This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] malloc: Correct malloc alignment on 32-bit architectures [BZ #6527]


On 05/13/2016 11:31 PM, DJ Delorie wrote:

fweimer@redhat.com (Florian Weimer) writes:
-#define MALLOC_STATE_VERSION (0 * 0x100l + 4l) /* major*0x100 + minor */
+#define MALLOC_STATE_VERSION (0 * 0x100l + 5l) /* major*0x100 + minor */

Do we record a history of what changed for each version bump, in case we
need to go back and somehow support an older one?

I see no reason to start this practice now, when the dumping functionality is rapidly approaching removal. (Undumping will still be supported, of course.)

If necessary, we can consult the Git history.

-#ifndef MALLOC_ALIGNMENT
-# if !SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_16)
-/* This is the correct definition when there is no past ABI to constrain it.
-
-   Among configurations with a past ABI constraint, it differs from
-   2*SIZE_SZ only on powerpc32.  For the time being, changing this is
-   causing more compatibility problems due to malloc_get_state and
-   malloc_set_state than will returning blocks not adequately aligned for
-   long double objects under -mlong-double-128.  */
-
-#  define MALLOC_ALIGNMENT       (2 *SIZE_SZ < __alignof__ (long double)      \
-                                  ? __alignof__ (long double) : 2 *SIZE_SZ)
-# else
-#  define MALLOC_ALIGNMENT       (2 *SIZE_SZ)
-# endif
-#endif
+#define MALLOC_ALIGNMENT       (2 * SIZE_SZ < __alignof__ (long double)      \
+				? __alignof__ (long double) : 2 * SIZE_SZ)

This drops the outer '#ifndef' which I think should remain.  The
ChangeLog mentions an override for HPPA.

No such override exists anymore.

If there is some platform-specific overrides for ppc32, we should take
those out instead.

The patch is doing that.

I've restored the override possibility for now. New version of the patch is attached.

Thanks,
Florian

malloc: Correct malloc alignment on 32-bit architectures [BZ #6527]

After the heap rewriting added in commit
4cf6c72fd2a482e7499c29162349810029632c3f (malloc: Rewrite dumped heap
for compatibility in __malloc_set_state), we can change malloc alignment
for new allocations because the alignment of old allocations no longer
matters.

We need to increase the malloc state version number, so that binaries
containing dumped heaps of the new layout will not try to run on
previous versions of glibc, resulting in obscure crashes.

This commit addresses a failure of tst-malloc-thread-fail on the
affected architectures (32-bit ppc and mips) because the test checks
pointer alignment.

2016-05-20  Florian Weimer  <fweimer@redhat.com>

	[BZ #6527]
	* malloc/malloc.c (MALLOC_ALIGNMENT): Use correct alignment
	unconditionally.
	* malloc/hooks.c (MALLOC_STATE_VERSION): Increase state version.

diff --git a/malloc/hooks.c b/malloc/hooks.c
index 45241f2..caa1e70 100644
--- a/malloc/hooks.c
+++ b/malloc/hooks.c
@@ -465,7 +465,7 @@ memalign_check (size_t alignment, size_t bytes, const void *caller)
    then the hooks are reset to 0.  */
 
 #define MALLOC_STATE_MAGIC   0x444c4541l
-#define MALLOC_STATE_VERSION (0 * 0x100l + 4l) /* major*0x100 + minor */
+#define MALLOC_STATE_VERSION (0 * 0x100l + 5l) /* major*0x100 + minor */
 
 struct malloc_save_state
 {
diff --git a/malloc/malloc.c b/malloc/malloc.c
index 44524ff..ead9a21 100644
--- a/malloc/malloc.c
+++ b/malloc/malloc.c
@@ -352,20 +352,8 @@ __malloc_assert (const char *assertion, const char *file, unsigned int line,
 
 
 #ifndef MALLOC_ALIGNMENT
-# if !SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_16)
-/* This is the correct definition when there is no past ABI to constrain it.
-
-   Among configurations with a past ABI constraint, it differs from
-   2*SIZE_SZ only on powerpc32.  For the time being, changing this is
-   causing more compatibility problems due to malloc_get_state and
-   malloc_set_state than will returning blocks not adequately aligned for
-   long double objects under -mlong-double-128.  */
-
-#  define MALLOC_ALIGNMENT       (2 *SIZE_SZ < __alignof__ (long double)      \
-                                  ? __alignof__ (long double) : 2 *SIZE_SZ)
-# else
-#  define MALLOC_ALIGNMENT       (2 *SIZE_SZ)
-# endif
+# define MALLOC_ALIGNMENT       (2 * SIZE_SZ < __alignof__ (long double) \
+				 ? __alignof__ (long double) : 2 * SIZE_SZ)
 #endif
 
 /* The corresponding bit mask value */

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]