Re: [Qemu-devel] [PATCH V2] build: remove compile warning

[Paolo Bonzini]

Il 22/06/2013 12:03, Stefan Weil ha scritto:
> Am 18.06.2013 12:13, schrieb Paolo Bonzini:
>> Il 07/06/2013 14:17, Markus Armbruster ha scritto:
>>>> diff --git a/util/iov.c b/util/iov.c
>>>> index cc6e837..b91cfb9 100644
>>>> --- a/util/iov.c
>>>> +++ b/util/iov.c
>>>> @@ -146,7 +146,7 @@ ssize_t iov_send_recv(int sockfd, struct iovec *iov, 
>>>> unsigned iov_cnt,
>>>>  {
>>>>      ssize_t total = 0;
>>>>      ssize_t ret;
>>>> -    size_t orig_len, tail;
>>>> +    size_t orig_len = 0, tail;
>>>>      unsigned niov;
>>>>  
>>>>      while (bytes > 0) {
>>> Here are the uses of orig_len:
>>>
>>>         if (tail) {
>>>             /* second, fixup the last element, and remember the original
>>>              * length */
>>>             assert(niov < iov_cnt);
>>>             assert(iov[niov].iov_len > tail);
>>>             orig_len = iov[niov].iov_len;
>>>             iov[niov++].iov_len = tail;
>>>         }
>>>
>>>         ret = do_send_recv(sockfd, iov, niov, do_send);
>>>
>>>         /* Undo the changes above before checking for errors */
>>>         if (tail) {
>>>             iov[niov-1].iov_len = orig_len;
>>>         }
>>>
>>>
> 
> I get this warning, too, when I run a normal cross compilation with
> MinGW-w64:
> 
> util/iov.c:190:33: warning: ‘orig_len’ may be used uninitialized in this
> function [-Wuninitialized]
> 
> My build environment:
> 
> Debian wheezy with packages gcc-mingw-w64-i686, gcc-mingw-w64-x86-64
> (4.6.3-14+8).
> 
> A complete build results in 5 warnings. Here are the other 4 of them:
> 
> hw/arm/spitz.c:280:0: warning: "MOD_SHIFT" redefined [enabled by default]
> hw/ppc/spapr.c:673:26: warning: cast from pointer to integer of
> different size [-Wpointer-to-int-cast]

Isn't this one a Win64 bug?

> hw/ppc/spapr_hcall.c:188:1: warning: control reaches end of non-void
> function [-Wreturn-type]
> hw/ppc/spapr_pci.c:454:1: warning: control reaches end of non-void
> function [-Wreturn-type]

I think you could report this to mingw.  GCC should handle "if (!0)
foo()" just fine if foo is noreturn, perhaps the "assertion failure"
runtime function is not noreturn in mingw.

> I already sent a patch for the MOD_SHIFT issue.
> 
> The remaining 3 warnings are also caused by code which makes it difficult
> for the compiler to detect that it is correct.

Not as hard as this one, though.

Anyway, I would be okay a fix that makes the code easier to follow for
compilers (and doesn't have too much duplication so that humans are also
happy).  But adding "= 0" is the worst, because smart compilers will not
detect a human's mistakes anymore.

Paolo