[libc] Improve memcpy copy loop

[gchatelet]

Summary

Rewriting loop so the terminating condition does not depend on the loop body

libc/src/string/memory_utils/memcpy_utils.h

 // Precondition: `count > 2 * kBlockSize` for efficiency.
 //               `count >= kBlockSize` for correctness.
 template <size_t kBlockSize>
 static void CopyAlignedBlocks(char *__restrict dst, const char *__restrict src,
                               size_t count) {
   CopyBlock<kBlockSize>(dst, src); // Copy first block
 
   // Copy aligned blocks
-  size_t offset = kBlockSize - offset_from_last_aligned<kBlockSize>(dst);
-  for (; offset + kBlockSize < count; offset += kBlockSize)
-    CopyBlock<kBlockSize>(dst + offset, src + offset);
+  const size_t ofla = offset_from_last_aligned<kBlockSize>(dst);
+  const size_t limit = count + ofla - kBlockSize;
+  for (size_t offset = kBlockSize; offset < limit; offset += kBlockSize)
+    CopyBlock<kBlockSize>(dst - ofla + offset, src - ofla + offset);
 
   CopyLastBlock<kBlockSize>(dst, src, count); // Copy last block
 }
 
 } // namespace __llvm_libc
 
 #endif //  LIBC_SRC_STRING_MEMORY_UTILS_MEMCPY_UTILS_H

libc/test/src/string/memory_utils/memcpy_utils_test.cpp

TEST(MemcpyUtilsTest, CopyBlockOverlap) {
   trace.Clear();
   CopyBlockOverlap<4>(I(2), I(1), 7);
   EXPECT_STREQ(trace.Write(), "001112111");
   EXPECT_STREQ(trace.Read(), "01112111");
 }
 
 TEST(MemcpyUtilsTest, CopyAlignedBlocks) {
   auto &trace = GetTrace();
-  // Destination is aligned already.
+  // Destination is aligned and multiple of alignment.
+  //   "1111"
+  trace.Clear();
+  CopyAlignedBlocks<4>(I(0), I(0), 4);
+  EXPECT_STREQ(trace.Write(), "2222");
+  EXPECT_STREQ(trace.Read(), "2222");
+
+  // Destination is aligned and multiple of alignment.
+  //   "11110000"
+  // + "00001111"
+  // = "11111111"
+  trace.Clear();
+  CopyAlignedBlocks<4>(I(0), I(0), 8);
+  EXPECT_STREQ(trace.Write(), "11111111");
+  EXPECT_STREQ(trace.Read(), "11111111");
+
+  // Destination is aligned already overlap at end.
   //   "1111000000000"
   // + "0000111100000"
   // + "0000000011110"
   // + "0000000001111"
   // = "1111111112221"
   trace.Clear();
   CopyAlignedBlocks<4>(I(0), I(0), 13);
   EXPECT_STREQ(trace.Write(), "1111111112221");
   EXPECT_STREQ(trace.Read(), "1111111112221");
 
-  // Misaligned destination
+  // Misaligned destination.
   //   "01111000000000"
   // + "00001111000000"
   // + "00000000111100"
   // + "00000000001111"
   // = "01112111112211"
   trace.Clear();
   CopyAlignedBlocks<4>(I(1), I(0), 13);
   EXPECT_STREQ(trace.Write(), "01112111112211");
   EXPECT_STREQ(trace.Read(), "1112111112211");
+
+  // Misaligned destination aligned at end.
+  //   "011110000000"
+  // + "000011110000"
+  // + "000000001111"
+  // = "011121111111"
+  trace.Clear();
+  CopyAlignedBlocks<4>(I(1), I(0), 11);
+  EXPECT_STREQ(trace.Write(), "011121111111");
+  EXPECT_STREQ(trace.Read(), "11121111111");
 }
 
 TEST(MemcpyUtilsTest, MaxReloads) {
   auto &trace = GetTrace();
   for (size_t alignment = 0; alignment < 32; ++alignment) {
     for (size_t count = 64; count < 768; ++count) {
       trace.Clear();
       // We should never reload more than twice when copying from count = 2x32.