qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH v4 0/2] block/qapi: refactor and optimize the qm


From: Dou Liyang
Subject: Re: [Qemu-devel] [PATCH v4 0/2] block/qapi: refactor and optimize the qmp_query_blockstats()
Date: Mon, 23 Jan 2017 13:31:59 +0800
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.4.0

Hi, Markus

Here is the non-RFC patch.
Any suggestions are welcome. :)

Thanks,
        Liyang.

At 01/15/2017 04:01 PM, Dou Liyang wrote:
Change log v3 -> v4:
 1. Develop these into the non-RFC patches.
 2. Fix some comments.
 3. do declarations first.

Change log v2 -> v3:
 1. Remove the unnecessary code for the bdrv_next_node().
 2. Remove the change of the locking rules.
    Even if this change can improve the performance, but it may
    effect the consistency.

The Qemu uses the qmp_query_blockstats() function to get the
blockstats. As the function has been changed several times, it
becomes more complex and longer.

For the multi-disks guest, if we want to execute I/O operations
and this function at the same time. we can use the dataplane
feature to hold I/O performance does not drop. But, Normally
without this feature, How to reduce the decline in performance?

These patches refactor the qmp_query_blockstats() to make the
code easier to follow, and shorter as follows:

From:

+--------------+      +---------------------+
 | 1            |      | 4.                  |
 |next_query_bds|      |bdrv_query_bds_stats +---+
 |              |      |                     |   |
 +--------^-----+      +-------------^-------+   |
          |                          |           |
+---------+----------+      +--------+-------+   |
| 0.                 |      | 2.             |   |
|qmp_query_blockstats+------>bdrv_query_stats<----
|                    |      |                |
+--------------------+      +--------+-------+
                                     |
                       +-------------v-------+
                       | 3.                  |
                       |bdrv_query_blk_stats |
                       |                     |
                       +---------------------+

To:

                                    +--------------+
                                    |              |
                           +--------v-----------+  |
                       +--->  3.                |  |
+-------------------+  |   |bdrv_query_bds_stats+--+
| 1.                +--+   |                    |
|                   +      +--------------------+
|qmp_query_blockstats--+
|                   |  |
+-------------------+  |   +--------------------+
                       |   | 2.                 |
                       +--->                    |
                           |bdrv_query_blk_stats|
                           |                    |
                           +--------------------+


They also optimize the fuction by reducing its running time.

1. The function running time

the time it takes(ns) in each requests:
the disk numbers     | 10    | 500
-------------------------------------
before these patches | 19429 | 667722
after these patches  | 18536 | 627945

2. For the performance
used the dd command likes this to test:
dd if=date_1.dat of=date_2.dat conv=fsync oflag=direct bs=1k count=100k.

the I/O performance is degraded(%) during the monitor:
the disk numbers     | 10    | 500
-------------------------------------
before these patches | 1.3   | 14.2
after these patches  | 1.0   | 11.3

Dou Liyang (2):
  block/qapi: reduce the coupling between the bdrv_query_stats and
    bdrv_query_bds_stats
  block/qapi: reduce the execution time of qmp_query_blockstats

 block/qapi.c | 95 ++++++++++++++++++++++++++----------------------------------
 1 file changed, 41 insertions(+), 54 deletions(-)






reply via email to

[Prev in Thread] Current Thread [Next in Thread]