From: supriya kannery <supriyak@in.ibm.com>
To: Stefan Hajnoczi <stefanha@gmail.com>
Cc: Kevin Wolf <kwolf@redhat.com>,
supriyak@linux.vnet.ibm.com,
Luiz Capitulino <lcapitulino@redhat.com>,
Christoph Hellwig <hch@lst.de>,
qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [v9 Patch 4/6]Qemu: Add commandline -drive option 'hostcache'
Date: Mon, 21 Nov 2011 17:58:34 +0530 [thread overview]
Message-ID: <4ECA43F2.7080506@in.ibm.com> (raw)
In-Reply-To: <CAJSP0QWZaCt2V-RGk_Ekomsw9XmgwJrawNU6RSQ3227Q2OrkrA@mail.gmail.com>
Stefan Hajnoczi wrote:
> On Thu, Nov 17, 2011 at 5:18 AM, Supriya Kannery
> <supriyak@linux.vnet.ibm.com> wrote:
>
>> On 11/17/2011 01:36 AM, Stefan Hajnoczi wrote:
>>
>>> On Fri, Nov 11, 2011 at 6:48 AM, Supriya Kannery
>>> <supriyak@linux.vnet.ibm.com> wrote:
>>>
>>>> + if ((hostcache = qemu_opt_get_bool(opts, "hostcache", -1)) !=
>>>> -1) {
>>>>
>>> This does not work. qemu_opt_get_bool() takes a bool default argument
>>> and returns a bool. (bool)-1 == true. But (int)true == 1 and you
>>> cannot expect it to ever equal -1.
>>>
>>> Try this:
>>>
>>> if (qemu_opt_get(opts, "hostcache")&&
>>> !qemu_opt_get_bool(opts, "hostcache", false)) {
>>> bdrv_flags |= BDRV_O_NOCACHE;
>>> }
>>>
>>> Stefan
>>>
>>>
>> Thanks! for pointing this.
>> Does the following look ok?
>>
>> if ((hostcache = qemu_opt_get_bool(opts, "hostcache", 1) == 0) {
>> bdrv_flags |= BDRV_O_NOCACHE;
>> }
>>
>> If either "hostcache" is not at all specified or it is specified
>> as "on", qemu_opt_get_bool will return 1, which can be ignored
>> as bdrv_flags is initialized to 0.
>>
>
> It depends on the overall way this should work. I think this captures
> all the cases:
>
> 1. cache= and hostcache= may not be used together.
> 2. cache= sets bdrv_flags.
> 3. hostcache= may |= BDRV_O_NOCACHE.
> 4. No option defaults to cache=writethrough (bdrv_flags &= ~BDRV_O_CACHE_MASK).
>
> The code you posted will work although I find it a bit weird how it
> also includes case #4. IMO it's cleanest to just do case #3 by
> testing whether or not the hostcache= option is set.
>
> BTW, is there a check for case #1 in your patch series. I thought I
> saw one earlier but now I can't find it.
>
Following is what I tried to accomplish:
1. cache= and hostcache= may be used together. Cache= will override hostcache= if specified together
This condition can be retained till cache-xx can be completely replaced by combinations of separate
cmdline options defined for setting hostcache, flush, and WCE
[I think I can change the current implementation to have Cache=xx ORed with hostcache=yy if
they are specified together instead of just ignoring hostcache= ]
2. If only cache= specified sets bdrv_flags.
3. If only hostcache= specified, bdrv_flags |= BDRV_O_NOCACHE
4. No option defaults to cache=writethrough (bdrv_flags &= ~BDRV_O_CACHE_MASK).
So the check I had earlier for case #1 in your list, I changed that to
allow both the options to co-exist.
-thanks, Supriya
> Stefan
>
>
next prev parent reply other threads:[~2011-11-21 12:49 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-11-11 6:47 [Qemu-devel] [v9 Patch 0/6]Qemu: Host pagecache setting from cmdline and monitor Supriya Kannery
2011-11-11 6:47 ` [Qemu-devel] [v9 Patch 1/6]Qemu: Enhance "info block" to display host cache setting Supriya Kannery
2011-11-11 10:09 ` [Qemu-devel] [v9 Patch 1/6 - updated]Qemu: " Supriya Kannery
2011-11-17 12:38 ` Luiz Capitulino
2011-11-18 9:14 ` supriya kannery
2011-11-11 6:47 ` [Qemu-devel] [v9 Patch 2/6]Qemu: Error classes for file reopen and data sync failure Supriya Kannery
2011-11-17 12:50 ` Luiz Capitulino
2011-11-18 9:29 ` supriya kannery
2011-11-18 12:09 ` Luiz Capitulino
2011-11-11 6:47 ` [Qemu-devel] [v9 Patch 3/6]Qemu: Cmd "block_set_hostcache" for dynamic cache change Supriya Kannery
2011-11-16 18:34 ` Stefan Hajnoczi
2011-11-17 5:45 ` Supriya Kannery
2011-11-17 13:14 ` Luiz Capitulino
2011-11-17 13:11 ` Luiz Capitulino
2011-11-18 10:44 ` supriya kannery
2011-11-11 6:48 ` [Qemu-devel] [v9 Patch 4/6]Qemu: Add commandline -drive option 'hostcache' Supriya Kannery
2011-11-16 20:06 ` Stefan Hajnoczi
2011-11-17 5:18 ` Supriya Kannery
2011-11-17 14:11 ` Stefan Hajnoczi
2011-11-21 12:28 ` supriya kannery [this message]
2011-11-21 14:03 ` Stefan Hajnoczi
2011-11-22 8:10 ` supriya kannery
2011-11-22 9:55 ` Kevin Wolf
2011-11-22 11:17 ` Stefan Hajnoczi
2011-11-22 11:31 ` Kevin Wolf
2011-11-11 6:48 ` [Qemu-devel] [v9 Patch 5/6]Qemu: Framework for reopening images safely Supriya Kannery
2011-11-17 13:16 ` Luiz Capitulino
2011-11-17 14:36 ` Stefan Hajnoczi
2011-11-21 12:13 ` supriya kannery
2011-11-21 14:31 ` Stefan Hajnoczi
2011-11-22 10:24 ` supriya kannery
2011-11-22 11:04 ` Kevin Wolf
2011-11-22 11:16 ` supriya kannery
2011-11-22 11:49 ` Stefan Hajnoczi
2011-11-23 3:52 ` Supriya Kannery
2011-11-11 6:48 ` [Qemu-devel] [v9 Patch 6/6]Qemu: raw posix implementation of reopen functions Supriya Kannery
2011-11-17 14:41 ` Stefan Hajnoczi
2011-11-21 12:30 ` supriya kannery
2011-11-22 9:45 ` supriya kannery
2011-11-22 11:32 ` Stefan Hajnoczi
2011-11-22 11:30 ` supriya kannery
2011-11-22 11:54 ` Stefan Hajnoczi
2011-11-22 11:36 ` Christoph Hellwig
2012-01-19 16:24 ` [Qemu-devel] [v9 Patch 0/6]Qemu: Host pagecache setting from cmdline and monitor Kevin Wolf
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=4ECA43F2.7080506@in.ibm.com \
--to=supriyak@in.ibm.com \
--cc=hch@lst.de \
--cc=kwolf@redhat.com \
--cc=lcapitulino@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@gmail.com \
--cc=supriyak@linux.vnet.ibm.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).