2.6.32-longterm review patch. If anyone has any objections, please let me know.
------------------
From: Catalin Marinas <catalin.marinas@xxxxxxx>
commit c5f3b1a51a591c18c8b33983908e7fdda6ae417e upstream.
The kmemleak scanning thread can run for minutes. Callbacks like
kmemleak_free() are allowed during this time, the race being taken care
of by the object->lock spinlock. Such lock also prevents a memory block
from being freed or unmapped while it is being scanned by blocking the
kmemleak_free() -> ... -> __delete_object() function until the lock is
released in scan_object().
When a kmemleak error occurs (e.g. it fails to allocate its metadata),
kmemleak_enabled is set and __delete_object() is no longer called on
freed objects. If kmemleak_scan is running at the same time,
kmemleak_free() no longer waits for the object scanning to complete,
allowing the corresponding memory block to be freed or unmapped (in the
case of vfree()). This leads to kmemleak_scan potentially triggering a
page fault.
This patch separates the kmemleak_free() enabling/disabling from the
overall kmemleak_enabled nob so that we can defer the disabling of the
object freeing tracking until the scanning thread completed. The
kmemleak_free_part() is deliberately ignored by this patch since this is
only called during boot before the scanning thread started.
Signed-off-by: Catalin Marinas <catalin.marinas@xxxxxxx>
Reported-by: Vignesh Radhakrishnan <vigneshr@xxxxxxxxxxxxxx>
Tested-by: Vignesh Radhakrishnan <vigneshr@xxxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
[bwh: Backported to 3.2:
- Adjust context
- Drop changes to kmemleak_free_percpu()]
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
(cherry picked from commit 3bc68ffc5b43468537a2f0aa415f3b57f3b19d16)
Signed-off-by: Willy Tarreau <w@xxxxxx>
---
mm/kmemleak.c | 14 +++++++++++++-
1 file changed, 13 insertions(+), 1 deletion(-)
diff --git a/mm/kmemleak.c b/mm/kmemleak.c
index c346660..e9bd6d5 100644
--- a/mm/kmemleak.c
+++ b/mm/kmemleak.c
@@ -191,6 +191,8 @@ static struct kmem_cache *scan_area_cache;
/* set if tracing memory operations is enabled */
static atomic_t kmemleak_enabled = ATOMIC_INIT(0);
+/* same as above but only for the kmemleak_free() callback */
+static int kmemleak_free_enabled;
/* set in the late_initcall if there were no errors */
static atomic_t kmemleak_initialized = ATOMIC_INIT(0);
/* enables or disables early logging of the memory operations */
@@ -870,7 +872,7 @@ void __ref kmemleak_free(const void *ptr)
{
pr_debug("%s(0x%p)\n", __func__, ptr);
- if (atomic_read(&kmemleak_enabled) && ptr && !IS_ERR(ptr))
+ if (kmemleak_free_enabled && ptr && !IS_ERR(ptr))
delete_object_full((unsigned long)ptr);
else if (atomic_read(&kmemleak_early_log))
log_early(KMEMLEAK_FREE, ptr, 0, 0, 0, 0);
@@ -1552,6 +1554,13 @@ static void kmemleak_do_cleanup(struct work_struct *work)
mutex_lock(&scan_mutex);
stop_scan_thread();
+ /*
+ * Once the scan thread has stopped, it is safe to no longer track
+ * object freeing. Ordering of the scan thread stopping and the memory
+ * accesses below is guaranteed by the kthread_stop() function.
+ */
+ kmemleak_free_enabled = 0;
+
rcu_read_lock();
list_for_each_entry_rcu(object, &object_list, object_list)
delete_object_full(object->pointer);
@@ -1578,6 +1587,8 @@ static void kmemleak_disable(void)
/* check whether it is too early for a kernel thread */
if (atomic_read(&kmemleak_initialized))
schedule_work(&cleanup_work);
+ else
+ kmemleak_free_enabled = 0;
pr_info("Kernel memory leak detector disabled\n");
}
@@ -1617,6 +1628,7 @@ void __init kmemleak_init(void)
if (!atomic_read(&kmemleak_error)) {
atomic_set(&kmemleak_enabled, 1);
atomic_set(&kmemleak_early_log, 0);
+ kmemleak_free_enabled = 1;
}
local_irq_restore(flags);
--
1.7.12.2.21.g234cd45.dirty
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html