Support: xxHash64: Zero extend byte values loaded from the input data.

[pcc]

Summary

This matches the reference implementation:
https://github.com/Cyan4973/xxHash/blob/3064d42e7d74b0921bdd1818395d9cb37bb8976a/xxhash.c#L707
https://github.com/Cyan4973/xxHash/blob/3064d42e7d74b0921bdd1818395d9cb37bb8976a/xxhash.c#L147

and makes the behaviour independent of whether we are compiled with
-funsigned-char.

[ruiu]

Hmm, P is a pointer to a unsigned char, so it is always sign-extended, no? What am I missing?

[pcc]

http://llvm-cs.pcc.me.uk/lib/Support/xxhash.cpp#74

It is a char pointer.

[ruiu]

In your diff (it's hidden but you can reveal it by clicking show all liens), it was an unsigned char. Anyways, LGTM.

[pcc]

Oh, it looks like @MaskRay fixed this in a different way in rL338128.

[MaskRay]

In D49923#1178580, @pcc wrote:

Oh, it looks like @MaskRay fixed this in a different way in rL338128.

I conveyed the message to you in a covert channel :)

llvm/lib/Support/xxhash.cpp

uint64_t llvm::xxHash64(StringRef Data) {
 
   if (P + 4 <= BEnd) {
     H64 ^= (uint64_t)(endian::read32le(P)) * PRIME64_1;
     H64 = rotl64(H64, 23) * PRIME64_2 + PRIME64_3;
     P += 4;
   }
 
   while (P < BEnd) {
-    H64 ^= (*P) * PRIME64_5;
+    H64 ^= uint8_t(*P) * PRIME64_5;
     H64 = rotl64(H64, 11) * PRIME64_1;
     P++;
   }
 
   H64 ^= H64 >> 33;
   H64 *= PRIME64_2;
   H64 ^= H64 >> 29;
   H64 *= PRIME64_3;
   H64 ^= H64 >> 32;
 
   return H64;
 }