qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Denis Plotnikov <dplotnikov@virtuozzo.com>
To: "qemu-block@nongnu.org" <qemu-block@nongnu.org>
Cc: "kwolf@redhat.com" <kwolf@redhat.com>,
	Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>,
	"berrange@redhat.com" <berrange@redhat.com>,
	"ehabkost@redhat.com" <ehabkost@redhat.com>,
	"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>,
	"mreitz@redhat.com" <mreitz@redhat.com>,
	"pbonzini@redhat.com" <pbonzini@redhat.com>,
	Denis Lunev <den@virtuozzo.com>
Subject: Re: [PATCH v0 2/2] block: allow to set 'drive' property on a realized block device
Date: Mon, 18 Nov 2019 10:50:30 +0000	[thread overview]
Message-ID: <c38c1301-a28f-baf1-d441-ae8c7205feb1@virtuozzo.com> (raw)
In-Reply-To: <c19343f9-73b2-89a4-2b7d-b101a0b20016@virtuozzo.com>



On 10.11.2019 22:08, Denis Plotnikov wrote:
>
> On 10.11.2019 22:03, Denis Plotnikov wrote:
>> This allows to change (replace) the file on a block device and is useful
>> to workaround exclusive file access restrictions, e.g. to implement VM
>> migration with a shared disk stored on some storage with the exclusive
>> file opening model: a destination VM is started waiting for incomming
>> migration with a fake image drive, and later, on the last migration
>> phase, the fake image file is replaced with the real one.
>>
>> Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
>> ---
>>   hw/core/qdev-properties-system.c | 89 +++++++++++++++++++++++++++-----
>>   1 file changed, 77 insertions(+), 12 deletions(-)
>>
>> diff --git a/hw/core/qdev-properties-system.c 
>> b/hw/core/qdev-properties-system.c
>> index c534590dcd..aaab1370a4 100644
>> --- a/hw/core/qdev-properties-system.c
>> +++ b/hw/core/qdev-properties-system.c
>> @@ -79,8 +79,55 @@ static void set_pointer(Object *obj, Visitor *v, 
>> Property *prop,
>>     /* --- drive --- */
>>   -static void do_parse_drive(DeviceState *dev, const char *str, void 
>> **ptr,
>> -                           const char *propname, bool iothread, 
>> Error **errp)
>> +static void do_parse_drive_realized(DeviceState *dev, const char *str,
>> +                                    void **ptr, const char *propname,
>> +                                    bool iothread, Error **errp)
>> +{
>> +    BlockBackend *blk = *ptr;
>> +    BlockDriverState *bs = bdrv_lookup_bs(NULL, str, NULL);
>> +    int ret;
>> +    bool blk_created = false;
>> +
>> +    if (!bs) {
>> +        error_setg(errp, "Can't find blockdev '%s'", str);
>> +        return;
>> +    }
>> +
>> +    if (!blk) {
>> +        AioContext *ctx = iothread ? bdrv_get_aio_context(bs) :
>> +                                     qemu_get_aio_context();
>> +        blk = blk_new(ctx, BLK_PERM_ALL, BLK_PERM_ALL);
>> +        blk_created = true;
>
> Actually, I have concerns about situation where blk=null.
>
> Is there any case when scsi-hd (or others) doesn't have a blk assigned 
> and it's legal?
>
>> +    } else {
>> +        if (blk_bs(blk)) {
>> +            blk_remove_bs(blk);
>> +        }
>> +    }
>> +
>> +    ret = blk_insert_bs(blk, bs, errp);
>> +
>> +    if (!ret && blk_created) {
>> +        if (blk_attach_dev(blk, dev) < 0) {
>> +            /*
>> +             * Shouldn't be any errors here since we just created
>> +             * the new blk because the device doesn't have any.
>> +             * Leave the message here in case blk_attach_dev is changed
>> +             */
>> +             error_setg(errp, "Can't attach drive '%s' to device '%s'",
>> +                        str, object_get_typename(OBJECT(dev)));
>> +        } else {
>> +            *ptr = blk;
>> +        }
>> +    }
Another problem here, is that the "size" of the device dev may not match 
after setting a drive.
So, we should update it after the drive setting.
It was found, that it could be done by calling 
BlockDevOps.bdrv_parent_cb_resize.

But I have some concerns about doing it so. In the case of virtio scsi 
disk we have the following callstack

     bdrv_parent_cb_resize calls() ->
         scsi_device_report_change(dev, SENSE_CODE(CAPACITY_CHANGED)) ->
             virtio_scsi_change ->
                 virtio_scsi_push_event(s, dev, VIRTIO_SCSI_T_PARAM_CHANGE,
                                                     sense.asc | 
(sense.ascq << 8));


virtio_scsi_change  pushes the event to the guest to make the guest ask 
for size refreshing.
If I'm not mistaken, here we can get a race condition when some another 
request is processed with an unchanged
size and then the size changing request is processed.

I didn't find a better way to update device size so any comments are 
welcome.

Thanks!

