| commit | fc6e910e69f49cd482966eb9af3f27c0b65bcf4b | |
| author | Volker Lendecke <vl@samba.org> | |
| Sat, 2 Oct 2010 09:13:34 +0000 (2 11:13 +0200) | ||
| committer | Volker Lendecke <vl@samba.org> | |
| Sat, 2 Oct 2010 09:17:49 +0000 (2 11:17 +0200) | ||
| tree | 46abee39db9a204b2268b8a199cfe18d62a08251 | treesnapshot (tar.gz zip) |
| parent | ed5e05a15d3a8bea3a4f428d1664c9c5077866c2 | commitdiff |
s3: Fix a 64-bit BUG
Quick explanation why this is a problem:
It is not safe to pass a pointer to a uint32_t into a routine that expects
a size_t *. The routine that this pointer is passed to has no chance to find
out that in reality it was passed a pointer to a uint32_t. On platforms where
sizeof(size_t)==8, we get a nice overwrite of a stack area that should not
be overwritten.
As this happens over an over again: Does someone know a means to make gcc
stop with an error in this case?
Volker
Quick explanation why this is a problem:
It is not safe to pass a pointer to a uint32_t into a routine that expects
a size_t *. The routine that this pointer is passed to has no chance to find
out that in reality it was passed a pointer to a uint32_t. On platforms where
sizeof(size_t)==8, we get a nice overwrite of a stack area that should not
be overwritten.
As this happens over an over again: Does someone know a means to make gcc
stop with an error in this case?
Volker
| lib/util/asn1.c | diffblobblamehistory |