From: Wenchao Xia <xiawenc@linux.vnet.ibm.com>
To: Blue Swirl <blauwirbel@gmail.com>
Cc: stefanha@gmail.com, aliguori@us.ibm.com, qemu-devel@nongnu.org,
pbonzini@redhat.com
Subject: Re: [Qemu-devel] [qemu-devel] [PATCH V2 0/3] [RFC] libqblock draft code v2
Date: Fri, 10 Aug 2012 16:04:38 +0800 [thread overview]
Message-ID: <5024C096.9040007@linux.vnet.ibm.com> (raw)
In-Reply-To: <CAAu8pHsJeuTCt3FzvTr_udktDa9ji1ze176++X1mUSMdt4AzBA@mail.gmail.com>
Thanks for your review, sorry I have forgot some fixing you
mentioned before, will correct them this time.
于 2012-8-10 1:12, Blue Swirl 写道:
> On Thu, Aug 9, 2012 at 10:10 AM, Wenchao Xia <xiawenc@linux.vnet.ibm.com> wrote:
>> This patch intrudce libqblock API, libqblock-test is used as a test case.
>>
>> V2:
>> Using struct_size and [xxx]_new [xxx]_free to keep ABI.
>> Using embbed structure to class format options in image creating.
>> Format specific options were brought to surface.
>> Image format was changed to enum interger instead of string.
>> Some API were renamed.
>> Internel error with errno was saved and with an API caller can get it.
>> ALL flags used were defined in libqblock.h.
>>
>> Something need discuss:
>> Embbed structure or union could make the model more friendly, but that
>> make ABI more difficult, because we need to check every embbed structure's
>> size and guess compiler's memory arrangement. This means #pragma pack(4)
>> or struct_size, offset_next in every structure. Any better way to solve it?
>> or make every structure a plain one?
>
> I'd still use accessor functions instead of structure passing, it
> would avoid these problems.
>
Do you mean some function like :
CreateOption_Filename_Set(const char *filename)
CreateOption_Format_Set(const char *filename)
It can solve the issue, with a cost of more small APIs in header
files that user should use. Not sure if there is a good way to make
it more friendly as an object language:
"oc.filename = name;" automatically call CreateOption_Filename_Set,
API CreateOption_Filename_Set is invisible to user.
> Packing can even introduce a new set of problems since we don't
> control the CFLAGS of the client of the library.
indeed, I tried to handle the difference in function qboo_adjust_o2n,
found that structure member alignment is hard to deal.
>> AIO is missing, need a prototype.
>>
>> Wenchao Xia (3):
>> adding libqblock
>> libqblock API
>> libqblock test case
>>
>> Makefile | 3 +
>> block.c | 2 +-
>> block.h | 1 +
>> libqblock-test.c | 197 ++++++++++++++++
>> libqblock.c | 670 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
>> libqblock.h | 447 ++++++++++++++++++++++++++++++++++++
>> 6 files changed, 1319 insertions(+), 1 deletions(-)
>> create mode 100644 libqblock-test.c
>> create mode 100644 libqblock.c
>> create mode 100644 libqblock.h
>>
>>
>>
>
--
Best Regards
Wenchao Xia
next prev parent reply other threads:[~2012-08-10 8:05 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-08-09 10:10 [Qemu-devel] [qemu-devel] [PATCH V2 0/3] [RFC] libqblock draft code v2 Wenchao Xia
2012-08-09 17:12 ` Blue Swirl
2012-08-10 8:04 ` Wenchao Xia [this message]
2012-08-11 12:18 ` Blue Swirl
2012-08-13 11:27 ` Wenchao Xia
2012-08-13 20:00 ` Blue Swirl
2012-08-14 8:52 ` Wenchao Xia
2012-08-14 18:25 ` Blue Swirl
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=5024C096.9040007@linux.vnet.ibm.com \
--to=xiawenc@linux.vnet.ibm.com \
--cc=aliguori@us.ibm.com \
--cc=blauwirbel@gmail.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@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).