[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PATCH] memchr: also detect Alpha bug
From: |
Eric Blake |
Subject: |
[PATCH] memchr: also detect Alpha bug |
Date: |
Tue, 14 Sep 2010 14:48:39 -0600 |
http://sourceware.org/bugzilla/show_bug.cgi?id=12019
* m4/memchr.m4 (gl_FUNC_MEMCHR): Detect glibc 2.11.2 failure on
Alpha.
* doc/posix-functions/memchr.texi (memchr): Tweak wording.
* tests/test-memchr.c (main): Enhance test.
Reported by Nelson H. F. Beebe.
Signed-off-by: Eric Blake <address@hidden>
---
Tested that this fixes test-strstr on Alpha; also that the enhanced
test-memchr passes on x86_64 before and after the rest of the patch;
a limit of 257 took about .25 seconds, while a limit of 500 took 2.5
seconds. Yes, the new test is O(n^3), so a sane limit is important.
Meanwhile, on Alpha, without this patch, the existing test-memchr
was too weak to detect failure; using just the tests/ fix exposes
the Alpha bug, and the rest of this patch again passes.
ChangeLog | 9 +++++++++
doc/posix-functions/memchr.texi | 2 +-
m4/memchr.m4 | 4 +++-
tests/test-memchr.c | 18 +++++++++++++++---
4 files changed, 28 insertions(+), 5 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index a9cb976..43bebf6 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,12 @@
+2010-09-14 Eric Blake <address@hidden>
+
+ memchr: also detect Alpha bug
+ * m4/memchr.m4 (gl_FUNC_MEMCHR): Detect glibc 2.11.2 failure on
+ Alpha.
+ * doc/posix-functions/memchr.texi (memchr): Tweak wording.
+ * tests/test-memchr.c (main): Enhance test.
+ Reported by Nelson H. F. Beebe.
+
2010-09-13 Paul Eggert <address@hidden>
fts, getcwd, glob: audit for dirfd returning -1
diff --git a/doc/posix-functions/memchr.texi b/doc/posix-functions/memchr.texi
index 69b693a..3c9cfcb 100644
--- a/doc/posix-functions/memchr.texi
+++ b/doc/posix-functions/memchr.texi
@@ -13,7 +13,7 @@ memchr
@item
This function dereferences too much memory on some platforms:
-glibc 2.10 on x86_64, IA-64, Alpha.
+glibc 2.10 on x86_64, IA-64; glibc 2.11 on Alpha.
@end itemize
Portability problems not fixed by Gnulib:
diff --git a/m4/memchr.m4 b/m4/memchr.m4
index ab773b8..b05a79a 100644
--- a/m4/memchr.m4
+++ b/m4/memchr.m4
@@ -1,4 +1,4 @@
-# memchr.m4 serial 8
+# memchr.m4 serial 9
dnl Copyright (C) 2002-2004, 2009-2010 Free Software Foundation, Inc.
dnl This file is free software; the Free Software Foundation
dnl gives unlimited permission to copy and/or distribute it,
@@ -62,6 +62,8 @@ AC_DEFUN_ONCE([gl_FUNC_MEMCHR],
strcpy (fence - 9, "12345678");
if (memchr (fence - 9, 0, 79) != fence - 1)
return 2;
+ if (memchr (fence - 1, 0, 3) != fence - 1)
+ return 3;
}
return 0;
]])], [gl_cv_func_memchr_works=yes], [gl_cv_func_memchr_works=no],
diff --git a/tests/test-memchr.c b/tests/test-memchr.c
index a801614..e7b9780 100644
--- a/tests/test-memchr.c
+++ b/tests/test-memchr.c
@@ -87,25 +87,37 @@ main (void)
/* Check that memchr() does not read past the first occurrence of the
byte being searched. See the Austin Group's clarification
- <http://www.opengroup.org/austin/docs/austin_454.txt>. */
+ <http://www.opengroup.org/austin/docs/austin_454.txt>.
+ Test both '\0' and something else, since some implementations
+ special-case searching for NUL.
+ */
{
char *page_boundary = (char *) zerosize_ptr ();
+ /* Too small, and we miss cache line boundary tests; too large,
+ and the test takes cubically longer to complete. */
+ int limit = 257;
if (page_boundary != NULL)
{
- for (n = 1; n <= 500; n++)
+ for (n = 1; n <= limit; n++)
{
char *mem = page_boundary - n;
memset (mem, 'X', n);
ASSERT (MEMCHR (mem, 'U', n) == NULL);
+ ASSERT (MEMCHR (mem, 0, n) == NULL);
{
size_t i;
+ size_t k;
for (i = 0; i < n; i++)
{
mem[i] = 'U';
- ASSERT (MEMCHR (mem, 'U', 4000) == mem + i);
+ for (k = i + 1; k < n + limit; k++)
+ ASSERT (MEMCHR (mem, 'U', k) == mem + i);
+ mem[i] = 0;
+ for (k = i + 1; k < n + limit; k++)
+ ASSERT (MEMCHR (mem, 0, k) == mem + i);
mem[i] = 'X';
}
}
--
1.7.2.2