rpms
/
systemd
2
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
systemd/18915.patch

155 lines
7.2 KiB
Diff
Raw Blame History

From 8b0f54c9290564e8c27c9c8ac464cdcc2c659ad5 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Sat, 6 Mar 2021 19:06:08 +0100
Subject: [PATCH 1/3] pid1: return varlink error on the right connection
---
src/core/core-varlink.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/core/core-varlink.c b/src/core/core-varlink.c
index d695106658b..b3df8cd893c 100644
--- a/src/core/core-varlink.c
+++ b/src/core/core-varlink.c
@@ -142,7 +142,7 @@ static int vl_method_subscribe_managed_oom_cgroups(
/* We only take one subscriber for this method so return an error if there's already an existing one.
* This shouldn't happen since systemd-oomd is the only client of this method. */
if (FLAGS_SET(flags, VARLINK_METHOD_MORE) && m->managed_oom_varlink_request)
- return varlink_error(m->managed_oom_varlink_request, VARLINK_ERROR_SUBSCRIPTION_TAKEN, NULL);
+ return varlink_error(link, VARLINK_ERROR_SUBSCRIPTION_TAKEN, NULL);
r = json_build(&arr, JSON_BUILD_EMPTY_ARRAY);
if (r < 0)
@@ -188,6 +188,7 @@ static int vl_method_subscribe_managed_oom_cgroups(
if (!FLAGS_SET(flags, VARLINK_METHOD_MORE))
return varlink_reply(link, v);
+ assert(!m->managed_oom_varlink_request);
m->managed_oom_varlink_request = varlink_ref(link);
return varlink_notify(m->managed_oom_varlink_request, v);
}
@@ -475,8 +476,7 @@ void manager_varlink_done(Manager *m) {
assert(m);
/* Send the final message if we still have a subscribe request open. */
- if (m->managed_oom_varlink_request)
- m->managed_oom_varlink_request = varlink_close_unref(m->managed_oom_varlink_request);
+ m->managed_oom_varlink_request = varlink_close_unref(m->managed_oom_varlink_request);
m->varlink_server = varlink_server_unref(m->varlink_server);
}
From 39ad3f1c092b5dffcbb4b1d12eb9ca407f010a3c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Sun, 7 Mar 2021 16:42:35 +0100
Subject: [PATCH 2/3] varlink: avoid using dangling ref in
varlink_close_unref()
Fixes #18025, https://bugzilla.redhat.com/show_bug.cgi?id=1931034.
We drop the reference stored in Manager.managed_oom_varlink_request in two code paths:
vl_disconnect() which is installed as a disconnect callback, and in manager_varlink_done().
But we also make a disconnect from manager_varlink_done(). So we end up with the following
call stack:
(gdb) bt
0 vl_disconnect (s=0x112c7b0, link=0xea0070, userdata=0xe9bcc0) at ../src/core/core-varlink.c:414
1 0x00007f1366e9d5ac in varlink_detach_server (v=0xea0070) at ../src/shared/varlink.c:1210
2 0x00007f1366e9d664 in varlink_close (v=0xea0070) at ../src/shared/varlink.c:1228
3 0x00007f1366e9d6b5 in varlink_close_unref (v=0xea0070) at ../src/shared/varlink.c:1240
4 0x0000000000524629 in manager_varlink_done (m=0xe9bcc0) at ../src/core/core-varlink.c:479
5 0x000000000048ef7b in manager_free (m=0xe9bcc0) at ../src/core/manager.c:1357
6 0x000000000042602c in main (argc=5, argv=0x7fff439c43d8) at ../src/core/main.c:2909
When we enter vl_disconnect(), m->managed_oom_varlink_request.n_ref==1.
When we exit from vl_discconect(), m->managed_oom_varlink_request==NULL. But
varlink_close_unref() has a copy of the pointer in *v. When we continue executing
varlink_close_unref(), this pointer is dangling, and the call to varlink_unref()
is done with an invalid pointer.
---
src/shared/varlink.c | 33 +++++++++++++++++++++++++--------
1 file changed, 25 insertions(+), 8 deletions(-)
diff --git a/src/shared/varlink.c b/src/shared/varlink.c
index 31128e02e06..6ed72075ba5 100644
--- a/src/shared/varlink.c
+++ b/src/shared/varlink.c
@@ -1206,8 +1206,9 @@ int varlink_close(Varlink *v) {
varlink_set_state(v, VARLINK_DISCONNECTED);
- /* Let's take a reference first, since varlink_detach_server() might drop the final (dangling) ref
- * which would destroy us before we can call varlink_clear() */
+ /* Let's take a reference first, since varlink_detach_server() might drop the final ref from the
+ * disconnect callback, which would invalidate the pointer we are holding before we can call
+ * varlink_clear(). */
varlink_ref(v);
varlink_detach_server(v);
varlink_clear(v);
@@ -1220,17 +1221,33 @@ Varlink* varlink_close_unref(Varlink *v) {
if (!v)
return NULL;
- (void) varlink_close(v);
+ /* A reference is given to us to be destroyed. But when calling varlink_close(), a callback might
+ * also drop a reference. We allow this, and will hold a temporary reference to the object to make
+ * sure that the object still exists when control returns to us. If there's just one reference
+ * remaining after varlink_close(), even though there were at least two right before, we'll handle
+ * that gracefully instead of crashing.
+ *
+ * In other words, this call drops the donated reference, but if the internal call to varlink_close()
+ * dropped a reference to, we don't drop the reference afain. This allows the caller to say:
+ * global_object->varlink = varlink_close_unref(global_object->varlink);
+ * even though there is some callback which has access to global_object and may drop the reference
+ * stored in global_object->varlink. Without this step, the same code would have to be written as:
+ * Varlink *t = TAKE_PTR(global_object->varlink);
+ * varlink_close_unref(t);
+ */
+ /* n_ref >= 1 */
+ varlink_ref(v); /* n_ref >= 2 */
+ varlink_close(v); /* n_ref >= 1 */
+ if (v->n_ref > 1)
+ v->n_ref--; /* n_ref >= 1 */
return varlink_unref(v);
}
Varlink* varlink_flush_close_unref(Varlink *v) {
- if (!v)
- return NULL;
+ if (v)
+ varlink_flush(v);
- (void) varlink_flush(v);
- (void) varlink_close(v);
- return varlink_unref(v);
+ return varlink_close_unref(v);
}
static int varlink_enqueue_json(Varlink *v, JsonVariant *m) {
From a19c1a4baaa1dadc80885e3ad41f19a6c6c450fd Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Mon, 8 Mar 2021 09:21:25 +0100
Subject: [PATCH 3/3] oomd: "downgrade" level of message
PID1 already logs about the service being started, so this line isn't necessary
in normal use. Also, by the time it is emitted, the service has already
signalled readiness, so let's not say "starting" but "started".
---
src/oom/oomd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/oom/oomd.c b/src/oom/oomd.c
index 674d53fdcfe..6e2a5889d1e 100644
--- a/src/oom/oomd.c
+++ b/src/oom/oomd.c
@@ -170,7 +170,7 @@ static int run(int argc, char *argv[]) {
notify_msg = notify_start(NOTIFY_READY, NOTIFY_STOPPING);
- log_info("systemd-oomd starting%s!", arg_dry_run ? " in dry run mode" : "");
+ log_debug("systemd-oomd started%s.", arg_dry_run ? " in dry run mode" : "");
r = sd_event_loop(m->event);
if (r < 0)
Reference in New Issue View Git Blame Copy Permalink