[Triple] Add llvm::Triple::isSPARC{,32,64}

[MaskRay]

Please sync past 3a1cb362370d223e09899d234726e15b52327b0e: the patch accidentally removes isDriverKit.

[XiaodongLoong]

rebase code

[XiaodongLoong]

In D120381#3339390, @MaskRay wrote:

Please sync past 3a1cb362370d223e09899d234726e15b52327b0e: the patch accidentally removes isDriverKit.

@MaskRay Sorry, it is OK now. Thanks!

[ro]

Looking at existing cases of comparisons with Triple::sparc and Triple::sparcv9, I guess we should have isSPARC32 and isSPARC64, too, like isMIPS32/isMIPS64 and isPPC32/isPPC64. When switching to isSPARC* from existing Triple::sparc* comparisons, one will have to be extra careful if they are meant to mean SPARC in general or 32-bit SPARC in particular. little-endian SPARC is another beast altogether: I didn't even know it existed beyond the ability of SPARCV9 CPUs to partially run in LE mode, so it will be hard to tell if LE SPARC is to be included. I guess only code review and testing by interested parties can tell.

llvm/include/llvm/ADT/Triple.h

828

I'd go for isSPARC instead: SPARC is an acronym, like MIPS, although LLVM is anything but consistent about the correct capitalization.

[XiaodongLoong]

update code for review comments.
rebase code.

[XiaodongLoong]

llvm/include/llvm/ADT/Triple.h

828

Thanks for the comment, I fixed it.

[XiaodongLoong]

@ro
Since sparcv9 is Bi-endianness (https://en.wikipedia.org/wiki/Endianness#Bi-endianness), I add "sparcv9el" like on MIPS. If it is not OK, I can delete it.
Could I do the switching to isSPARC* from existing Triple::sparc* in another patch?

[brad]

In D120381#3342008, @XiaodongLoong wrote:

Since sparcv9 is Bi-endianness (https://en.wikipedia.org/wiki/Endianness#Bi-endianness), I add "sparcv9el" like on MIPS. If it is not OK, I can delete it.

I don't think this makes sense. No one runs SPARCv9 with little endian.

[XiaodongLoong]

@brad delete little endian SPARC V9

[ro]

In D120381#3342217, @brad wrote:

I don't think this makes sense. No one runs SPARCv9 with little endian.

Not only that: this isn't a full fledged little-endian mode like mipsel. See "Oracle SPARC Architecture 2015", p.20:

3.3.1.2 Addressing Conventions
[...]
The Oracle SPARC Architecture also supports little-endian byte order for data accesses only: the
address of a quadword, doubleword, word, or halfword is the address of its least significant byte.
Increasing the address means increasing the significance of the data unit being accessed.

[ro]

In D120381#3342008, @XiaodongLoong wrote:

Could I do the switching to isSPARC* from existing Triple::sparc* in another patch?

Given the evaluation and testing requirements I'd mentioned, I guess it's better to start using isSPARC* for new or modified code right now, but handle existing comparisons to Triple::sparc* separately. I don't think that's your duty, though. You've done way enough already. Thanks.

[ro]

llvm/include/llvm/ADT/Triple.h

828

I think all those clang-format reformatting suggestions should be ignored: while they might be appropriate for new code, the new methods would be formatted differently from every other one in this file.

[XiaodongLoong]

format new code

[ro]

llvm/include/llvm/ADT/Triple.h

833

That's the opposite of what I suggested: please follow existing formatting.

836

Same here.

[XiaodongLoong]

format code

[XiaodongLoong]

llvm/include/llvm/ADT/Triple.h

833

I am so sorry. I misunderstood your words. I format back.

836

Done.

[ro]

LGTM

[MaskRay]

clang-format

update subject

llvm/include/llvm/ADT/Triple.h

bool isPPC64() const {
     return getArch() == Triple::ppc64 || getArch() == Triple::ppc64le;
   }
 
   /// Tests whether the target is RISC-V (32- and 64-bit).
   bool isRISCV() const {
     return getArch() == Triple::riscv32 || getArch() == Triple::riscv64;
   }
 
+  /// Tests whether the target is Sparc.
+  bool isSparc() const {
+    return getArch() == Triple::sparc || getArch() == Triple::sparcv9 ||
+           getArch() == Triple::sparcel;
+  }
+
   /// Tests whether the target is SystemZ.
   bool isSystemZ() const {
     return getArch() == Triple::systemz;
   }
 
   /// Tests whether the target is x86 (32- or 64-bit).
   bool isX86() const {
     return getArch() == Triple::x86 || getArch() == Triple::x86_64;
   }
 
   /// Tests whether the target is VE
   bool isVE() const {