Re: [linux-stable-rc:linux-4.19.y 9916/9999] drivers/pinctrl/pinctrl-ingenic.o: warning: objtool: ingenic_pinconf_group_get()+0x12: unsupported stack pointer realignment

[Philip Li <philip.li@intel.com>]

[-- Attachment #1: Type: text/plain, Size: 3367 bytes --]

On Tue, Apr 07, 2020 at 04:34:12PM -0700, Nick Desaulniers wrote:
> On Tue, Apr 7, 2020 at 4:31 PM Philip Li <philip.li@intel.com> wrote:
> >
> > On Tue, Apr 07, 2020 at 10:51:38AM -0700, Nick Desaulniers wrote:
> > > On Tue, Apr 7, 2020 at 9:11 AM kbuild test robot <lkp@intel.com> wrote:
> > > >
> > > > Hi Dirk,
> > > >
> > > > First bad commit (maybe != root cause):
> > > >
> > > > tree:   https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-4.19.y
> > > > head:   6e19c6f2d2653aebd7d307ec43293f579e21604f
> > > > commit: 621f2ded601546119fabccd1651b1ae29d26cd38 [9916/9999] scripts/dtc: Remove redundant YYLOC global declaration
> > > > config: x86_64-randconfig-d003-20200407 (attached as .config)
> > > > compiler: clang version 11.0.0 (https://github.com/llvm/llvm-project 3185881d69022e03b300e189838b0599ed417be5)
> > > > reproduce:
> > > >         wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
> > > >         chmod +x ~/bin/make.cross
> > > >         git checkout 621f2ded601546119fabccd1651b1ae29d26cd38
> > > >         # save the attached .config to linux build tree
> > > >         COMPILER=clang make.cross ARCH=x86_64
> > > >
> > > > If you fix the issue, kindly add following tag as appropriate
> > > > Reported-by: kbuild test robot <lkp@intel.com>
> > > >
> > > > All warnings (new ones prefixed by >>):
> > > >
> > > > >> drivers/pinctrl/pinctrl-ingenic.o: warning: objtool: ingenic_pinconf_group_get()+0x12: unsupported stack pointer realignment
> > > > >> drivers/pinctrl/pinctrl-ingenic.o: warning: objtool: ingenic_pinconf_group_get()+0x0: stack state mismatch: cfa1=7+136 cfa2=7+8
> > >
> > > Apologies that the bot keeps converging on this patch; it's not the
> > > source of the regression.  The build was red with ToT LLVM before this
> > > patch landed; after the patch landed this warning seemed new, so the
> > > bot converged its bisection on this commit.
> > Sorry for the false positive bisection, we will think of this further.
> 
> > Hi Nick, do you think the bot need blacklist this or send similar objtool
> > to clang mailing list only?
> 
> Maybe for now with the goal to eventually revert, if it's easy for you
> to do so, it might be a good idea.  In general, developers should pay
> attention to new objtool warnings they introduce, but so far all
> reports from objtool with clang builds have been clang bugs in my
> experience.  Once we're in better shape, then maybe we can remove the
> diversion of the failure reports?
got it, thanks for advice, we will update bot logic to send objtool
related report to clang list only.

> 
> >
> > >
> > > Also, we recently turned on Clang builds for 0day bot, so it's digging
> > > up interesting bugs via randconfig testing.  This is useful and we
> > > appreciate the reports, especially ones from objtool frequently find
> > > compiler bugs.
> > >
> > > I've filed https://github.com/ClangBuiltLinux/linux/issues/977 for me
> > > to follow up on.
> > >
> > > --
> > > Thanks,
> > > ~Nick Desaulniers
> > > _______________________________________________
> > > kbuild-all mailing list -- kbuild-all(a)lists.01.org
> > > To unsubscribe send an email to kbuild-all-leave(a)lists.01.org
> 
> 
> 
> -- 
> Thanks,
> ~Nick Desaulniers