[PATCH] possible use-after-free of bio

There is a possibility that a bio will be accessed after it has been freed
on SCSI.  It happens if you submit a bio with BIO_SYNC marked and the
auto-unplugging kicks the request_fn, SCSI re-enables interrupts in-between
so if the request completes between the add_request() in __make_request()
and the bio_sync() call, we could be looking at a dead bio.  It's a slim
race, but it has been triggered in the Real World.

So assign bio_sync() to a local variable instead.

Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/drivers/block/ll_rw_blk.c b/drivers/block/ll_rw_blk.c
index 02242e8ba996872e0aa78aa78af662cdf83c732c..2d6934a028674fb0df6ed187643686159cdacc84 100644 (file)
--- a/drivers/block/ll_rw_blk.c
+++ b/drivers/block/ll_rw_blk.c
@@ -2559,7 +2559,7 @@ EXPORT_SYMBOL(__blk_attempt_remerge);
 static int __make_request(request_queue_t *q, struct bio *bio)
 {
        struct request *req, *freereq = NULL;
-       int el_ret, rw, nr_sectors, cur_nr_sectors, barrier, err;
+       int el_ret, rw, nr_sectors, cur_nr_sectors, barrier, err, sync;
        sector_t sector;
 
        sector = bio->bi_sector;
@@ -2567,6 +2567,7 @@ static int __make_request(request_queue_t *q, struct bio *bio)
        cur_nr_sectors = bio_cur_sectors(bio);
 
        rw = bio_data_dir(bio);
+       sync = bio_sync(bio);
 
        /*
         * low level driver can indicate that it wants pages above a
@@ -2698,7 +2699,7 @@ get_rq:
 out:
        if (freereq)
                __blk_put_request(q, freereq);
-       if (bio_sync(bio))
+       if (sync)
                __generic_unplug_device(q);
 
        spin_unlock_irq(q->queue_lock);