[llvm] r320404 - [Hexagon] Add support for Hexagon V65
Krzysztof Parzyszek via llvm-commits
llvm-commits at lists.llvm.org
Tue Mar 20 12:26:37 PDT 2018
I committed r328021, I hope it fixes all of them, but I'll do a clean
build later today.
-Krzysztof
On 3/20/2018 2:10 PM, Eric Christopher wrote:
> It appears to have shown up recently. I've also pointed Richard at it.
>
> I just did a clean release build and a clean release+asserts build of
> llvm and it showed up.
>
> -eric
>
> On Tue, Mar 20, 2018 at 12:07 PM Krzysztof Parzyszek
> <kparzysz at codeaurora.org <mailto:kparzysz at codeaurora.org>> wrote:
>
> This is a surprise to me. Are you seeing it on some buildbot as well?
>
> In any case, I'll take a look.
>
> -Krzysztof
>
> On 3/20/2018 1:56 PM, Eric Christopher wrote:
> > I'm sure.
> >
> > That said, you've got some warnings in the code with a ToT clang now:
> >
> > [1961/3415] Building CXX object
> >
> lib/Target/Hexagon/MCTargetDesc/CMakeFiles/LLVMHexagonDesc.dir/HexagonMCDuplexInfo.cpp.o
> >
> /usr/local/google/home/echristo/sources/llvm/lib/Target/Hexagon/MCTargetDesc/HexagonMCDuplexInfo.cpp:
> > In function ‘llvm::MCInst
> llvm::HexagonMCInstrInfo::deriveSubInst(const
> > llvm::MCInst&)’:
> >
> /usr/local/google/home/echristo/sources/llvm/lib/Target/Hexagon/MCTargetDesc/HexagonMCDuplexInfo.cpp:786:5:
> > warning: this statement may fall through [-Wimplicit-fallthrough=]
> > if (Value == 2) {
> > ^~
> >
> /usr/local/google/home/echristo/sources/llvm/lib/Target/Hexagon/MCTargetDesc/HexagonMCDuplexInfo.cpp:792:3:
> > note: here
> > case Hexagon::A4_combineir:
> > ^~~~
> >
> /usr/local/google/home/echristo/sources/llvm/lib/Target/Hexagon/MCTargetDesc/HexagonMCDuplexInfo.cpp:898:12:
> > warning: this statement may fall through [-Wimplicit-fallthrough=]
> > } else if (Value == 1) {
> > ^~
> >
> /usr/local/google/home/echristo/sources/llvm/lib/Target/Hexagon/MCTargetDesc/HexagonMCDuplexInfo.cpp:904:3:
> > note: here
> > case Hexagon::S2_storerb_io:
> > ^~~~
> >
> /usr/local/google/home/echristo/sources/llvm/lib/Target/Hexagon/MCTargetDesc/HexagonMCDuplexInfo.cpp:934:12:
> > warning: this statement may fall through [-Wimplicit-fallthrough=]
> > } else if (Inst.getOperand(0).getReg() == Hexagon::R29) {
> > ^~
> >
> /usr/local/google/home/echristo/sources/llvm/lib/Target/Hexagon/MCTargetDesc/HexagonMCDuplexInfo.cpp:940:3:
> > note: here
> > case Hexagon::S2_storeri_io:
> > ^~~~
> > [2042/3415] Building CXX object
> >
> lib/Target/Hexagon/AsmParser/CMakeFiles/LLVMHexagonAsmParser.dir/HexagonAsmParser.cpp.o
> > In file included from
> >
> /usr/local/google/home/echristo/sources/llvm/include/llvm/ADT/StringRef.h:13:0,
> > from
> >
> /usr/local/google/home/echristo/sources/llvm/include/llvm/ADT/StringMap.h:17,
> > from
> >
> /usr/local/google/home/echristo/sources/llvm/include/llvm/Support/Host.h:17,
> > from
> >
> /usr/local/google/home/echristo/sources/llvm/include/llvm/ADT/Hashing.h:49,
> > from
> >
> /usr/local/google/home/echristo/sources/llvm/include/llvm/ADT/ArrayRef.h:13,
> > from
> >
> /usr/local/google/home/echristo/sources/llvm/include/llvm/Support/CommandLine.h:23,
> > from
> >
> /usr/local/google/home/echristo/sources/llvm/lib/Target/Hexagon/MCTargetDesc/HexagonMCTargetDesc.h:17,
> > from
> >
> /usr/local/google/home/echristo/sources/llvm/lib/Target/Hexagon/Hexagon.h:45,
> > from
> >
> /usr/local/google/home/echristo/sources/llvm/lib/Target/Hexagon/AsmParser/HexagonAsmParser.cpp:12:
> >
> /usr/local/google/home/echristo/sources/llvm/include/llvm/ADT/STLExtras.h:886:6:
> > warning: mangled name for ‘decltype (llvm::adl_begin(Range))
> > llvm::remove_if(R&&, UnaryPredicate) [with R =
> > std::__cxx11::basic_string<char>&; UnaryPredicate = int (*)(int)
> throw
> > ()]’ will change in C++17 because the exception specification is
> part of
> > a function type [-Wnoexcept-type]
> > auto remove_if(R &&Range, UnaryPredicate P) ->
> > decltype(adl_begin(Range)) {
> > ^~~~~~~~~
> >
> > I looked at trying to fix, but I don't know how your processor
> works. :)
> >
> > -eric
> >
> > On Tue, Mar 20, 2018 at 11:53 AM Krzysztof Parzyszek
> > <kparzysz at codeaurora.org <mailto:kparzysz at codeaurora.org>
> <mailto:kparzysz at codeaurora.org <mailto:kparzysz at codeaurora.org>>>
> wrote:
> >
> > I've tried. Going from V62 to V65 was quite a large change,
> and locally
> > we had lots of interdependent patches implementing it,
> intertwined with
> > unrelated changes to the same components. In the end it saved
> me a lot
> > of time to just lump it all together.
> >
> > -Krzysztof
> >
> > On 3/20/2018 1:44 PM, Eric Christopher wrote:
> > > For the record you really should have broken this up into
> > separate commits.
> > >
> > > 69 files changed, 13364 insertions(+), 5905 deletions(-)
> > >
> > > -eric
> > >
> >
> > --
> > Qualcomm Innovation Center, Inc. is a member of Code Aurora
> Forum,
> > hosted by The Linux Foundation
> >
>
> --
> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
> hosted by The Linux Foundation
>
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
hosted by The Linux Foundation
More information about the llvm-commits
mailing list