From: Olivier Bonvalet <ceph.list@daevel.fr>
To: Alex Elder <elder@ieee.org>
Cc: Ilya Dryomov <ilya.dryomov@inktank.com>,
Ceph Development <ceph-devel@vger.kernel.org>
Subject: Re: Issue #5876 : assertion failure in rbd_img_obj_callback()
Date: Wed, 26 Mar 2014 05:00:47 +0100 [thread overview]
Message-ID: <1395806447.2076.70.camel@localhost> (raw)
In-Reply-To: <53324F79.1080108@ieee.org>
Le mardi 25 mars 2014 à 22:54 -0500, Alex Elder a écrit :
> On 03/25/2014 09:45 PM, Olivier Bonvalet wrote:
> > Le mardi 25 mars 2014 à 21:42 -0500, Alex Elder a écrit :
> >> PS I thought you said you were going to stop for the night!
> >
> > Yes, I would love ! But my phone doesn't stop ring about ceph crash :D
> >
>
> OK, one more thing to try and I'm going to bed.
>
> I'm hoping that an image request spanning multiple objects
> is an unusual case, enough so that the following won't
> overwhelm with output. I'd avoid putting it on a production
> system (that's the case for all this testing, really) if
> possible.
>
> Basically I'm trying to catch an image object request being
> either submitted more than once, or completing more than
> once. So if an image request has more than one object
> request I produce some informative output.
>
> This patch fixes two warnings in the previous debug patch,
> and adds to it (so use it instead of the last one).
>
> If you get a chance to try this I'll want to see the output.
> But first, I shall sleep.
>
> Thank you.
>
> -Alex
>
> Index: b/drivers/block/rbd.c
> ===================================================================
> --- a/drivers/block/rbd.c
> +++ b/drivers/block/rbd.c
> @@ -1484,6 +1484,10 @@ static void rbd_img_request_complete(str
> xferred += obj_request->xferred;
> img_request->xferred = xferred;
> }
> + if (img_request->obj_request_count > 1)
> + printk("%s: img_request %p count %u result %d xferred %llu\n",
> + __func__, img_request, img_request->obj_request_count,
> + img_request->result, img_request->xferred);
>
> if (img_request->callback)
> img_request->callback(img_request);
> @@ -2129,9 +2133,43 @@ static void rbd_img_obj_callback(struct
> rbd_assert(which != BAD_WHICH);
> rbd_assert(which < img_request->obj_request_count);
>
> + if (img_request->obj_request_count > 1)
> + printk("%s: obj_request %p (%llu/%llu)\n", __func__,
> + obj_request, obj_request->offset, obj_request->length);
> spin_lock_irq(&img_request->completion_lock);
> if (which > img_request->next_completion)
> goto out;
> + if (which != img_request->next_completion) {
> + printk("%s: bad image object request information:\n", __func__);
> + printk("obj_request %p\n", obj_request);
> + printk(" ->object_name <%s>\n", obj_request->object_name);
> + printk(" ->offset %llu\n", obj_request->offset);
> + printk(" ->length %llu\n", obj_request->length);
> + printk(" ->type 0x%x\n", (u32)obj_request->type);
> + printk(" ->flags 0x%lx\n", obj_request->flags);
> + printk(" ->img_request %p\n", obj_request->img_request);
> + printk(" ->which %u\n", obj_request->which);
> + printk(" ->xferred %llu\n", obj_request->xferred);
> + printk(" ->result %d\n", obj_request->result);
> + printk(" ->kref %d\n",
> + atomic_read(&obj_request->kref.refcount));
> +
> + printk("img_request %p\n", img_request);
> + printk(" ->snap 0x%016llx\n", img_request->snap_id);
> + printk(" ->offset %llu\n", img_request->offset);
> + printk(" ->length %llu\n", img_request->length);
> + printk(" ->flags 0x%lx\n", img_request->flags);
> + printk(" ->obj_request_count %u\n",
> + img_request->obj_request_count);
> + printk(" ->next_completion %u\n",
> + img_request->next_completion);
> + printk(" ->xferred %llu\n", img_request->xferred);
> + printk(" ->result %d\n", img_request->result);
> + printk(" ->obj_requests head %p\n",
> + img_request->obj_requests.next);
> + printk(" ->kref %d\n",
> + atomic_read(&img_request->kref.refcount));
> + }
> rbd_assert(which == img_request->next_completion);
>
> for_each_obj_request_from(img_request, obj_request) {
> @@ -2697,11 +2735,21 @@ static int rbd_img_request_submit(struct
> {
> struct rbd_obj_request *obj_request;
> struct rbd_obj_request *next_obj_request;
> + bool verbose = false;
>
> dout("%s: img %p\n", __func__, img_request);
> + if (img_request->obj_request_count > 1) {
> + printk("%s: img_request %p count %u (%llu/%llu)\n", __func__,
> + img_request, img_request->offset, img_request->length);
> + verbose = true;
> + }
> for_each_obj_request_safe(img_request, obj_request, next_obj_request) {
> int ret;
>
> + if (verbose)
> + printk("%s: obj_request %p (%llu/%llu)\n", __func__,
> + obj_request, obj_request->offset,
> + obj_request->length);
> ret = rbd_img_obj_request_submit(obj_request);
> if (ret)
> return ret;
>
>
Thanks ! I will check that tomorrow.
