[PATCH] eal: fix unaligned loads/stores in rte_memcpy_generic

[Luc Pelletier <lucp.at.work@gmail.com>]

Calls to rte_memcpy_generic could result in unaligned loads/stores for
1 < n < 16. This is undefined behavior according to the C standard,
and it gets flagged by the clang undefined behavior sanitizer.

rte_memcpy_generic is called with unaligned src and dst addresses.
When 1 < n < 16, the code would cast both src and dst to a qword,
dword or word pointer, without verifying the alignment of src/dst. The
code was changed to use inline assembly for the load/store operations.
Unaligned load/store operations are permitted in x64 assembly.

Fixes: d35cc1fe6a7a ("eal/x86: revert select optimized memcpy at run-time")
Cc: Xiaoyun Li <xiaoyun.li@intel.com>
Cc: stable@dpdk.org

Signed-off-by: Luc Pelletier <lucp.at.work@gmail.com>
---

I have just submitted another patch related to rte_memcpy. The changes
in the other patch are straightforward. The changes in this patch are
likely to require more discussion, so I've decided to keep both patches
completely separate. Please let me know if I should have done otherwise.
Based on the benchmarks of rte_memcpy that are in the dpdk tree, the
changes in this patch do not negatively affect performance.

 lib/eal/x86/include/rte_memcpy.h | 106 +++++++++++++------------------
 1 file changed, 44 insertions(+), 62 deletions(-)

diff --git a/lib/eal/x86/include/rte_memcpy.h b/lib/eal/x86/include/rte_memcpy.h
index 1b6c6e585f..d2a6099200 100644
--- a/lib/eal/x86/include/rte_memcpy.h
+++ b/lib/eal/x86/include/rte_memcpy.h
@@ -45,6 +45,47 @@ extern "C" {
 static __rte_always_inline void *
 rte_memcpy(void *dst, const void *src, size_t n);
 
+/**
+ * Copy bytes from one location to another,
+ * locations should not overlap.
+ * Use with unaligned src/dst, and n <= 15.
+ */
+static __rte_always_inline void *
+rte_mov15_or_less_unaligned(void *dst, const void *src, size_t n)
+{
+	void *ret = dst;
+	if (n & 0x08) {
+		asm ("movq (%[src]), %%rax\n"
+		"movq %%rax, (%[dst])"
+		:
+		: [dst] "r" (dst), [src] "r" (src)
+		: "rax", "memory");
+		src = (const uint64_t *)src + 1;
+		dst = (uint64_t *)dst + 1;
+	}
+	if (n & 0x04) {
+		asm ("movl (%[src]), %%eax\n"
+		"movl %%eax, (%[dst])"
+		:
+		: [dst] "r" (dst), [src] "r" (src)
+		: "eax", "memory");
+		src = (const uint32_t *)src + 1;
+		dst = (uint32_t *)dst + 1;
+	}
+	if (n & 0x02) {
+		asm ("movw (%[src]), %%ax\n"
+		"movw %%ax, (%[dst])"
+		:
+		: [dst] "r" (dst), [src] "r" (src)
+		: "ax", "memory");
+		src = (const uint16_t *)src + 1;
+		dst = (uint16_t *)dst + 1;
+	}
+	if (n & 0x01)
+		*(uint8_t *)dst = *(const uint8_t *)src;
+	return ret;
+}
+
 #if defined __AVX512F__ && defined RTE_MEMCPY_AVX512
 
