Re: [stable:linux-4.14.y 4459/9999] lib/mpi/mpih-div.c:142:20: error: invalid use of a cast in a inline asm context requiring an lvalue: remove the cast or build with -fheinous-gnu-extensions

["Liu, Yujie" <yujie.liu@intel.com>]

On Fri, 2023-08-04 at 08:41 -0700, Nick Desaulniers wrote:
> On Fri, Aug 4, 2023 at 8:37 AM Nathan Chancellor <nathan@kernel.org>
> wrote:
> > 
> > On Fri, Aug 04, 2023 at 05:32:55AM +0800, kernel test robot wrote:
> > > Hi Fangrui,
> > > 
> > > First bad commit (maybe != root cause):
> > > 
> > > tree:  
> > > https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git
> > >  linux-4.14.y
> > > head:   60a6e3043cc8b918c989707a5eba5fd6830a08a4
> > > commit: 71749b1181eca12af96654a781cabcd058c37c8f [4459/9999]
> > > Makefile: Fix GCC_TOOLCHAIN_DIR prefix for Clang cross
> > > compilation
> > > config: mips-randconfig-r035-20230731
> > > (https://download.01.org/0day-ci/archive/20230804/202308040553.nM
> > > R9oB8v-lkp@intel.com/config)
> > > compiler: clang version 17.0.0
> > > (https://github.com/llvm/llvm-project.git 4a5ac14ee968ff0ad5d2cc1
> > > ffa0299048db4c88a)
> > > reproduce:
> > > (https://download.01.org/0day-ci/archive/20230804/202308040553.nM
> > > R9oB8v-lkp@intel.com/reproduce)
> > > 
> > > If you fix the issue in a separate patch/commit (i.e. not just a
> > > new version of
> > > the same patch/commit), kindly add following tags
> > > > Reported-by: kernel test robot <lkp@intel.com>
> > > > Closes:
> > > > https://lore.kernel.org/oe-kbuild-all/202308040553.nMR9oB8v-lkp@intel.com/
> > > 
> > > All errors (new ones prefixed by >>):
> > > 
> > > > > lib/mpi/mpih-div.c:142:20: error: invalid use of a cast in a
> > > > > inline asm context requiring an lvalue: remove the cast or
> > > > > build with -fheinous-gnu-extensions
> > >      142 |                                         umul_ppmm(n1,
> > > n0, d0, q);
> > >          |                                        
> > > ~~~~~~~~~~~~~~^~~~~~~~~~
> > >    lib/mpi/longlong.h:652:20: note: expanded from macro
> > > 'umul_ppmm'
> > >      652 |         : "=l" ((USItype)(w0)), \
> > >          |                 ~~~~~~~~~~^~~
> > 
> > This is the same as [1] but the upstream solution (b0c091ae04f6)
> > could
> > cause breakage in 4.14 for compilers older than GCC 4.4; 4.14
> > claims to
> > support GCC 3.2 as a minimum version. An alternative solution if we
> > care
> > about fixing this issue in 4.14 without any regression possiblity
> > would
> > be something like:
> 
> In our own CI, we don't test mips until 5.4.
> https://github.com/ClangBuiltLinux/continuous-integration2/blob/main/generator.yml#L1691-L1698
> 
> Another solution is perhaps to ask kindly that the kernel test robot
> team not test branches of stable w/ clang for ARCH=mips.

Thanks for this info. We've configured the kernel test robot to do so.

--
Best Regards,
Yujie

> 
> > 
> > diff --git a/lib/mpi/longlong.h b/lib/mpi/longlong.h
> > index 6c5229f98c9e..42076c124e23 100644
> > --- a/lib/mpi/longlong.h
> > +++ b/lib/mpi/longlong.h
> > @@ -639,7 +639,7 @@ do { \
> >         **************  MIPS  *****************
> >         ***************************************/
> >  #if defined(__mips__) && W_TYPE_SIZE == 32
> > -#if (__GNUC__ >= 5) || (__GNUC__ >= 4 && __GNUC_MINOR__ >= 4)
> > +#if (__GNUC__ >= 5) || (__GNUC__ >= 4 && __GNUC_MINOR__ >= 4) ||
> > defined(__clang__)
> >  #define umul_ppmm(w1, w0, u, v)                        \
> >  do {                                           \
> >         UDItype __ll = (UDItype)(u) * (v);      \
> > 
> > [1]:
> > https://lore.kernel.org/20230801202602.GA2589710@dev-arch.thelio-3990X/
> > 
> > Cheers,
> > Nathan
> > 
> 
>