Re: [PATCH] migration: Silence compiler warning in global_state_store_running()

[Thomas Huth]

On 17/09/2020 09.18, Paolo Bonzini wrote:
> On 17/09/20 08:30, Thomas Huth wrote:
>> On 16/09/2020 19.19, Thomas Huth wrote:
>>> GCC 9.3.0 on Ubuntu complains:
>>>
>>> In file included from /usr/include/string.h:495,
>>>                  from /home/travis/build/huth/qemu/include/qemu/osdep.h:87,
>>>                  from ../migration/global_state.c:13:
>>> In function ‘strncpy’,
>>>     inlined from ‘global_state_store_running’ at 
>>> ../migration/global_state.c:47:5:
>>> /usr/include/x86_64-linux-gnu/bits/string_fortified.h:106:10: error: 
>>> ‘__builtin_strncpy’ specified bound 100 equals destination size 
>>> [-Werror=stringop-truncation]
>>>   106 |   return __builtin___strncpy_chk (__dest, __src, __len, __bos 
>>> (__dest));
>>>       |          
>>> ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>>
>>> ... but we apparently really want to do the strncpy here. Silence the
>>> warning with QEMU_NONSTRING.
>>>
>>> Signed-off-by: Thomas Huth <thuth@redhat.com>
>>> ---
>>>  migration/global_state.c | 4 ++--
>>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/migration/global_state.c b/migration/global_state.c
>>> index 25311479a4..f1355d7d97 100644
>>> --- a/migration/global_state.c
>>> +++ b/migration/global_state.c
>>> @@ -43,9 +43,9 @@ int global_state_store(void)
>>>  void global_state_store_running(void)
>>>  {
>>>      const char *state = RunState_str(RUN_STATE_RUNNING);
>>> +    QEMU_NONSTRING char *dest = (char *)global_state.runstate;
>>>      assert(strlen(state) < sizeof(global_state.runstate));
>>> -    strncpy((char *)global_state.runstate,
>>> -           state, sizeof(global_state.runstate));
>>> +    strncpy(dest, state, sizeof(global_state.runstate));
>>>  }
>>
>> Darn, I was sending too fast here, sorry, but seems like this does *not*
>> fix the issue with GCC 9.3:
>>
>>  https://travis-ci.com/github/huth/qemu/jobs/385871010#L2930
>>
>> ... so maybe we should simply switch to strpadcpy() instead?
> 
> Yes, and probably do so everywhere that strncpy is used.

I think the trick with QEMU_NONSTRING should work fine in all cases
where we have a real array, see e.g. buf[] in find_vdi_name() in
block/sheepdog.c. It just does not seem to work in case you have a char*
pointer instead of an array...

 Thomas