[Qemu-devel] [PATCH] gluster: Abort on AIO completion failure

[Qemu-devel] [PATCH] gluster: Abort on AIO completion failure

[Bharata B Rao]

Currently if gluster AIO callback thread fails to notify the QEMU thread about
AIO completion, we try graceful recovery by marking the disk drive as
inaccessible. This error recovery code is race-prone as found by Asias and
Stefan. However as found out by Paolo, this kind of error is impossible and
hence simplify the code that handles this error recovery.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
---
 block/gluster.c | 15 ++-------------
 1 file changed, 2 insertions(+), 13 deletions(-)

diff --git a/block/gluster.c b/block/gluster.c
index 46f36f8..942154a 100644
--- a/block/gluster.c
+++ b/block/gluster.c
@@ -427,20 +427,9 @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret, void *arg)
         /*
          * Gluster AIO callback thread failed to notify the waiting
          * QEMU thread about IO completion.
-         *
-         * Complete this IO request and make the disk inaccessible for
-         * subsequent reads and writes.
          */
-        error_report("Gluster failed to notify QEMU about IO completion");
-
-        qemu_mutex_lock_iothread(); /* We are in gluster thread context */
-        acb->common.cb(acb->common.opaque, -EIO);
-        qemu_aio_release(acb);
-        close(s->fds[GLUSTER_FD_READ]);
-        close(s->fds[GLUSTER_FD_WRITE]);
-        qemu_aio_set_fd_handler(s->fds[GLUSTER_FD_READ], NULL, NULL, NULL);
-        bs->drv = NULL; /* Make the disk inaccessible */
-        qemu_mutex_unlock_iothread();
+        error_report("Gluster AIO completion failed");
+        abort();
     }
 }
 
-- 
1.7.11.7

Re: [Qemu-devel] [PATCH] gluster: Abort on AIO completion failure

[Paolo Bonzini]

Il 27/08/2013 07:53, Bharata B Rao ha scritto:
> Currently if gluster AIO callback thread fails to notify the QEMU thread about
> AIO completion, we try graceful recovery by marking the disk drive as
> inaccessible. This error recovery code is race-prone as found by Asias and
> Stefan. However as found out by Paolo, this kind of error is impossible and
> hence simplify the code that handles this error recovery.
> 
> Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
> ---
>  block/gluster.c | 15 ++-------------
>  1 file changed, 2 insertions(+), 13 deletions(-)
> 
> diff --git a/block/gluster.c b/block/gluster.c
> index 46f36f8..942154a 100644
> --- a/block/gluster.c
> +++ b/block/gluster.c
> @@ -427,20 +427,9 @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret, void *arg)
>          /*
>           * Gluster AIO callback thread failed to notify the waiting
>           * QEMU thread about IO completion.
> -         *
> -         * Complete this IO request and make the disk inaccessible for
> -         * subsequent reads and writes.
>           */
> -        error_report("Gluster failed to notify QEMU about IO completion");
> -
> -        qemu_mutex_lock_iothread(); /* We are in gluster thread context */
> -        acb->common.cb(acb->common.opaque, -EIO);
> -        qemu_aio_release(acb);
> -        close(s->fds[GLUSTER_FD_READ]);
> -        close(s->fds[GLUSTER_FD_WRITE]);
> -        qemu_aio_set_fd_handler(s->fds[GLUSTER_FD_READ], NULL, NULL, NULL);
> -        bs->drv = NULL; /* Make the disk inaccessible */
> -        qemu_mutex_unlock_iothread();
> +        error_report("Gluster AIO completion failed");

Still, considering everyone can be wrong, adding the errno here is
probably not a bad idea. :)

Paolo

> +        abort();
>      }
>  }
>  
> 

Re: [Qemu-devel] [PATCH] gluster: Abort on AIO completion failure

[Bharata B Rao]

On Tue, Aug 27, 2013 at 08:58:05AM +0200, Paolo Bonzini wrote:
> > diff --git a/block/gluster.c b/block/gluster.c
> > index 46f36f8..942154a 100644
> > --- a/block/gluster.c
> > +++ b/block/gluster.c
> > @@ -427,20 +427,9 @@ static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret, void *arg)
> >          /*
> >           * Gluster AIO callback thread failed to notify the waiting
> >           * QEMU thread about IO completion.
> > -         *
> > -         * Complete this IO request and make the disk inaccessible for
> > -         * subsequent reads and writes.
> >           */
> > -        error_report("Gluster failed to notify QEMU about IO completion");
> > -
> > -        qemu_mutex_lock_iothread(); /* We are in gluster thread context */
> > -        acb->common.cb(acb->common.opaque, -EIO);
> > -        qemu_aio_release(acb);
> > -        close(s->fds[GLUSTER_FD_READ]);
> > -        close(s->fds[GLUSTER_FD_WRITE]);
> > -        qemu_aio_set_fd_handler(s->fds[GLUSTER_FD_READ], NULL, NULL, NULL);
> > -        bs->drv = NULL; /* Make the disk inaccessible */
> > -        qemu_mutex_unlock_iothread();
> > +        error_report("Gluster AIO completion failed");
> 
> Still, considering everyone can be wrong, adding the errno here is
> probably not a bad idea. :)

Sent v1 with errno added.

Regards,
Bharata.