Xen project Mailing List

2014-03-11 19:10 GMT+08:00 Ian Campbell <Ian.Campbell@xxxxxxxxxx>:

On Fri, 2014-02-28 at 22:50 +0800, gordongong0350@xxxxxxxxx wrote:
> Fix vreq with error of -EBUSY fails 100 times too soon, that return

Sorry, but your attempts to send this patch are all whitespace mangled
and are lacking a signed-off-by. Please see
http://wiki.xen.org/wiki/Submitting_Xen_Patches

Ian.

> EIO to td device layer. If this td
> device request is metadata of filesystem, the result is not good at
> all.
>
>
> To reproduce it witch "./iozone -I -s 2G -r 512k -r 1m -r 2m -r 4m -i
> 0 -i 1 -f /data/iozone_test.tmp",
> the result is input/output error & iozone is stopped.
>
>
>
>
>
>
>
>
> From ccc0ce3aede1f5c920036847ddb16b25969be7ba Mon Sep 17 00:00:00 2001
> From: Xiaodong Gong <gordongong0350@xxxxxxxxx>
> Date: Fri, 28 Feb 2014 03:01:19 -0500
> Subject: [PATCH] fix vreq with error of -EBUSY fails 100 times too
> soon,
> return EIO to td device layer.
>
>
> 1) all free list is consumed by vreq in one or two block that
> is seted in bat entry.
> 2) vreq in fail queue, get a chance to run quickly, such as 90us.
> ---
> tools/blktap2/drivers/tapdisk-vbd.c | 17 ++++++++++++++---
> tools/blktap2/drivers/tapdisk-vbd.h | 1 +
> 2 files changed, 15 insertions(+), 3 deletions(-)
>
>
> diff --git a/tools/blktap2/drivers/tapdisk-vbd.c
> b/tools/blktap2/drivers/tapdisk-vbd.c
> index c665f27..529eb91 100644
> --- a/tools/blktap2/drivers/tapdisk-vbd.c
> +++ b/tools/blktap2/drivers/tapdisk-vbd.c
> @@ -1081,7 +1081,7 @@ tapdisk_vbd_check_state(td_vbd_t *vbd)
> td_vbd_request_t *vreq, *tmp;
>
> tapdisk_vbd_for_each_request(vreq, tmp, &vbd->failed_requests)
> - if (vreq->num_retries >= TD_VBD_MAX_RETRIES)
> + if (vreq->num_retries >= TD_VBD_MAX_RETRIES && vreq->busy_looping !=
> 1 )
> tapdisk_vbd_complete_vbd_request(vbd, vreq);
>
> if (!list_empty(&vbd->new_requests) ||
> @@ -1168,7 +1168,7 @@ tapdisk_vbd_complete_vbd_request(td_vbd_t *vbd,
> td_vbd_request_t *vreq)
> {
> if (!vreq->submitting && !vreq->secs_pending) {
> if (vreq->status == BLKIF_RSP_ERROR &&
> - vreq->num_retries < TD_VBD_MAX_RETRIES &&
> + (vreq->num_retries < TD_VBD_MAX_RETRIES || vreq->busy_looping ==
> 1)&&
> !td_flag_test(vbd->state, TD_VBD_DEAD) &&
> !td_flag_test(vbd->state, TD_VBD_SHUTDOWN_REQUESTED))
> tapdisk_vbd_move_request(vreq, &vbd->failed_requests);
> @@ -1450,17 +1450,28 @@ tapdisk_vbd_reissue_failed_requests(td_vbd_t
> *vbd)
> gettimeofday(&now, NULL);
>
> tapdisk_vbd_for_each_request(vreq, tmp, &vbd->failed_requests) {
> + uint64_t delta = 0;
> +
> if (vreq->secs_pending)
> continue;
>
> if (td_flag_test(vbd->state, TD_VBD_SHUTDOWN_REQUESTED))
> goto fail;
>
> + if (vreq->num_retries > TD_VBD_MAX_RETRIES - 10) {
> + delta = (now.tv_sec - vreq->last_try.tv_sec) * 1000000\
> + + now.tv_usec - vreq->last_try.tv_usec;
> + if (delta * vreq->num_retries < TD_VBD_RETRY_INTERVAL)
> + vreq->busy_looping = 1;
> + else
> + vreq->busy_looping = 0;
> + }
> +
> if (vreq->error != -EBUSY &&
> now.tv_sec - vreq->last_try.tv_sec < TD_VBD_RETRY_INTERVAL)
> continue;
>
> - if (vreq->num_retries >= TD_VBD_MAX_RETRIES) {
> + if (vreq->num_retries >= TD_VBD_MAX_RETRIES && vreq->busy_looping !=
> 1) {
> fail:
> DBG(TLOG_INFO, "req %"PRIu64"retried %d times\n",
> vreq->req.id, vreq->num_retries);
> diff --git a/tools/blktap2/drivers/tapdisk-vbd.h
> b/tools/blktap2/drivers/tapdisk-vbd.h
> index be084b2..9e5f5f6 100644
> --- a/tools/blktap2/drivers/tapdisk-vbd.h
> +++ b/tools/blktap2/drivers/tapdisk-vbd.h
> @@ -73,6 +73,7 @@ struct td_vbd_request {
> int submitting;
> int secs_pending;
> int num_retries;
> + int busy_looping;
> struct timeval last_try;
>
> td_vbd_t *vbd;
>
>

> --
> 1.8.3.1
>
>
>
> ______________________________________________________________________
> gordongong0350@xxxxxxxxx

Re: [Xen-devel] [tapdisk2][PATCH1/1] Fix vreq with error of -EBUSY fails 100 times too soon