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2014-03-26 4:00 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-03-25 8:39 Issue #5876 : assertion failure in rbd_img_obj_callback() Olivier Bonvalet
2014-03-25 9:04 ` Ilya Dryomov
[not found] ` <1395739214.2823.34.camel@localhost>
2014-03-25 9:52 ` Ilya Dryomov
2014-03-25 11:48 ` Alex Elder
2014-03-25 12:34 ` Ilya Dryomov
2014-03-25 12:51 ` Alex Elder
2014-03-25 12:57 ` Ilya Dryomov
2014-03-25 13:18 ` Olivier Bonvalet
2014-03-25 13:29 ` Alex Elder
2014-03-25 13:31 ` Alex Elder
2014-03-25 14:01 ` Olivier Bonvalet
2014-03-25 17:15 ` Olivier Bonvalet
2014-03-25 17:21 ` Alex Elder
2014-03-25 18:53 ` Olivier Bonvalet
2014-03-25 17:43 ` Alex Elder
2014-03-25 18:53 ` Olivier Bonvalet
2014-03-25 19:03 ` Alex Elder
2014-03-25 20:18 ` Ilya Dryomov
2014-03-25 20:21 ` Olivier Bonvalet
2014-03-25 20:24 ` Alex Elder
2014-03-25 20:29 ` Olivier Bonvalet
2014-03-25 20:44 ` Alex Elder
2014-03-25 21:03 ` Olivier Bonvalet
2014-03-25 20:41 ` Alex Elder
2014-03-25 20:53 ` Olivier Bonvalet
2014-03-25 21:10 ` Olivier Bonvalet
2014-03-25 21:20 ` Ilya Dryomov
[not found] ` <1395782577.2076.23.camel@localhost>
2014-03-25 21:25 ` Ilya Dryomov
2014-03-25 21:41 ` Olivier Bonvalet
2014-03-25 21:49 ` Ilya Dryomov
2014-03-25 21:54 ` Olivier Bonvalet
2014-03-25 22:17 ` Olivier Bonvalet
2014-03-25 22:46 ` Alex Elder
2014-03-25 23:04 ` Olivier Bonvalet
2014-03-26 0:00 ` Alex Elder
2014-03-26 1:33 ` Olivier Bonvalet
2014-03-26 1:50 ` Olivier Bonvalet
2014-03-26 1:55 ` Alex Elder
2014-03-26 2:40 ` Olivier Bonvalet
2014-03-26 2:42 ` Alex Elder
2014-03-26 2:45 ` Olivier Bonvalet
2014-03-26 3:54 ` Alex Elder
2014-03-26 4:00 ` Olivier Bonvalet [this message]
2014-03-26 5:00 ` Alex Elder
2014-03-26 11:13 ` Alex Elder
2014-03-26 11:43 ` Ilya Dryomov
2014-03-26 11:47 ` Alex Elder
2014-03-26 12:05 ` Ilya Dryomov
2014-03-26 20:58 ` Alex Elder
2014-03-27 7:48 ` Olivier Bonvalet
2014-03-27 8:45 ` Ilya Dryomov
2014-03-27 8:49 ` Olivier Bonvalet
2014-03-26 2:35 ` Olivier Bonvalet
2014-03-26 2:54 ` Alex Elder
2014-03-26 3:58 ` Olivier Bonvalet
2014-04-05 1:16 ` Olivier Bonvalet
2014-04-05 1:57 ` Alex Elder
2014-04-05 8:09 ` Olivier Bonvalet
2014-04-05 13:08 ` Alex Elder
2014-04-25 11:37 ` Olivier Bonvalet
2014-04-25 12:17 ` Alex Elder
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1395806447.2076.70.camel@localhost \
--to=ceph.list@daevel.fr \
--cc=ceph-devel@vger.kernel.org \
--cc=elder@ieee.org \
--cc=ilya.dryomov@inktank.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox