Messages in this thread | | | Date | Mon, 7 Jan 2013 14:15:21 -0800 | From | Kent Overstreet <> | Subject | Re: [PATCH 04/32] aio: remove retry-based AIO |
| |
On Sat, Dec 29, 2012 at 03:47:37PM +0800, Hillf Danton wrote: > On Thu, Dec 27, 2012 at 9:59 AM, Kent Overstreet <koverstreet@google.com> wrote: > > @@ -1585,18 +1278,27 @@ static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb, > > * don't see ctx->dead set here, io_destroy() waits for our IO to > > * finish. > > */ > > - if (ctx->dead) { > > - spin_unlock_irq(&ctx->ctx_lock); > > + if (ctx->dead) > > ret = -EINVAL; > > + spin_unlock_irq(&ctx->ctx_lock); > > + if (ret) > > goto out_put_req; > > + > > + if (unlikely(kiocbIsCancelled(req))) { > > + ret = -EINTR; > > + } else { > > + ret = req->ki_retry(req); > > } > > - aio_run_iocb(req); > > - if (!list_empty(&ctx->run_list)) { > > - /* drain the run list */ > > - while (__aio_run_iocbs(ctx)) > > - ; > > + if (ret != -EIOCBQUEUED) { > > + /* > > + * There's no easy way to restart the syscall since other AIO's > > + * may be already running. Just fail this IO with EINTR. > > + */ > > + if (unlikely(ret == -ERESTARTSYS || ret == -ERESTARTNOINTR || > > + ret == -ERESTARTNOHAND || ret == -ERESTART_RESTARTBLOCK)) > > + ret = -EINTR; > > + aio_complete(req, ret, 0); > > } > > - spin_unlock_irq(&ctx->ctx_lock); > > > > aio_put_req(req); /* drop extra ref to req */ > > return 0; > > return ret; yes?
No - and this code is _really_ confusing. If we get an error we can return it in two different ways:
1) By making the io_submit() call fail 2) Through the io_event that aio_complete() delivers
Note the return 0 coming after the aio_complete() call - we don't want to return ret because then we'd be delivering an error twice.
| |