct unix_edge *edge;
+
+	/* SCC containing multiple vertices ? */
+	if (!list_is_singular(scc))
+		return true;
+
+	vertex = list_first_entry(scc, typeof(*vertex), scc_entry);
+
+	/* Self-reference or a embryo-listener circle ? */
+	list_for_each_entry(edge, &vertex->edges, vertex_entry) {
+		if (unix_edge_successor(edge) == vertex)
+			return true;
+	}
+
+	return false;
+}
+
 static LIST_HEAD(unix_visited_vertices);
 static unsigned long unix_vertex_grouped_index = UNIX_VERTEX_INDEX_MARK2;
 
@@ -353,6 +391,9 @@ static void __unix_walk_scc(struct unix_vertex *vertex)
 			vertex->index = unix_vertex_grouped_index;
 		}
 
+		if (!unix_graph_maybe_cyclic)
+			unix_graph_maybe_cyclic = unix_scc_cyclic(&scc);
+
 		list_del(&scc);
 	}
 
@@ -363,6 +404,8 @@ static void __unix_walk_scc(struct unix_vertex *vertex)
 
 static void unix_walk_scc(void)
 {
+	unix_graph_maybe_cyclic = false;
+
 	/* Visit every vertex exactly once.
 	 * __unix_walk_scc() moves visited vertices to unix_visited_vertices.
 	 */
@@ -524,6 +567,9 @@ static void __unix_gc(struct work_struct *work)
 
 	spin_lock(&unix_gc_lock);
 
+	if (!unix_graph_maybe_cyclic)
+		goto skip_gc;
+
 	unix_walk_scc();
 
 	/* First, select candidates for garbage collection.  Only
@@ -617,7 +663,7 @@ static void __unix_gc(struct work_struct *work)
 
 	/* All candidates should have been detached by now. */
 	WARN_ON_ONCE(!list_empty(&gc_candidates));
-
+skip_gc:
 	/* Paired with READ_ONCE() in wait_for_unix_gc(). */
 	WRITE_ONCE(gc_in_progress, false);
 
-- 
2.30.2


  parent reply	other threads:[~2024-03-25 20:29 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-03-25 20:24 [PATCH v5 net-next 00/15] af_unix: Rework GC Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 01/15] af_unix: Allocate struct unix_vertex for each inflight AF_UNIX fd Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 02/15] af_unix: Allocate struct unix_edge " Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 03/15] af_unix: Link struct unix_edge when queuing skb Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 04/15] af_unix: Bulk update unix_tot_inflight/unix_inflight " Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 05/15] af_unix: Iterate all vertices by DFS Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 06/15] af_unix: Detect Strongly Connected Components Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 07/15] af_unix: Save listener for embryo socket Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 08/15] af_unix: Fix up unix_edge.successor " Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 09/15] af_unix: Save O(n) setup of Tarjan's algo Kuniyuki Iwashima
2024-03-25 20:24 ` Kuniyuki Iwashima [this message]
2024-03-25 20:24 ` [PATCH v5 net-next 11/15] af_unix: Avoid Tarjan's algorithm if unnecessary Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 12/15] af_unix: Assign a unique index to SCC Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 13/15] af_unix: Detect dead SCC Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 14/15] af_unix: Replace garbage collection algorithm Kuniyuki Iwashima
2024-03-25 20:24 ` [PATCH v5 net-next 15/15] selftest: af_unix: Test GC for SCM_RIGHTS Kuniyuki Iwashima
2024-03-29  9:55 ` [PATCH v5 net-next 00/15] af_unix: Rework GC Paolo Abeni
2024-03-29 15:50 ` patchwork-bot+netdevbpf

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20240325202425.60930-11-kuniyu@amazon.com \
    --to=kuniyu@amazon.com \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=kuba@kernel.org \
    --cc=kuni1840@gmail.com \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.