[Qemu-devel] [PATCH v5 2/4] timer: protect timers_state's clock with seqlock

[Liu Ping Fan <qemulist@gmail.com>]

QEMU_CLOCK_VIRTUAL may be read outside BQL. This will make its
foundation, i.e. cpu_clock_offset exposed to race condition.
Using private lock to protect it.

After this patch, reading QEMU_CLOCK_VIRTUAL is thread safe
unless use_icount is true, in which case the existing callers
still rely on the BQL

Lock rule: private lock innermost, ie BQL->"this lock"

Signed-off-by: Liu Ping Fan <pingfank@linux.vnet.ibm.com>
---
 cpus.c               | 41 +++++++++++++++++++++++++++++++++--------
 include/qemu/timer.h |  2 ++
 2 files changed, 35 insertions(+), 8 deletions(-)

diff --git a/cpus.c b/cpus.c
index e566297..5baa76d 100644
--- a/cpus.c
+++ b/cpus.c
@@ -37,6 +37,7 @@
 #include "sysemu/qtest.h"
 #include "qemu/main-loop.h"
 #include "qemu/bitmap.h"
+#include "qemu/seqlock.h"
 
 #ifndef _WIN32
 #include "qemu/compatfd.h"
@@ -112,6 +113,11 @@ static int64_t qemu_icount;
 typedef struct TimersState {
     int64_t cpu_ticks_prev;
     int64_t cpu_ticks_offset;
+    /* cpu_clock_offset will be read out of BQL, so protect it with private
+     * lock. As for cpu_ticks_*, no requirement to read it outside BQL yet.
+     * Lock rule: innermost
+     */
+    QemuSeqLock cpu_clock_offset_seqlock;
     int64_t cpu_clock_offset;
     int32_t cpu_ticks_enabled;
     int64_t dummy;
@@ -137,6 +143,7 @@ int64_t cpu_get_icount(void)
 }
 
 /* return the host CPU cycle counter and handle stop/restart */
+/* Caller must hold the BQL */
 int64_t cpu_get_ticks(void)
 {
     if (use_icount) {
@@ -161,33 +168,50 @@ int64_t cpu_get_ticks(void)
 int64_t cpu_get_clock(void)
 {
     int64_t ti;
-    if (!timers_state.cpu_ticks_enabled) {
-        return timers_state.cpu_clock_offset;
-    } else {
-        ti = get_clock();
-        return ti + timers_state.cpu_clock_offset;
-    }
+    unsigned start;
+
+    do {
+        start = seqlock_read_begin(&timers_state.cpu_clock_offset_seqlock);
+        if (!timers_state.cpu_ticks_enabled) {
+            ti = timers_state.cpu_clock_offset;
+        } else {
+            ti = get_clock();
+            ti += timers_state.cpu_clock_offset;
+        }
+    } while (seqlock_read_retry(&timers_state.cpu_clock_offset_seqlock, start));
+
+    return ti;
 }
 
-/* enable cpu_get_ticks() */
+/* enable cpu_get_ticks()
+ * Caller must hold BQL which server as mutex for cpu_clock_offset_seqlock.
+ */
 void cpu_enable_ticks(void)
 {
+    /* Here, the really thing protected by seqlock is cpu_clock_offset. */
+    seqlock_write_lock(&timers_state.cpu_clock_offset_seqlock);
     if (!timers_state.cpu_ticks_enabled) {
         timers_state.cpu_ticks_offset -= cpu_get_real_ticks();
         timers_state.cpu_clock_offset -= get_clock();
         timers_state.cpu_ticks_enabled = 1;
     }
+    seqlock_write_unlock(&timers_state.cpu_clock_offset_seqlock);
 }
 
 /* disable cpu_get_ticks() : the clock is stopped. You must not call
-   cpu_get_ticks() after that.  */
+ * cpu_get_ticks() after that.
+ * Caller must hold BQL which server as mutex for cpu_clock_offset_seqlock.
+ */
 void cpu_disable_ticks(void)
 {
+    /* Here, the really thing protected by seqlock is cpu_clock_offset. */
+    seqlock_write_lock(&timers_state.cpu_clock_offset_seqlock);
     if (timers_state.cpu_ticks_enabled) {
         timers_state.cpu_ticks_offset = cpu_get_ticks();
         timers_state.cpu_clock_offset = cpu_get_clock();
         timers_state.cpu_ticks_enabled = 0;
     }
+    seqlock_write_unlock(&timers_state.cpu_clock_offset_seqlock);
 }
 
 /* Correlation between real and virtual time is always going to be
@@ -371,6 +395,7 @@ static const VMStateDescription vmstate_timers = {
 
 void configure_icount(const char *option)
 {
+    seqlock_init(&timers_state.cpu_clock_offset_seqlock, NULL);
     vmstate_register(NULL, 0, &vmstate_timers, &timers_state);
     if (!option) {
         return;
diff --git a/include/qemu/timer.h b/include/qemu/timer.h
index e4934dd..bb1de23 100644
--- a/include/qemu/timer.h
+++ b/include/qemu/timer.h
@@ -636,7 +636,9 @@ static inline int64_t qemu_soonest_timeout(int64_t timeout1, int64_t timeout2)
 void init_clocks(void);
 
 int64_t cpu_get_ticks(void);
+/* Caller must hold BQL */
 void cpu_enable_ticks(void);
+/* Caller must hold BQL */
 void cpu_disable_ticks(void);
 
 static inline int64_t get_ticks_per_sec(void)
-- 
1.8.1.4