When allocating the slab, the code accidentally computed the array
size from s->slab (an elemtype**).  The slab is an array of elemtype*,
however, so we should take the size of *s->slab.

Noticed-by: Nguyễn Thái Ngọc Duy <pclo...@gmail.com>
Signed-off-by: Thomas Rast <t...@thomasrast.ch>
[I hope this comes through clean.  git-send-email is currently broken
for me, and I'm still investigating, so I have to kludge around it.]

I browsed around for a while, and couldn't find out whether any
architecture actually has any hope of running git (i.e. is at least
mostly POSIX conformant) but still violates the assumption that all
pointers[*] are the same size.

The comp.lang.c FAQ has some interesting examples of wildly different
pointer representations at:


Consider the cases mentioned there where void* and char* have
different representations from, e.g., int*.  Then if elemtype is char,
the slab will be too small before this patch.

But I have no idea if any of those are POSIXish.

One interesting, though orthogonal, tidbit is that POSIX actually
requires _function_ pointers to have the same representation as void*.
>From the specification of dlsym(), which depends on this to be able to
return function pointers:

  [...] Indeed, the ISO C standard does not require that an object of
  type void * can hold a pointer to a function. Implementations
  supporting the XSI extension, however, do require that an object of
  type void * can hold a pointer to a function.

Thank god for POSIX.  So much craziness averted.

[*] Note that C++ method pointers are yet another story.  This only
applies to the kinds of pointers that C supports.

 commit-slab.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/commit-slab.h b/commit-slab.h
index d068e2d..cc114b5 100644
--- a/commit-slab.h
+++ b/commit-slab.h
@@ -91,7 +91,7 @@ struct slabname {                                             
        if (s->slab_count <= nth_slab) {                                \
                int i;                                                  \
                s->slab = xrealloc(s->slab,                             \
-                                  (nth_slab + 1) * sizeof(s->slab));   \
+                                  (nth_slab + 1) * sizeof(*s->slab));  \
                stat_ ##slabname## realloc++;                           \
                for (i = s->slab_count; i <= nth_slab; i++)             \
                        s->slab[i] = NULL;                              \
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to