Re: [Qemu-devel] [PATCH] linux-user: Fix possible realloc memory leak

[Peter Maydell <peter.maydell@linaro.org>]

On 18 January 2011 17:51, Markus Armbruster <armbru@redhat.com> wrote:
> Stefan Weil <weil@mail.berlios.de> writes:
>> Am 18.01.2011 09:26, schrieb Markus Armbruster:
>>> Stefan Weil <weil@mail.berlios.de> writes:
>>>> Extract from "man realloc":
>>>> "If realloc() fails the original block is left untouched;
>>>> it is not freed or moved."

>>> Sidestep the problem via qemu_realloc() instead?
>>
>> The same change was applied to bsd-user/elfload.c.
>>
>> As symbol loading is not essential in most applications,
>> returning after out-of-memory should be better than
>> aborting (that's what qemu_realloc does).
>
> Unless the requested size is *really* large, I'd expect this to stave
> off the out-of-memory failure for a few microseconds at best.

Yeah, but the patch is OK, it fixes an actual bug and
it does so in line with the malloc-failure handling of
the rest of the function. It doesn't seem to me to
be important enough an issue to worry about one way
or the other.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

-- PMM