llvm dev - Dec 2018 - Should intrinsics llvm.eh.sjlj.setjmp be with isBarrier flag?

Hi,

I meet an issue when I verify machineinstrs for Powerpc testcases in llvm.

llc -mtriple=powerpc64-unknown-linux-gnu <
llvm/llvm/test/CodeGen/PowerPC/sj-ctr-loop.ll -verify-machineinstrs

Bad machine code: MBB exits via unconditional fall-through but ends with a
barrier instruction! ***
      function: main
      basic block: %bb.2 for.body.lr.ph (0x100275437e8)

Content in block BB.2:

BB#2: derived from LLVM BB %for.end
    Predecessors according to CFG: BB#1
        %vreg2<def> = ADDIStocHA %X2, <ga:@env_sigill>;
G8RC_and_G8RC_NOX0:%vreg2
        %vreg3<def> = LDtocL <ga:@env_sigill>, %vreg2<kill>;
mem:LD8[GOT]
G8RC:%vreg3 G8RC_and_G8RC_NOX0:%vreg2
        %vreg4<def> = EH_SjLj_SetJmp64 %vreg3<kill>,
%CTR8<imp-def,dead>;
GPRC:%vreg4 G8RC:%vreg3

Currently Powerpc sets EH_SjLj_SetJmp64 with flag isBarrier. But it is also
a fall-through instruction. So it fails in machineinstr verifying.

I checked other platforms like X86 and ARM, they also set this intrinsics
as barrier. For my understanding, a barrier instruction means control flow
can not fall through like llvm.eh.sjlj.longjmp and unconditional jumps. But
setjmp should fall through.

Is there any misunderstanding? Is there any specification reason for
setting llvm.eh.sjlj.setjmp as a barrier?

Thanks in advance.

BRS//
Chen Zheng
Power Compiler Backend Developer
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20181207/3dc0e49d/attachment.html>

On Fri, 7 Dec 2018 at 18:25, Zheng CZ Chen via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
> Is there any misunderstanding? Is there any specification reason for
setting llvm.eh.sjlj.setjmp as a barrier?
I agree with you, the setjmp ones shouldn't be isBarrier. I suspect
the original commit mistook the name "isBarrier" as referring to a
memory fence-like construct. It's a fairly common mistake and could
easily apply here.

Cheers.

Tim.

Thanks for your reply.

Patch is created https://reviews.llvm.org/D55499.

Thanks.

BRS//
Chen Zheng
Power Compiler Backend Developer




From:	Tim Northover <t.p.northover at gmail.com>
To:	czhengsz at cn.ibm.com
Cc:	LLVM Developers Mailing List <llvm-dev at lists.llvm.org>
Date:	2018/12/08 06:07 PM
Subject:	Re: [llvm-dev] Should intrinsics llvm.eh.sjlj.setjmp be with
            isBarrier flag?



On Fri, 7 Dec 2018 at 18:25, Zheng CZ Chen via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
> Is there any misunderstanding? Is there any specification reason for
setting llvm.eh.sjlj.setjmp as a barrier?

I agree with you, the setjmp ones shouldn't be isBarrier. I suspect
the original commit mistook the name "isBarrier" as referring to a
memory fence-like construct. It's a fairly common mistake and could
easily apply here.

Cheers.

Tim.



-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20181210/a741d0cd/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20181210/a741d0cd/attachment.gif>