summaryrefslogtreecommitdiffstats
path: root/sys/libkern
diff options
context:
space:
mode:
authored <ed@FreeBSD.org>2012-01-01 20:26:11 +0000
committered <ed@FreeBSD.org>2012-01-01 20:26:11 +0000
commita201da71788983b3f0445aa46bca32751fc7393c (patch)
treeaf6bd36dd5b255501e8018e450b5f00a332c52cb /sys/libkern
parent4a643207b240e975c07808e969d7545d1d8c0b31 (diff)
downloadFreeBSD-src-a201da71788983b3f0445aa46bca32751fc7393c.zip
FreeBSD-src-a201da71788983b3f0445aa46bca32751fc7393c.tar.gz
Introducing memcchr(3).
It seems two of the file system drivers we have in the tree, namely ufs and ext3, use a function called `skpc()'. The meaning of this function does not seem to be documented in FreeBSD, but it turns out one needs to be a VAX programmer to understand what it does. SPKC is an instruction on the VAX that does the opposite of memchr(). It searches for the non-equal character. Add a new function called memcchr() to the tree that has the following advantages over skpc(): - It has a name that makes more sense than skpc(). Just like strcspn() matches the complement of strspn(), memcchr() is the complement of memchr(). - It is faster than skpc(). Similar to our strlen() in libc, it compares entire words, instead of single bytes. It seems that for this routine this yields a sixfold performance increase on amd64. - It has a man page.
Diffstat (limited to 'sys/libkern')
-rw-r--r--sys/libkern/memcchr.c115
1 files changed, 115 insertions, 0 deletions
diff --git a/sys/libkern/memcchr.c b/sys/libkern/memcchr.c
new file mode 100644
index 0000000..eebb23a
--- /dev/null
+++ b/sys/libkern/memcchr.c
@@ -0,0 +1,115 @@
+/*-
+ * Copyright (c) 2012 Ed Schouten <ed@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/libkern.h>
+#include <sys/limits.h>
+#include <sys/param.h>
+
+/*
+ * memcchr(): find first character in buffer not matching `c'.
+ *
+ * This function performs the complement of memchr(). To provide decent
+ * performance, this function compares data from the buffer one word at
+ * a time.
+ *
+ * This code is inspired by libc's strlen(), written by Xin Li.
+ */
+
+#if LONG_BIT != 32 && LONG_BIT != 64
+#error Unsupported word size
+#endif
+
+#define LONGPTR_MASK (sizeof(long) - 1)
+
+#define TESTBYTE \
+ do { \
+ if (*p != (unsigned char)c) \
+ goto done; \
+ p++; \
+ } while (0)
+
+void *
+memcchr(const void *begin, int c, size_t n)
+{
+ const unsigned long *lp;
+ const unsigned char *p, *end;
+ unsigned long word;
+
+ /* Four or eight repetitions of `c'. */
+ word = (unsigned char)c;
+ word |= word << 8;
+ word |= word << 16;
+#if LONG_BIT >= 64
+ word |= word << 32;
+#endif
+
+ /* Don't perform memory I/O when passing a zero-length buffer. */
+ if (n == 0)
+ return (NULL);
+
+ /*
+ * First determine whether there is a character unequal to `c'
+ * in the first word. As this word may contain bytes before
+ * `begin', we may execute this loop spuriously.
+ */
+ lp = (const unsigned long *)((uintptr_t)begin & ~LONGPTR_MASK);
+ end = (const unsigned char *)begin + n;
+ if (*lp++ != word)
+ for (p = begin; p < (const unsigned char *)lp;)
+ TESTBYTE;
+
+ /* Now compare the data one word at a time. */
+ for (; (const unsigned char *)lp < end; lp++) {
+ if (*lp != word) {
+ p = (const unsigned char *)lp;
+ TESTBYTE;
+ TESTBYTE;
+ TESTBYTE;
+#if LONG_BIT >= 64
+ TESTBYTE;
+ TESTBYTE;
+ TESTBYTE;
+ TESTBYTE;
+#endif
+ goto done;
+ }
+ }
+
+ return (NULL);
+
+done:
+ /*
+ * If the end of the buffer is not word aligned, the previous
+ * loops may obtain an address that's beyond the end of the
+ * buffer.
+ */
+ if (p < end)
+ return (__DECONST(void *, p));
+ return (NULL);
+}
OpenPOWER on IntegriCloud