qemu/plugins-make-qemu_plugin_user_exit-s-locking-order-c.patch
Fei Xu f9d45693eb QEMU update to version 6.2.0-71(master)
-vhost-user-blk: fix the resize crash
 -plugins: make qemu_plugin_user_exit's locking order consistent with fork_start's
 -linux-user: fix strace build w/out munlockall
 -ui: fix crash on serial reset, during init
 -qga/win/vss: requester_freeze changes
 -migration: fix populate_vfio_info
 -block/rbd: workaround for ceph issue #53784
 -target/i386: add FZRM, FSRS, FSRC
 -i386: Add new CPU model SapphireRapids
 -core/cpu-common: Fix the wrong '#ifdef __aarch64__'

Signed-off-by: Fei Xu <xufei30@huawei.com>
2023-04-22 21:43:51 +08:00

67 lines
2.1 KiB
Diff

From a729d2730d9d30f6610e43f070cedd1d60ba022f Mon Sep 17 00:00:00 2001
From: qihao <qihao@cmss.chinamobile.com>
Date: Thu, 30 Mar 2023 17:58:32 +0800
Subject: [PATCH] plugins: make qemu_plugin_user_exit's locking order
consistent with fork_start's
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
cheery-pick from 2bbbc1be8d9a21b25d0c80b9a7345074d54abd51
To fix potential deadlocks as reported by tsan.
Signed-off-by: qihao_yewu <qihao_yewu@cmss.chinamobile.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Emilio Cota <cota@braap.org>
Message-Id: <20230111151628.320011-6-cota@braap.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230124180127.1881110-31-alex.bennee@linaro.org>
---
plugins/core.c | 14 +++++++++++---
1 file changed, 11 insertions(+), 3 deletions(-)
diff --git a/plugins/core.c b/plugins/core.c
index 792262da08..e935e3c0c9 100644
--- a/plugins/core.c
+++ b/plugins/core.c
@@ -500,10 +500,18 @@ void qemu_plugin_user_exit(void)
enum qemu_plugin_event ev;
CPUState *cpu;
- QEMU_LOCK_GUARD(&plugin.lock);
+ /*
+ * Locking order: we must acquire locks in an order that is consistent
+ * with the one in fork_start(). That is:
+ * - start_exclusive(), which acquires qemu_cpu_list_lock,
+ * must be called before acquiring plugin.lock.
+ * - tb_flush(), which acquires mmap_lock(), must be called
+ * while plugin.lock is not held.
+ */
start_exclusive();
+ qemu_rec_mutex_lock(&plugin.lock);
/* un-register all callbacks except the final AT_EXIT one */
for (ev = 0; ev < QEMU_PLUGIN_EV_MAX; ev++) {
if (ev != QEMU_PLUGIN_EV_ATEXIT) {
@@ -514,12 +522,12 @@ void qemu_plugin_user_exit(void)
}
}
- tb_flush(current_cpu);
-
CPU_FOREACH(cpu) {
qemu_plugin_disable_mem_helpers(cpu);
}
+ qemu_rec_mutex_unlock(&plugin.lock);
+ tb_flush(current_cpu);
end_exclusive();
/* now it's safe to handle the exit case */
--
2.27.0