Re: [Qemu-devel] [PATCH] replace functions which are only available in glib-2.24

[Thomas Huth <thuth@redhat.com>]

On 15.05.2018 10:41, Paolo Bonzini wrote:
> On 15/05/2018 09:42, Markus Armbruster wrote:
>> Olaf Hering <olaf@aepfle.de> writes:
>>
>>> Currently the minimal supported version of glib is 2.22.
>>> Since testing is done with a glib that claims to be 2.22, but in fact
>>> has APIs from newer version of glib, this bug was not caught during
>>> submit of the patch referenced below.
>>>
>>> Replace g_realloc_n, which is available only since 2.24, with g_renew.
>>
>> We're about to bump our required GLib version to 2.34.  See
>> https://lists.gnu.org/archive/html/qemu-devel/2018-05/msg02562.html
>> Message-ID: <751a3b4f-55ef-dec0-aec9-6b48922b7843@redhat.com>
>>
>> However, the patch may make sense for stable anyway.
> 
> It's a good idea even for non-stable.  In the end, g_*alloc*_n are
> rarely needed, and almost always g_*new* is better as it provides better
> type safety.  I'm queuing the patch, and we should consider adding a
> checkpatch rule to suggest not using these functions.

Please do not queue this version of the patch - Olaf sent a v2 where he
fixed the obvious bug with "IOVAMapping *" vs. "IOVAMapping".

> Paolo
> 
>>> Fixes commit 418026ca43 ("util: Introduce vfio helpers")
>>>
>>> Signed-off-by: Olaf Hering <olaf@aepfle.de>
>>> ---
>>>
>>> This must go into stable-2.12.
>>
>> Please see docs/devel/stable-process.rst for how to get a patch into a
>> stable release.
>>
>>>  util/vfio-helpers.c | 6 ++----
>>>  1 file changed, 2 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
>>> index 006674c916..5fd4f5f212 100644
>>> --- a/util/vfio-helpers.c
>>> +++ b/util/vfio-helpers.c
>>> @@ -522,8 +522,7 @@ static IOVAMapping *qemu_vfio_add_mapping(QEMUVFIOState *s,
>>>  
>>>      assert(index >= 0);
>>>      s->nr_mappings++;
>>> -    s->mappings = g_realloc_n(s->mappings, sizeof(s->mappings[0]),
>>> -                              s->nr_mappings);
>>> +    s->mappings = g_renew(IOVAMapping *, s->mappings, s->nr_mappings);
>>
>> I'm afraid this switches from allocating a bunch of IOVAMapping to a
>> bunch of IOVAMapping *.  My gcc duly warns.

See v2 of the patch where Olaf fixed this already.

 Thomas