mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] firmware loader: Fix the concurrent request_firmware() race for kref_get/put
@ 2012-11-09 12:28 Chuansheng Liu
  2012-11-09  8:16 ` Ming Lei
  2012-11-09 17:27 ` [PATCH V2] " Chuansheng Liu
  0 siblings, 2 replies; 4+ messages in thread
From: Chuansheng Liu @ 2012-11-09 12:28 UTC (permalink / raw)
  To: ming.lei, gregkh; +Cc: linux-kernel, chuansheng.liu


There is one race that both request_firmware() with the same
firmware name.

The race scenerio is as below:
CPU1                                                  CPU2
request_firmware() -->
_request_firmware_load() return err                   another request_firmware() is coming -->
_request_firmware_cleanup is called -->               _request_firmware_prepare -->
release_firmware --->                                 fw_lookup_and_allocate_buf -->
                                                      spin_lock(&fwc->lock)
...                                                   __fw_lookup_buf() return true
fw_free_buf() will be called -->                      ...
kref_put -->
decrease the refcount to 0
                                                      kref_get(&tmp->ref) ==> it will trigger warning
                                                                              due to refcount == 0
__fw_free_buf() -->
...                                                   spin_unlock(&fwc->lock)
spin_lock(&fwc->lock)
list_del(&buf->list)
spin_unlock(&fwc->lock)
kfree(buf)
                                                      After that, the freed buf will be used.

The key race is decreasing refcount to 0 and list_del is not protected together by
fwc->lock, and it is possible another thread try to get it between refcount==0
and list_del.

Fix it here to protect it together.

Signed-off-by: liu chuansheng <chuansheng.liu@intel.com>
---
 drivers/base/firmware_class.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c
index b44ed35..7df32cd 100644
--- a/drivers/base/firmware_class.c
+++ b/drivers/base/firmware_class.c
@@ -246,7 +246,6 @@ static void __fw_free_buf(struct kref *ref)
 		 __func__, buf->fw_id, buf, buf->data,
 		 (unsigned int)buf->size);
 
-	spin_lock(&fwc->lock);
 	list_del(&buf->list);
 	spin_unlock(&fwc->lock);
 
@@ -263,7 +262,10 @@ static void __fw_free_buf(struct kref *ref)
 
 static void fw_free_buf(struct firmware_buf *buf)
 {
-	kref_put(&buf->ref, __fw_free_buf);
+	struct firmware_cache *fwc = buf->fwc;
+	spin_lock(&fwc->lock);
+	if(!kref_put(&buf->ref, __fw_free_buf))
+		spin_unlock(&fwc->lock);
 }
 
 /* direct firmware loading support */
-- 
1.7.0.4




^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2012-11-09  8:26 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-11-09 12:28 [PATCH] firmware loader: Fix the concurrent request_firmware() race for kref_get/put Chuansheng Liu
2012-11-09  8:16 ` Ming Lei
2012-11-09  8:21   ` Liu, Chuansheng
2012-11-09 17:27 ` [PATCH V2] " Chuansheng Liu

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®