[LLVMdev] Issue with std::call_once in PPC64 platform

[Hal Finkel]

----- Original Message -----
> From: "Jiangning Liu" <liujiangning1 at gmail.com>
> To: "Chris Bieneman" <beanz at apple.com>
> Cc: "Bill Schmidt" <wschmidt at linux.vnet.ibm.com>, "Samuel F Antao" <sfantao at us.ibm.com>, azanella at linux.vnet.ibm.com,
> "Hal Finkel" <hfinkel at anl.gov>, "LLVM Dev" <llvmdev at cs.uiuc.edu>, "ulrich weigand" <ulrich.weigand at de.ibm.com>
> Sent: Tuesday, November 4, 2014 6:56:07 PM
> Subject: Re: Issue with std::call_once in PPC64 platform
> 
> It seems the crash of llvm/clang build on aarch64 Debian has been
> fixed by r220941.

That's scary. Is there any actual code-generation change from that commit?

 -Hal

> 
> 
> Thanks,
> -Jiangning
> 
> 
> 
> 
> 2014-11-05 8:45 GMT+08:00 Jiangning Liu < liujiangning1 at gmail.com > :
> 
> 
> 
> The versions I'm using right now are
> 
> 
> * gcc: (Debian/Linaro 4.9.1-14) 4.9.1
> * libstdc++: libstdc++.so.6.0.20
> 
> 
> Thanks,
> -Jiangning
> 
> 
> 
> 
> 2014-11-05 4:46 GMT+08:00 Chris Bieneman < beanz at apple.com > :
> 
> 
> Adding Jiangning Liu to the thread.
> 
> Jiangning reported a similar issue on the llvm-commits list on Debian
> aarch64.
> 
> In general it sounds like std::call_once may not really be bug free.
> 
> Jiangning, can you please provide your gcc/libstdc++ version?
> 
> Thanks,
> -Chris
> 
> 
> 
> 
> > On Nov 4, 2014, at 9:38 AM, Bill Schmidt <
> > wschmidt at linux.vnet.ibm.com > wrote:
> > 
> > On Tue, 2014-11-04 at 12:17 -0500, Samuel F Antao wrote:
> >> Ok, I'll put a patch together to fix this later today. I'll
> >> probably
> >> do what Reid was suggesting and use what is already in there for
> >> Windows.
> > 
> > Hm, better hold off on that. Ulrich Weigand has noted that there is
> > invalid code generation occurring around a TLS call. Looks like is
> > is
> > probably a PPC back end problem at the moment.
> > 
> > Thanks,
> > Bill
> > 
> >> 
> >> Thanks,
> >> Samuel
> >> 
> >> Bill Schmidt < wschmidt at linux.vnet.ibm.com > wrote on 11/04/2014
> >> 12:11:08 PM:
> >> 
> >>> From: Bill Schmidt < wschmidt at linux.vnet.ibm.com >
> >>> To: Samuel F Antao/Watson/IBM at IBMUS
> >>> Cc: azanella at linux.vnet.ibm.com , beanz at apple.com , Hal Finkel
> >>> < hfinkel at anl.gov >, LLVM Dev < llvmdev at cs.uiuc.edu >
> >>> Date: 11/04/2014 12:11 PM
> >>> Subject: Re: Issue with std::call_once in PPC64 platform
> >>> 
> >>> On Tue, 2014-11-04 at 11:59 -0500, Samuel F Antao wrote:
> >>>> Hi Bill,
> >>>> 
> >>>> You can find the same issue in the buildbot:
> >>>> 
> >>>> http://lab.llvm.org:8011/builders/clang-ppc64-elf-linux2/builds/
> >>> 16444/steps/compile.llvm.stage2/logs/stdio
> >>>> 
> >>>> It is failing for me both in BE (gcc 4.8.2) and LE(4.9.1). I am
> >>>> compiling with clang 3.5, but those are the gcc toolchains I am
> >> using.
> >>>> 
> >>>> What do you think is the best way to fix this?
> >>> 
> >>> I don't know what the best fix is, but clearly a fix or revert is
> >> needed
> >>> since the patch is breaking some platforms.
> >>> 
> >>> Thanks,
> >>> Bill
> >>> 
> >>>> 
> >>>> Thanks!
> >>>> Samuel
> >>>> 
> >>>> Bill Schmidt < wschmidt at linux.vnet.ibm.com > wrote on 11/04/2014
> >>>> 11:45:06 AM:
> >>>> 
> >>>>> From: Bill Schmidt < wschmidt at linux.vnet.ibm.com >
> >>>>> To: Samuel F Antao/Watson/IBM at IBMUS
> >>>>> Cc: LLVM Dev < llvmdev at cs.uiuc.edu >, beanz at apple.com , Hal
> >>>>> Finkel
> >>>>> < hfinkel at anl.gov >, azanella at linux.vnet.ibm.com
> >>>>> Date: 11/04/2014 11:45 AM
> >>>>> Subject: Re: Issue with std::call_once in PPC64 platform
> >>>>> 
> >>>>> Hi Samuel,
> >>>>> 
> >>>>> I haven't encountered this problem. What gcc/libstdc++ are you
> >>>> using?
> >>>>> BE or LE?
> >>>>> 
> >>>>> Copying Adhemerval in case he has heard of any related issues.
> >>>>> 
> >>>>> Thanks,
> >>>>> Bill
> >>>>> 
> >>>>> On Mon, 2014-11-03 at 20:19 -0500, Samuel F Antao wrote:
> >>>>>> Hi all,
> >>>>>> 
> >>>>>> I observe that r220932 (Removing the static initializer in
> >>>>>> ManagedStatic.cpp by using llvm_call_once to initialize the
> >>>>>> ManagedStatic mutex.) is causing tablegen to segfault in PPC
> >>>> platforms
> >>>>>> during static initialization. The crash happens while calling
> >>>>>> std::call_once introduced by this patch in the wrapper used
> >> in
> >>>>>> getManagedStaticMutex.
> >>>>>> 
> >>>>>> I understand this call is buggy for some platforms (or
> >> probably
> >>>> some
> >>>>>> oldish libstdc++ versions). Is this a known issue? Should we
> >> guard
> >>>>>> LLVM_DEFINE_ONCE_FLAG definition with platform specific macros
> >> to
> >>>>>> avoid the crash, at least for the moment?
> >>>>>> 
> >>>>>> Thanks!
> >>>>>> Samuel
> >>>>>> 
> >>>>>> 
> >>>>> 
> >>>>> 
> >>>> 
> >>>> 
> >>> 
> >>> 
> >> 
> >> 
> > 
> > 
> 
> 
> 
> 

-- 
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory