qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Stefan Hajnoczi <stefanha@gmail.com>
To: Sam Li <faithilikerun@gmail.com>
Cc: qemu-devel <qemu-devel@nongnu.org>,
	 Damien Le Moal <damien.lemoal@opensource.wdc.com>,
	Markus Armbruster <armbru@redhat.com>,
	 Dmitry Fomichev <dmitry.fomichev@wdc.com>,
	Stefan Hajnoczi <stefanha@redhat.com>,
	 Hanna Reitz <hreitz@redhat.com>,
	qemu block <qemu-block@nongnu.org>,
	 Eric Blake <eblake@redhat.com>, Kevin Wolf <kwolf@redhat.com>,
	Fam Zheng <fam@euphon.net>, Hannes Reinecke <hare@suse.de>
Subject: Re: [RFC v4 7/9] config: add check to block layer
Date: Wed, 27 Jul 2022 10:50:38 -0400	[thread overview]
Message-ID: <CAJSP0QVTGa0smONqrFXJW=tVpDcax+M4j3-tQFYeK23QL3FBnA@mail.gmail.com> (raw)
In-Reply-To: <20220712021345.8530-8-faithilikerun@gmail.com>

On Mon, 11 Jul 2022 at 22:22, Sam Li <faithilikerun@gmail.com> wrote:
>
> Putting zoned/non-zoned BlockDrivers on top of each other is not
> allowed.
>
> Signed-off-by: Sam Li <faithilikerun@gmail.com>
> ---
>  block.c                          | 7 +++++++
>  block/file-posix.c               | 2 ++
>  include/block/block_int-common.h | 5 +++++
>  3 files changed, 14 insertions(+)
>
> diff --git a/block.c b/block.c
> index 2c00dddd80..0e24582c7d 100644
> --- a/block.c
> +++ b/block.c
> @@ -7945,6 +7945,13 @@ void bdrv_add_child(BlockDriverState *parent_bs, BlockDriverState *child_bs,
>          return;
>      }
>
> +    if (parent_bs->drv->is_zoned != child_bs->drv->is_zoned) {
> +        error_setg(errp, "Cannot add a %s child to a %s parent",
> +                   child_bs->drv->is_zoned ? "zoned" : "non-zoned",
> +                   parent_bs->drv->is_zoned ? "zoned" : "non-zoned");
> +        return;
> +    }

Please explain the rationale:

/*
 * Non-zoned block drivers do not follow zoned storage constraints
(i.e. sequential writes
 * to zones). Refuse mixing zoned and non-zoned drivers in a graph.
 */

> +
>      if (!QLIST_EMPTY(&child_bs->parents)) {
>          error_setg(errp, "The node %s already has a parent",
>                     child_bs->node_name);
> diff --git a/block/file-posix.c b/block/file-posix.c
> index 42708012ff..e9ad1d8e1e 100644
> --- a/block/file-posix.c
> +++ b/block/file-posix.c
> @@ -3924,6 +3924,7 @@ static BlockDriver bdrv_host_device = {
>      .format_name        = "host_device",
>      .protocol_name        = "host_device",
>      .instance_size      = sizeof(BDRVRawState),
> +    .is_zoned = false,

In C static struct fields are automatically initialized to 0/false.
This line can be omitted.

>      .bdrv_needs_filename = true,
>      .bdrv_probe_device  = hdev_probe_device,
>      .bdrv_parse_filename = hdev_parse_filename,
> @@ -3971,6 +3972,7 @@ static BlockDriver bdrv_zoned_host_device = {
>          .format_name = "zoned_host_device",
>          .protocol_name = "zoned_host_device",
>          .instance_size = sizeof(BDRVRawState),
> +        .is_zoned = true,
>          .bdrv_needs_filename = true,
>          .bdrv_probe_device  = hdev_probe_device,
>          .bdrv_parse_filename = hdev_parse_filename,
> diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h
> index 6037871089..29f1ec9184 100644
> --- a/include/block/block_int-common.h
> +++ b/include/block/block_int-common.h
> @@ -141,6 +141,11 @@ struct BlockDriver {
>       */
>      bool is_format;
>
> +    /*
> +     * Set to true if the BlockDriver is a zoned block driver.
> +     */
> +    bool is_zoned;

This isn't powerful enough to express the constraints.
block/raw-format.c supports both non-zoned and zoned children (after
your patch) but it won't pass the check.

