qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
To: Zhi Yong Wu <zwu.kernel@gmail.com>
Cc: kwolf@redhat.com, pair@us.ibm.com, kvm@vger.kernel.org,
	Stefan Hajnoczi <stefanha@gmail.com>,
	mtosatti@redhat.com, qemu-devel@nongnu.org, ryanh@us.ibm.com,
	Zhi Yong Wu <wuzhy@linux.vnet.ibm.com>,
	luowenj@cn.ibm.com
Subject: Re: [Qemu-devel] [PATCH v5 2/4] block: add the block queue support
Date: Wed, 10 Aug 2011 10:37:11 +0100	[thread overview]
Message-ID: <20110810093711.GD28852@stefanha-thinkpad.localdomain> (raw)
In-Reply-To: <CAEH94Lh2c4ysT1Grtw+M5E+7zfktuWyRXWF_9TZd8pbKESH20w@mail.gmail.com>

On Wed, Aug 10, 2011 at 01:54:33PM +0800, Zhi Yong Wu wrote:
> On Tue, Aug 9, 2011 at 8:49 PM, Stefan Hajnoczi <stefanha@gmail.com> wrote:
> > On Tue, Aug 9, 2011 at 5:17 AM, Zhi Yong Wu <wuzhy@linux.vnet.ibm.com> wrote:
> >> +BlockDriverAIOCB *qemu_block_queue_enqueue(BlockQueue *queue,
> >> +                        BlockDriverState *bs,
> >> +                        BlockRequestHandler *handler,
> >> +                        int64_t sector_num,
> >> +                        QEMUIOVector *qiov,
> >> +                        int nb_sectors,
> >> +                        BlockDriverCompletionFunc *cb,
> >> +                        void *opaque)
> >> +{
> >> +    BlockIORequest *request;
> >> +    BlockDriverAIOCB *acb;
> >> +
> >> +    request = qemu_malloc(sizeof(BlockIORequest));
> >> +    request->bs = bs;
> >> +    request->handler = handler;
> >> +    request->sector_num = sector_num;
> >> +    request->qiov = qiov;
> >> +    request->nb_sectors = nb_sectors;
> >> +    request->cb = cb;
> >> +    request->opaque = opaque;
> >> +
> >> +    QTAILQ_INSERT_TAIL(&queue->requests, request, entry);
> >
> > It would be simpler to define BlockQueueAIOCB and using it as our acb
> > instead of managing an extra BlockIORequest structure.  That way you
> > don't need to worry about extra mallocs and frees.
> Sorry, i don't get what you mean. how to define it? Can you elaborate?

BlockDriverAIOCB is designed to be embedded inside a bigger struct.  For
example, QEDAIOCB is a larger struct that contains BlockDriverAIOCB as
its first field:

typedef struct QEDAIOCB {
    BlockDriverAIOCB common;
    ...
} QEDAIOCB;

And the QED AIOPool contains the size of QEDAIOCB so that qemu_aio_get() can
allocate the full QEDAIOCB struct:

static AIOPool qed_aio_pool = {
    .aiocb_size         = sizeof(QEDAIOCB),
    .cancel             = qed_aio_cancel,
};

This allows QED to store per-request state in QEDAIOCB for the lifetime of a
request:

QEDAIOCB *acb = qemu_aio_get(&qed_aio_pool, bs, cb, opaque);

acb->is_write = is_write;
acb->finished = NULL;
acb->qiov = qiov;
...

I suggest creating a BlockQueueAIOCB that contains the fields from
BlockIORequest (which is no longer needed as a separate struct):

typedef struct BlockQueueAIOCB {
    BlockDriverAIOCB common;
    BlockRequestHandler *handler;
    int64_t sector_num;
    QEMUIOVector *qiov;
    int nb_sectors;
    QTAILQ_ENTRY(BlockQueueAIOCB) entry; /* pending request queue */
} BlockQueueAIOCB;

Now you can drop the malloc and simply qemu_aio_get() a new
BlockQueueAIOCB.

Stefan

  reply	other threads:[~2011-08-10  9:37 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-08-09  4:17 [Qemu-devel] [PATCH v5 0/4] The intro of QEMU block I/O throttling Zhi Yong Wu
2011-08-09  4:17 ` [Qemu-devel] [PATCH v5 1/4] block: add the command line support Zhi Yong Wu
2011-08-09 12:25   ` Stefan Hajnoczi
2011-08-10  5:20     ` Zhi Yong Wu
2011-08-10  9:27       ` Stefan Hajnoczi
2011-08-11  4:44         ` Zhi Yong Wu
2011-08-11  5:35           ` Stefan Hajnoczi
2011-08-09  4:17 ` [Qemu-devel] [PATCH v5 2/4] block: add the block queue support Zhi Yong Wu
2011-08-09  8:46   ` Ram Pai
2011-08-09  8:53     ` Zhi Yong Wu
2011-08-09 12:49   ` Stefan Hajnoczi
2011-08-10  5:54     ` Zhi Yong Wu
2011-08-10  9:37       ` Stefan Hajnoczi [this message]
2011-08-11  4:59         ` Zhi Yong Wu
2011-08-11  5:36         ` Zhi Yong Wu
2011-08-12  4:40         ` Zhi Yong Wu
2011-08-12  4:50           ` Stefan Hajnoczi
2011-08-12  8:10     ` Zhi Yong Wu
2011-08-12  8:42       ` Stefan Hajnoczi
2011-08-12  9:11         ` Zhi Yong Wu
2011-08-09  4:17 ` [Qemu-devel] [PATCH v5 3/4] block: add block timer and block throttling algorithm Zhi Yong Wu
2011-08-09  8:57   ` Ram Pai
2011-08-09  9:06     ` Zhi Yong Wu
2011-08-12  5:35     ` Zhi Yong Wu
2011-08-12  5:47       ` Stefan Hajnoczi
2011-08-12  6:00         ` Zhi Yong Wu
2011-08-09 15:19   ` Stefan Hajnoczi
2011-08-10  6:57     ` Zhi Yong Wu
2011-08-10 11:00       ` Stefan Hajnoczi
2011-08-12  5:00         ` Zhi Yong Wu
2011-08-12  5:06           ` Stefan Hajnoczi
2011-08-12  5:23             ` Zhi Yong Wu
2011-08-09  4:17 ` [Qemu-devel] [PATCH v5 4/4] qmp/hmp: add block_set_io_throttle Zhi Yong Wu
2011-08-09 12:08 ` [Qemu-devel] [PATCH v5 0/4] The intro of QEMU block I/O throttling Stefan Hajnoczi
2011-08-10  5:09   ` Zhi Yong Wu
2011-08-10  9:39     ` Stefan Hajnoczi

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=20110810093711.GD28852@stefanha-thinkpad.localdomain \
    --to=stefanha@linux.vnet.ibm.com \
    --cc=kvm@vger.kernel.org \
    --cc=kwolf@redhat.com \
    --cc=luowenj@cn.ibm.com \
    --cc=mtosatti@redhat.com \
    --cc=pair@us.ibm.com \
    --cc=qemu-devel@nongnu.org \
    --cc=ryanh@us.ibm.com \
    --cc=stefanha@gmail.com \
    --cc=wuzhy@linux.vnet.ibm.com \
    --cc=zwu.kernel@gmail.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).