Denis
>> +
>> +    if (blk_created) {
>> +        blk_unref(blk);
>> +    }
>> +}
>> +
>> +static void do_parse_drive_unrealized(DeviceState *dev, const char 
>> *str,
>> +                                      void **ptr, const char *propname,
>> +                                      bool iothread, Error **errp)
>>   {
>>       BlockBackend *blk;
>>       bool blk_created = false;
>> @@ -137,18 +184,34 @@ fail:
>>       }
>>   }
>>   -static void parse_drive(DeviceState *dev, const char *str, void 
>> **ptr,
>> -                        const char *propname, Error **errp)
>> -{
>> -    do_parse_drive(dev, str, ptr, propname, false, errp);
>> -}
>> -
>> -static void parse_drive_iothread(DeviceState *dev, const char *str, 
>> void **ptr,
>> +static void parse_drive_realized(DeviceState *dev, const char *str, 
>> void **ptr,
>>                                    const char *propname, Error **errp)
>>   {
>> -    do_parse_drive(dev, str, ptr, propname, true, errp);
>> +    do_parse_drive_realized(dev, str, ptr, propname, false, errp);
>>   }
>>   +static void parse_drive_realized_iothread(DeviceState *dev, const 
>> char *str,
>> +                                          void **ptr, const char 
>> *propname,
>> +                                          Error **errp)
>> +{
>> +    do_parse_drive_realized(dev, str, ptr, propname, true, errp);
>> +}
>> +
>> +static void parse_drive_unrealized(DeviceState *dev, const char *str,
>> +                                   void **ptr, const char *propname,
>> +                                   Error **errp)
>> +{
>> +    do_parse_drive_unrealized(dev, str, ptr, propname, false, errp);
>> +}
>> +
>> +static void parse_drive_unrealized_iothread(DeviceState *dev, const 
>> char *str,
>> +                                            void **ptr, const char 
>> *propname,
>> +                                            Error **errp)
>> +{
>> +    do_parse_drive_unrealized(dev, str, ptr, propname, true, errp);
>> +}
>> +
>> +
>>   static void release_drive(Object *obj, const char *name, void *opaque)
>>   {
>>       DeviceState *dev = DEVICE(obj);
>> @@ -188,13 +251,15 @@ static void get_drive(Object *obj, Visitor *v, 
>> const char *name, void *opaque,
>>   static void set_drive(Object *obj, Visitor *v, const char *name, 
>> void *opaque,
>>                         Error **errp)
>>   {
>> -    set_pointer(obj, v, opaque, NULL, parse_drive, name, errp);
>> +    set_pointer(obj, v, opaque, parse_drive_realized, 
>> parse_drive_unrealized,
>> +                name, errp);
>>   }
>>     static void set_drive_iothread(Object *obj, Visitor *v, const 
>> char *name,
>>                                  void *opaque, Error **errp)
>>   {
>> -    set_pointer(obj, v, opaque, NULL, parse_drive_iothread, name, 
>> errp);
>> +    set_pointer(obj, v, opaque, parse_drive_realized_iothread,
>> +                parse_drive_unrealized_iothread, name, errp);
>>   }
>>     const PropertyInfo qdev_prop_drive = {


  reply	other threads:[~2019-11-18 10:51 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-10 19:03 [PATCH v0 0/2] allow to set 'drive' property on a realized block device Denis Plotnikov
2019-11-10 19:03 ` [PATCH v0 1/2] qdev-properties-system: extend set_pionter for unrealized devices Denis Plotnikov
2019-11-18 18:54   ` Eduardo Habkost
2019-11-22 11:36     ` Denis Plotnikov
2019-11-25 15:30       ` Eduardo Habkost
2019-11-26  6:49         ` Denis Plotnikov
2019-11-26 16:38           ` Kevin Wolf
2019-11-10 19:03 ` [PATCH v0 2/2] block: allow to set 'drive' property on a realized block device Denis Plotnikov
2019-11-10 19:08   ` Denis Plotnikov
2019-11-18 10:50     ` Denis Plotnikov [this message]
2019-12-13  7:30       ` [PING]Re: " Denis Plotnikov
2019-12-13 10:32       ` Kevin Wolf
2019-12-16 14:51         ` Denis Plotnikov
2019-12-16 15:38           ` Kevin Wolf
2019-12-16 15:58             ` Denis Plotnikov
2019-11-18 10:30 ` [PATCH v0 0/2] " Denis Plotnikov
2020-03-02 13:38 ` Kevin Wolf
2020-03-02 13:55   ` Denis Plotnikov
2020-03-02 15:39     ` Kevin Wolf
2020-03-03  7:43       ` Vladimir Sementsov-Ogievskiy

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=c38c1301-a28f-baf1-d441-ae8c7205feb1@virtuozzo.com \
    --to=dplotnikov@virtuozzo.com \
    --cc=berrange@redhat.com \
    --cc=den@virtuozzo.com \
    --cc=ehabkost@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=mreitz@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=vsementsov@virtuozzo.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;
as well as URLs for NNTP newsgroup(s).