| Message ID | d0186f22982fb49e88bd5ee2273ba90946318ba5.1499711846.git.shli@fb.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, Jul 10, 2017 at 11:40:17AM -0700, Shaohua Li wrote: > bio_free isn't a good place to free cgroup info. There are a > lot of cases bio is allocated in special way (for example, in stack) and > never gets called by bio_put hence bio_free, we are leaking memory. This > patch moves the free to bio endio, which should be called anyway. The > bio_uninit call in bio_free is kept, in case the bio never gets called > bio endio. > > This assumes ->bi_end_io() doesn't access cgroup info, which seems true > in my audit. > > This along with Christoph's integrity patch should fix the memory leak > issue. Can we please go ahead and: (a) remove the now unneeded call in __blkdev_direct_IO_simple (b) kill bio_uninit - it's just a wrapper around bio_disassociate_task (c) check if we can remove the call in bio_free, which I think we can as bi_ioc and bi_css are only assigned from generic_make_request
On Wed, Jul 12, 2017 at 09:25:17AM +0200, Christoph Hellwig wrote: > On Mon, Jul 10, 2017 at 11:40:17AM -0700, Shaohua Li wrote: > > bio_free isn't a good place to free cgroup info. There are a > > lot of cases bio is allocated in special way (for example, in stack) and > > never gets called by bio_put hence bio_free, we are leaking memory. This > > patch moves the free to bio endio, which should be called anyway. The > > bio_uninit call in bio_free is kept, in case the bio never gets called > > bio endio. > > > > This assumes ->bi_end_io() doesn't access cgroup info, which seems true > > in my audit. > > > > This along with Christoph's integrity patch should fix the memory leak > > issue. > > Can we please go ahead and: > > (a) remove the now unneeded call in __blkdev_direct_IO_simple > (b) kill bio_uninit - it's just a wrapper around bio_disassociate_task > (c) check if we can remove the call in bio_free, which I think we can > as bi_ioc and bi_css are only assigned from generic_make_request Sure, will do these soon.
diff --git a/block/bio.c b/block/bio.c index 9cabf5d..9a63597 100644 --- a/block/bio.c +++ b/block/bio.c @@ -1833,6 +1833,8 @@ void bio_endio(struct bio *bio) } blk_throtl_bio_endio(bio); + /* release cgroup info */ + bio_uninit(bio); if (bio->bi_end_io) bio->bi_end_io(bio); }
bio_free isn't a good place to free cgroup info. There are a lot of cases bio is allocated in special way (for example, in stack) and never gets called by bio_put hence bio_free, we are leaking memory. This patch moves the free to bio endio, which should be called anyway. The bio_uninit call in bio_free is kept, in case the bio never gets called bio endio. This assumes ->bi_end_io() doesn't access cgroup info, which seems true in my audit. This along with Christoph's integrity patch should fix the memory leak issue. Cc: Christoph Hellwig <hch@lst.de> Signed-off-by: Shaohua Li <shli@fb.com> --- block/bio.c | 2 ++ 1 file changed, 2 insertions(+)