Message ID | 20130606141941.GA5382@thunk.org |
---|---|
State | Accepted, archived |
Headers | show |
Hi Ted, On Thu, 6 Jun 2013 10:19:41 -0400 Theodore Ts'o <tytso@mit.edu> wrote: > > On Thu, Jun 06, 2013 at 11:56:53AM +1000, Stephen Rothwell wrote: > > > > In tracking this down, I followed the call chains and discovered that > > io_submit_init_bio() only ever returned 0. Switching that to return > > void and following back up the chain lead to the following patch. > > Actually, that's a bug. We're missing an check in that function. > Thanks for looking at this code. (I'll fix the uninitialized err > warning another way.) Glad to be of service.
diff --git a/fs/ext4/page-io.c b/fs/ext4/page-io.c index ce8c15a..48786cd 100644 --- a/fs/ext4/page-io.c +++ b/fs/ext4/page-io.c @@ -372,6 +372,8 @@ static int io_submit_init_bio(struct ext4_io_submit *io, struct bio *bio; bio = bio_alloc(GFP_NOIO, min(nvecs, BIO_MAX_PAGES)); + if (!bio) + return -ENOMEM; bio->bi_sector = bh->b_blocknr * (bh->b_size >> 9); bio->bi_bdev = bh->b_bdev; bio->bi_end_io = ext4_end_bio;