From: "Alex Bennée" <alex.bennee@linaro.org>
To: Thomas Huth <thuth@redhat.com>
Cc: Mahmoud Mandour <ma.mandourr@gmail.com>, qemu-devel@nongnu.org
Subject: Re: [PATCH 5/8] util/compatfd.c: Replaced a malloc with GLib's variant
Date: Mon, 15 Mar 2021 16:15:54 +0000 [thread overview]
Message-ID: <87v99s8k3p.fsf@linaro.org> (raw)
In-Reply-To: <3bff00dd-34e7-10bb-ee74-45a7856be030@redhat.com>
Thomas Huth <thuth@redhat.com> writes:
> On 15/03/2021 07.43, Mahmoud Mandour wrote:
>> If it's unrelated, then maybe better do it in a separate patch.
>> I thought so but I didn't know whether it was a so-small change
>> that it didn't require its own patch or not. I will amend that.
>> Since this is only a very small allocation, I think it would be
>> better to
>> use g_malloc() here and then simply remove the "if (info == NULL) ..." part.
>> I was thinking of always maintaining the semantics of the existing
>> code and since g_malloc() does not behave like malloc() on
>> error, I refrained from using g_malloc() anywhere, but of course
>> I'll do it since it's the better thing to do.
>
> Keeping the semantics is normally a good idea, but the common sense in
> the QEMU project is to rather use g_malloc() for small allocations (if
> allocating some few bytes already fails, then the system is pretty
> much dead anyway), and only g_try_malloc() for huge allocations that
> really might fail on a healthy system, too.
>
> We should likely add some text to our coding style document to make
> this more obvious...
So while there are some places where we may try to dynamically scale the
memory we allocate on failure of a large allocation generally memory
allocation failure is considered fatal (ergo g_malloc, no NULL check).
However some care has to be taken depending on where we are - for
example calling abort() because something the guest did triggered us to
try an allocate more memory than we could is a no no.
We could certainly be clearer in style.rst though.
>> I will split the patches to a two-patch series regarding the
>> util/compactfd.c file (one for the style change and one for
>> changing the malloc() call into g_malloc()) and send them
>> again, is that ok?
>
> Sounds good, thanks!
>
> Thomas
--
Alex Bennée
next prev parent reply other threads:[~2021-03-15 16:21 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-14 3:23 [PATCH 0/8] Replacing malloc and the like with GLib's variants Mahmoud Mandour
2021-03-14 3:23 ` [PATCH 1/8] bsd-user/elfload.c: Replaced calls to malloc/free with GLib variants Mahmoud Mandour
2021-03-15 16:07 ` Alex Bennée
2021-03-15 16:22 ` Mahmoud Mandour
2021-03-15 17:56 ` Alex Bennée
2021-03-14 3:23 ` [PATCH 2/8] hw/audio/fmopl.c: Fixing some style errors Mahmoud Mandour
2021-03-14 3:23 ` [PATCH 3/8] hw/audio/fmopl.c: Replaced calls to malloc with GLib's variants Mahmoud Mandour
2021-03-15 16:12 ` Alex Bennée
2021-03-15 17:35 ` Mahmoud Mandour
2021-03-14 3:23 ` [PATCH 4/8] target/xtensa: Replaced malloc/free " Mahmoud Mandour
2021-03-14 15:38 ` Max Filippov
2021-03-14 3:23 ` [PATCH 5/8] util/compatfd.c: Replaced a malloc with GLib's variant Mahmoud Mandour
2021-03-15 6:10 ` Thomas Huth
2021-03-15 6:43 ` Mahmoud Mandour
2021-03-15 7:29 ` Thomas Huth
2021-03-15 16:15 ` Alex Bennée [this message]
2021-03-14 3:23 ` [PATCH 6/8] tools/virtiofsd/buffer.c: replaced a calloc call with GLib's g_try_new0 Mahmoud Mandour
2021-03-15 9:53 ` Stefan Hajnoczi
2021-05-25 19:01 ` Dr. David Alan Gilbert
2021-03-14 3:23 ` [PATCH 7/8] tools/virtiofsd/fuse_opt.c: Replaced a malloc with GLib's g_try_malloc Mahmoud Mandour
2021-03-15 9:54 ` Stefan Hajnoczi
2021-03-14 3:23 ` [PATCH 8/8] tools/virtiofsd: Replacing malloc-like calls with GLib's variants Mahmoud Mandour
2021-03-15 10:01 ` Stefan Hajnoczi
2021-03-15 10:46 ` Mahmoud Mandour
2021-03-16 17:23 ` 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=87v99s8k3p.fsf@linaro.org \
--to=alex.bennee@linaro.org \
--cc=ma.mandourr@gmail.com \
--cc=qemu-devel@nongnu.org \
--cc=thuth@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).