[llvm-dev] Tail merging "undef" with a defined register: wrong code

[Quentin Colombet via llvm-dev]

Why do you say the code is wrong?

I can see why the MachineVerifier would complain, but the code seems to do the same thing as it was doing before the transformation, isn’t it?

> On Jul 27, 2017, at 11:40 AM, Krzysztof Parzyszek via llvm-dev <llvm-dev at lists.llvm.org> wrote:
> 
> The comment in test/CodeGen/X86/branchfolding-undef.mir states that such merging is legal, however doing so can actually generate wrong code:
> 
> Consider this (valid code):
> 
> ---
> name: fred
> tracksRegLiveness: true
> 
> body: |
>  bb.0:
>    successors: %bb.1, %bb.2
>      J2_jumpt undef %p0, %bb.2, implicit-def %pc
>      J2_jump %bb.1, implicit-def %pc
> 
>  bb.1:
>    successors: %bb.3
>      %r0 = L2_loadruh_io undef %r0, 0
>      PS_storerhabs 0, killed %r0
>      J2_jump %bb.3, implicit-def %pc
> 
>  bb.2:
>    successors: %bb.3
>      PS_storerhabs 0, undef %r0
>      J2_jump %bb.3, implicit-def %pc
> 
>  bb.3:
>      PS_jmpret killed %r31, implicit-def %pc
> ...
> 
> 
> Run:
> llc -march=hexagon -run-pass branch-folder test.mir -o -
> 
> 
> We get an invalid code:
> 
> body: |
>  bb.0:
>    successors: %bb.1(0x40000000), %bb.2(0x40000000)
>    J2_jumpt undef %p0, %bb.2, implicit-def %pc
> 
>  bb.1:
>    successors: %bb.2(0x80000000)
>    %r0 = L2_loadruh_io undef %r0, 0
> 
>  bb.2:
>    liveins: %r0       ; Invalid: r0 is not live out of bb.0
>    PS_storerhabs 0, %r0
>    PS_jmpret killed %r31, implicit-def %pc
> ...
> 
> 
> Should we disallow merging undef with non-undef?
> 
> -Krzysztof
> 
> 
> -- 
> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev