qemu-devel
[Top][All Lists]
Advanced

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

Re: [Qemu-devel] [PATCH] Fix error checking and cleaning in path.c V2


From: Jean-Christophe Dubois
Subject: Re: [Qemu-devel] [PATCH] Fix error checking and cleaning in path.c V2
Date: Fri, 4 Sep 2009 21:15:30 +0200
User-agent: KMail/1.11.2 (Linux/2.6.28-15-generic; KDE/4.2.2; i686; ; )

On Friday 04 September 2009 19:54:31 Anthony Liguori wrote:
> Jean-Christophe DUBOIS wrote:
> > Very little error checking was done in path.c and we were
> > leaking some memories in case of error.
> >
> > This patch rely on the abort() behavior of qemu_malloc and
> > friends
> >
> > Signed-off-by: Jean-Christophe Dubois <address@hidden>
>
> This breaks the linux-user build.

Strange !!! On my system, it doesn't break the build (everything is compiling 
fine) but the linux-user binaries are indeed wrong as they are missing the 
qemu_strdup() function ... and they fail to run ...

How could the link succeed ????

I'll look into it.

However, note that this is related to a question I asked on this list back in 
june. Why are xxx-user (in xxx-user/mmap.c files) implemetation of qemu_malloc 
and friends not on par with the main qemu implementation? They are not the 
same on a content point of view (qemu_stdup and qemu_strndup are missing) and 
on a behavior point of view (they don't abort on failure).

A lot of code in qemu is now assuming qemu_malloc will abort on error but this 
is not the case for the xxx-user build. Is this perceived as a problem?

JC

>
> Regards,
>
> Anthony Liguori






reply via email to

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