Perhaps add bool supports_zoned_children and change the check to:

if (child_bs->drv->is_zoned &&
!parent_bs->drv->supports_zoned_children) { ...refuse... }

Then raw-format would work on top of zoned_host_device as well as
still working on non-zoned BDSes.

Stefan


  reply	other threads:[~2022-07-27 14:55 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-07-12  2:13 [RFC v4 0/9] Add support for zoned device Sam Li
2022-07-12  2:13 ` [RFC v4 1/9] block: add block layer APIs resembling Linux ZonedBlockDevice ioctls Sam Li
2022-07-12  6:10   ` Hannes Reinecke
2022-07-12  6:17     ` Sam Li
2022-07-12  7:35   ` Damien Le Moal
2022-07-13  0:54     ` Sam Li
2022-07-12 15:49   ` Stefan Hajnoczi
2022-07-12 22:12     ` Damien Le Moal
2022-07-13  6:22       ` Stefan Hajnoczi
2022-07-13  0:51     ` Sam Li
2022-07-13  6:19       ` Stefan Hajnoczi
2022-07-12  2:13 ` [RFC v4 2/9] qemu-io: add zoned block device operations Sam Li
2022-07-12  6:14   ` Hannes Reinecke
2022-07-12  7:44   ` Damien Le Moal
2022-07-27 14:13   ` Stefan Hajnoczi
2022-07-28  1:57     ` Damien Le Moal
2022-07-12  2:13 ` [RFC v4 3/9] file-posix: introduce get_sysfs_long_val for a block queue of sysfs attribute Sam Li
2022-07-12  6:16   ` Hannes Reinecke
2022-07-12  7:37   ` Damien Le Moal
2022-07-12  2:13 ` [RFC v4 4/9] file-posix: introduce get_sysfs_str_val for device zoned model Sam Li
2022-07-12  6:17   ` Hannes Reinecke
2022-07-12  6:35     ` Damien Le Moal
2022-07-12  7:42   ` Damien Le Moal
2022-07-12  2:13 ` [RFC v4 5/9] qemu-iotests: test new zone operations Sam Li
2022-07-27 14:34   ` Stefan Hajnoczi
2022-07-27 14:59     ` Ming Lei
2022-07-27 15:12       ` Stefan Hajnoczi
2022-07-12  2:13 ` [RFC v4 6/9] raw-format: add " Sam Li
2022-07-27 14:39   ` Stefan Hajnoczi
2022-07-12  2:13 ` [RFC v4 7/9] config: add check to block layer Sam Li
2022-07-27 14:50   ` Stefan Hajnoczi [this message]
2022-07-12  2:13 ` [RFC v4 8/9] include: add support for zoned block devices Sam Li
2022-07-27 14:52   ` Stefan Hajnoczi
2022-07-12  2:13 ` [RFC v4 9/9] qapi: add support for zoned host device Sam Li
2022-07-12  7:48   ` Damien Le Moal
2022-07-27 14:53   ` Stefan Hajnoczi
2022-07-12  5:47 ` [RFC v4 0/9] Add support for zoned device Markus Armbruster
2022-07-12  5:59   ` Sam Li
2022-07-18 10:53     ` Markus Armbruster
2022-07-27 14:55 ` Stefan Hajnoczi
2022-07-27 15:06 ` Stefan Hajnoczi
2022-07-27 15:14   ` Sam Li

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='CAJSP0QVTGa0smONqrFXJW=tVpDcax+M4j3-tQFYeK23QL3FBnA@mail.gmail.com' \
    --to=stefanha@gmail.com \
    --cc=armbru@redhat.com \
    --cc=damien.lemoal@opensource.wdc.com \
    --cc=dmitry.fomichev@wdc.com \
    --cc=eblake@redhat.com \
    --cc=faithilikerun@gmail.com \
    --cc=fam@euphon.net \
    --cc=hare@suse.de \
    --cc=hreitz@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@redhat.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).