On Mon, Mar 10, 2014 at 04:15:33PM +0800, Tang Chen wrote: > IO ring page migration has been implemented by the following patch: > > https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/fs/aio.c?id=36bc08cc01709b4a9bb563b35aa530241ddc63e3 > > In this patch, ctx->completion_lock is used to prevent other processes > from accessing the ring page being migrated. > > But in aio_setup_ring(), ioctx_add_table() and aio_read_events_ring(), > when writing to the ring page, they didn't take ctx->completion_lock. > As a result, for example, we have the following problem: ... > As above, the new ring page will not be updated. > > The solution is taking ctx->completion_lock in thread 2, which means, > in aio_setup_ring(), ioctx_add_table() and aio_read_events_ring() when > writing to ring pages. Upon review, there are still two accesses of ->ring_pages that are not protected by any spinlocks which could potentially race with migration. One is in aio_setup_ring(), which can be easily resolved by moving the assignment of ->ring_pages above the unlock_page(). Another spot is in aio_read_events_ring() where head and tail are fetched from the ring without any locking. I also fear we'll be introducing new performance issues with all the additonal spinlock bouncing, despite the fact that is only ever needed for migration. I'm going to continue looking into this today and will try to send out a followup to this email later. -ben -- "Thought is the essence of where you are now." -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
