virtualization.lists.linux-foundation.org archive mirror
 help / color / mirror / Atom feed
From: Asias He <asias@redhat.com>
To: Rusty Russell <rusty@rustcorp.com.au>
Cc: mst@redhat.com, kvm@vger.kernel.org,
	virtualization@lists.linux-foundation.org
Subject: Re: [PATCH] virtio-blk: Don't free ida when disk is in use
Date: Thu, 20 Dec 2012 16:46:48 +0800	[thread overview]
Message-ID: <50D2D078.2070408@redhat.com> (raw)
In-Reply-To: <878v8t1hri.fsf@rustcorp.com.au>

On 12/20/2012 12:15 PM, Rusty Russell wrote:
> Alexander Graf <agraf@suse.de> writes:
> 
>> When a file system is mounted on a virtio-blk disk, we then remove it
>> and then reattach it, the reattached disk gets the same disk name and
>> ids as the hot removed one.
>>
>> This leads to very nasty effects - mostly rendering the newly attached
>> device completely unusable.
>>
>> Trying what happens when I do the same thing with a USB device, I saw
>> that the sd node simply doesn't get free'd when a device gets forcefully
>> removed.
>>
>> Imitate the same behavior for vd devices. This way broken vd devices
>> simply are never free'd and newly attached ones keep working just fine.
>>
>> Signed-off-by: Alexander Graf <agraf@suse.de>
> 
> I think deserves a CC:stable, no?
> 
> I've put it in my pending queue for *next* merge window for now...

Thanks for looking into this, Alexander!

I also noticed this problem. The problem is that, if you hot-unplug a
mounted or opened disk, the disk is in opened state. Next time, when you
hotplug the same disk. The kernel thought it was opened already. The
driver will use the wrong gendisk data structure in bdev.

blkdev_open
   blkdev_get
      __blkdev_get
         if (!bdev->bd_openers) {    <-- Here, bd_disk not got updated
                                         still points to old one
           bdev->bd_disk = disk;
           bdev->bd_queue = disk->queue;
           ...

I tried something like this:

@@ -854,6 +862,19 @@ static int __devinit virtblk_probe(struct
virtio_device *vdev)
                blk_queue_io_opt(q, blk_size * opt_io_size);

        add_disk(vblk->disk);
+
+       for (i = 0; i < 1 << PART_BITS; i++) {
+               bdev = bdget_disk(vblk->disk, i);
+               if (bdev) {
+                       bdev->bd_disk = vblk->disk;
+                       bdev->bd_queue = q;
+                       bdput(bdev);
+               }
+       }


1) Before:
---> hot-plug
[   35.730183] virtio_blk: virtblk_probe: vblk=ffff880078b0e000,
disk=ffff880078d54c00, q=ffff88007f88b3d8
[   35.735352] virtio_blk:    virtblk_ioctl: vblk=ffff880078b0e000,
disk=ffff880078d54c00, bdev=ffff88007c45cc00, q=ffff88007f88b3d8
---> hot-unplug

---> hot-plug
[   83.570480] virtio_blk: virtblk_probe: vblk=ffff880078b0e000,
disk=ffff880078d55800, q=ffff88007f88bb40
[   83.575614] virtio_blk:   virtblk_ioctl: vblk=ffff880078b0e000,
disk=ffff880078d54c00, bdev=ffff88007c45cc00, q=ffff88007f88b3d8

The disk points to old one ffff880078d54c00. The queue also points to
old one ffff88007f88b3d8.

2) After:
---> hot-plug
[   68.035063] virtio_blk: virtblk_probe: vblk=ffff880079b20000,
disk=ffff88007f9ebc00, q=ffff8800784d8ed0
[   68.041140] virtio_blk:    virtblk_ioctl: vblk=ffff880079b20000,
disk=ffff88007f9ebc00, bdev=ffff88007ab2c000, q=ffff8800784d8ed0
---> hot-unplug

---> hot-plug
[   86.317706] virtio_blk: virtblk_probe: vblk=ffff880079b20000,
disk=ffff88007f9eb000, q=ffff8800784d9638
[   86.322535] virtio_blk:   virtblk_ioctl: vblk=ffff880079b20000,
disk=ffff88007f9eb000, bdev=ffff88007ab2c000, q=ffff8800784d9638

The disk and queue are updated correctly. The attached disk works and
still uses the old name.

-- 
Asias

  reply	other threads:[~2012-12-20  8:46 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-12-19 19:40 [PATCH] virtio-blk: Don't free ida when disk is in use Alexander Graf
2012-12-20  4:15 ` Rusty Russell
2012-12-20  8:46   ` Asias He [this message]
2012-12-20  9:41     ` Alexander Graf
2012-12-21  1:48       ` Asias He
2013-01-02  5:09       ` Rusty Russell
2012-12-20 10:54 ` Michael S. Tsirkin
2012-12-20 11:27   ` Alexander Graf
2012-12-21  1:57   ` Asias He
     [not found]   ` <62D6A704-CB88-4A8C-A5F3-6BD3C267895F@suse.de>
2012-12-20 11:38     ` Michael S. Tsirkin
2012-12-20 11:47       ` Alexander Graf
2012-12-21  2:02         ` Asias He
2012-12-21  1:58     ` Asias He

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=50D2D078.2070408@redhat.com \
    --to=asias@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=mst@redhat.com \
    --cc=rusty@rustcorp.com.au \
    --cc=virtualization@lists.linux-foundation.org \
    /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;
as well as URLs for NNTP newsgroup(s).