diff --git a/kern/libkern/string.c b/kern/libkern/string.c
new file mode 100644
index 0000000..99a0ea2
--- /dev/null
+++ b/kern/libkern/string.c
@@ -0,0 +1,127 @@
+#include <string.h>
+
+void *memcpy(void *s1, const void *s2, size_t n) {
+    const char *f = s2;
+    char       *t = s1;
+
+    while (n-- > 0) *t++ = *f++;
+    return s1;
+}
+
+void *memmove(void *s1, const void *s2, size_t n) {
+    const char *f = s2;
+    char       *t = s1;
+
+    if (f < t) {
+        f += n;
+        t += n;
+        while (n-- > 0) *--t = *--f;
+    } else
+        while (n-- > 0) *t++ = *f++;
+    return s1;
+}
+
+void *memset(void *dest, int c, size_t n) {
+    unsigned char *s = dest;
+    size_t         k;
+
+    /* Fill head and tail with minimal branching. Each
+     * conditional ensures that all the subsequently used
+     * offsets are well-defined and in the dest region. */
+
+    if (!n)
+        return dest;
+    s[0] = c;
+    s[n - 1] = c;
+    if (n <= 2)
+        return dest;
+    s[1] = c;
+    s[2] = c;
+    s[n - 2] = c;
+    s[n - 3] = c;
+    if (n <= 6)
+        return dest;
+    s[3] = c;
+    s[n - 4] = c;
+    if (n <= 8)
+        return dest;
+
+    /* Advance pointer to align it at a 4-byte boundary,
+     * and truncate n to a multiple of 4. The previous code
+     * already took care of any head/tail that get cut off
+     * by the alignment. */
+
+    k = -(uintptr_t)s & 3;
+    s += k;
+    n -= k;
+    n &= -4;
+
+#ifdef __GNUC__
+    typedef uint32_t __attribute__((__may_alias__)) u32;
+    typedef uint64_t __attribute__((__may_alias__)) u64;
+
+    u32 c32 = ((u32)-1) / 255 * (unsigned char)c;
+
+    /* In preparation to copy 32 bytes at a time, aligned on
+     * an 8-byte bounary, fill head/tail up to 28 bytes each.
+     * As in the initial byte-based head/tail fill, each
+     * conditional below ensures that the subsequent offsets
+     * are valid (e.g. !(n<=24) implies n>=28). */
+
+    *(u32 *)(s + 0) = c32;
+    *(u32 *)(s + n - 4) = c32;
+    if (n <= 8)
+        return dest;
+    *(u32 *)(s + 4) = c32;
+    *(u32 *)(s + 8) = c32;
+    *(u32 *)(s + n - 12) = c32;
+    *(u32 *)(s + n - 8) = c32;
+    if (n <= 24)
+        return dest;
+    *(u32 *)(s + 12) = c32;
+    *(u32 *)(s + 16) = c32;
+    *(u32 *)(s + 20) = c32;
+    *(u32 *)(s + 24) = c32;
+    *(u32 *)(s + n - 28) = c32;
+    *(u32 *)(s + n - 24) = c32;
+    *(u32 *)(s + n - 20) = c32;
+    *(u32 *)(s + n - 16) = c32;
+
+    /* Align to a multiple of 8 so we can fill 64 bits at a time,
+     * and avoid writing the same bytes twice as much as is
+     * practical without introducing additional branching. */
+
+    k = 24 + ((uintptr_t)s & 4);
+    s += k;
+    n -= k;
+
+    /* If this loop is reached, 28 tail bytes have already been
+     * filled, so any remainder when n drops below 32 can be
+     * safely ignored. */
+
+    u64 c64 = c32 | ((u64)c32 << 32);
+    for (; n >= 32; n -= 32, s += 32) {
+        *(u64 *)(s + 0) = c64;
+        *(u64 *)(s + 8) = c64;
+        *(u64 *)(s + 16) = c64;
+        *(u64 *)(s + 24) = c64;
+    }
+#else
+    /* Pure C fallback with no aliasing violations. */
+    for (; n; n--, s++) *s = c;
+#endif
+
+    return dest;
+}
+
+int memcmp(const void *s1, const void *s2, size_t n) {
+    if (n != 0) {
+        const unsigned char *p1 = s1, *p2 = s2;
+
+        do {
+            if (*p1++ != *p2++)
+                return (*--p1 - *--p2);
+        } while (--n != 0);
+    }
+    return (0);
+}
diff --git a/kern/libkern/string.h b/kern/libkern/string.h
new file mode 100644
index 0000000..01d3853
--- /dev/null
+++ b/kern/libkern/string.h
@@ -0,0 +1,7 @@
+#pragma once
+#include <types.h>
+
+void *memcpy(void *s1, const void *s2, size_t n);
+void *memmove(void *s1, const void *s2, size_t n);
+void *memset(void *dest, int c, size_t n);
+int   memcmp(const void *s1, const void *s2, size_t n);