 #define ALIGNMENT_MASK 0x3F
@@ -171,8 +212,6 @@ rte_mov512blocks(uint8_t *dst, const uint8_t *src, size_t n)
 static __rte_always_inline void *
 rte_memcpy_generic(void *dst, const void *src, size_t n)
 {
-	uintptr_t dstu = (uintptr_t)dst;
-	uintptr_t srcu = (uintptr_t)src;
 	void *ret = dst;
 	size_t dstofss;
 	size_t bits;
@@ -181,24 +220,7 @@ rte_memcpy_generic(void *dst, const void *src, size_t n)
 	 * Copy less than 16 bytes
 	 */
 	if (n < 16) {
-		if (n & 0x01) {
-			*(uint8_t *)dstu = *(const uint8_t *)srcu;
-			srcu = (uintptr_t)((const uint8_t *)srcu + 1);
-			dstu = (uintptr_t)((uint8_t *)dstu + 1);
-		}
-		if (n & 0x02) {
-			*(uint16_t *)dstu = *(const uint16_t *)srcu;
-			srcu = (uintptr_t)((const uint16_t *)srcu + 1);
-			dstu = (uintptr_t)((uint16_t *)dstu + 1);
-		}
-		if (n & 0x04) {
-			*(uint32_t *)dstu = *(const uint32_t *)srcu;
-			srcu = (uintptr_t)((const uint32_t *)srcu + 1);
-			dstu = (uintptr_t)((uint32_t *)dstu + 1);
-		}
-		if (n & 0x08)
-			*(uint64_t *)dstu = *(const uint64_t *)srcu;
-		return ret;
+		return rte_mov15_or_less_unaligned(dst, src, n);
 	}
 
 	/**
@@ -379,8 +401,6 @@ rte_mov128blocks(uint8_t *dst, const uint8_t *src, size_t n)
 static __rte_always_inline void *
 rte_memcpy_generic(void *dst, const void *src, size_t n)
 {
-	uintptr_t dstu = (uintptr_t)dst;
-	uintptr_t srcu = (uintptr_t)src;
 	void *ret = dst;
 	size_t dstofss;
 	size_t bits;
@@ -389,25 +409,7 @@ rte_memcpy_generic(void *dst, const void *src, size_t n)
 	 * Copy less than 16 bytes
 	 */
 	if (n < 16) {
-		if (n & 0x01) {
-			*(uint8_t *)dstu = *(const uint8_t *)srcu;
-			srcu = (uintptr_t)((const uint8_t *)srcu + 1);
-			dstu = (uintptr_t)((uint8_t *)dstu + 1);
-		}
-		if (n & 0x02) {
-			*(uint16_t *)dstu = *(const uint16_t *)srcu;
-			srcu = (uintptr_t)((const uint16_t *)srcu + 1);
-			dstu = (uintptr_t)((uint16_t *)dstu + 1);
-		}
-		if (n & 0x04) {
-			*(uint32_t *)dstu = *(const uint32_t *)srcu;
-			srcu = (uintptr_t)((const uint32_t *)srcu + 1);
-			dstu = (uintptr_t)((uint32_t *)dstu + 1);
-		}
-		if (n & 0x08) {
-			*(uint64_t *)dstu = *(const uint64_t *)srcu;
-		}
-		return ret;
+		return rte_mov15_or_less_unaligned(dst, src, n);
 	}
 
 	/**
@@ -672,8 +674,6 @@ static __rte_always_inline void *
 rte_memcpy_generic(void *dst, const void *src, size_t n)
 {
 	__m128i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7, xmm8;
-	uintptr_t dstu = (uintptr_t)dst;
-	uintptr_t srcu = (uintptr_t)src;
 	void *ret = dst;
 	size_t dstofss;
 	size_t srcofs;
@@ -682,25 +682,7 @@ rte_memcpy_generic(void *dst, const void *src, size_t n)
 	 * Copy less than 16 bytes
 	 */
 	if (n < 16) {
-		if (n & 0x01) {
-			*(uint8_t *)dstu = *(const uint8_t *)srcu;
-			srcu = (uintptr_t)((const uint8_t *)srcu + 1);
-			dstu = (uintptr_t)((uint8_t *)dstu + 1);
-		}
-		if (n & 0x02) {
-			*(uint16_t *)dstu = *(const uint16_t *)srcu;
-			srcu = (uintptr_t)((const uint16_t *)srcu + 1);
-			dstu = (uintptr_t)((uint16_t *)dstu + 1);
-		}
-		if (n & 0x04) {
-			*(uint32_t *)dstu = *(const uint32_t *)srcu;
-			srcu = (uintptr_t)((const uint32_t *)srcu + 1);
-			dstu = (uintptr_t)((uint32_t *)dstu + 1);
-		}
-		if (n & 0x08) {
-			*(uint64_t *)dstu = *(const uint64_t *)srcu;
-		}
-		return ret;
+		return rte_mov15_or_less_unaligned(dst, src, n);
 	}
 
 	/**
-- 
2.25.1