blob: 28038b7115b493664f49d96661c82c81b9a30d5e (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
|
From: Niko Tyni <ntyni@debian.org>
Subject: Prevent gcc from optimizing the alignment test away on armel
Bug-Debian: http://bugs.debian.org/289884
Author: Marc Pignat <marc.pignat@hevs.ch>
Origin: upstream, http://perl5.git.perl.org/perl.git/commit/f1c7503b9028d20741c9a01345ba8704998ea381
As hunted down by Marc Pignat, gcc optimizations make the check for
u32align (U32_ALIGNMENT_REQUIRED) a no-op on armel, breaking the MD5
module.
---
Configure | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/Configure b/Configure
index 01fa3c0..f39b2f9 100755
--- a/Configure
+++ b/Configure
@@ -18533,6 +18533,7 @@ $cat <<EOM
Checking to see whether you can access character data unalignedly...
EOM
+: volatile so that the compiler does not optimize the test away
case "$d_u32align" in
'') $cat >try.c <<EOCP
#include <stdio.h>
@@ -18549,8 +18550,8 @@ $signal_t bletch(int s) { exit(4); }
#endif
int main() {
#if BYTEORDER == 0x1234 || BYTEORDER == 0x4321
- U8 buf[8];
- U32 *up;
+ $volatile U8 buf[8];
+ $volatile U32 *up;
int i;
if (sizeof(U32) != 4) {
--
tg: (daf8b46..) fixes/arm-alignment (depends on: upstream)
|
