| commit | 215c7ada006d3badc4fc28bd1b8f5f03fddf04b3 | |
| author | Dan Carpenter <dan.carpenter@oracle.com> | |
| Tue, 6 Mar 2018 14:14:48 +0000 (6 17:14 +0300) | ||
| committer | Dan Carpenter <dan.carpenter@oracle.com> | |
| Tue, 6 Mar 2018 14:14:48 +0000 (6 17:14 +0300) | ||
| tree | 19331745418e02c7668567604a69e2683f30d482 | treesnapshot (tar.gz zip) |
| parent | 4d6a2546b5a98d3386edfc159dad5ca9b9bb6a70 | commitdiff |
return_to_param: fix a crashing bug
This started happening when I upgraded to the latest Sparse.
So I'm not totally sure why I did this HACK ALERT... And I also can't
swear that this fix is correct because the test cases aren't there...
But what happens is that we use the left_sym, because it gets passed to
smatch_extra mod expression. So the hack says we don't use it but that's
wrong. Then check_rosenberg.c hooks into smatch_extra and looks up the
type of the symbol which doesn't work since it's an expression pointer.
And it cause a crash.
I think this fix will work 99 times out of a hundred and it's probably
harmless the remaining time.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
This started happening when I upgraded to the latest Sparse.
So I'm not totally sure why I did this HACK ALERT... And I also can't
swear that this fix is correct because the test cases aren't there...
But what happens is that we use the left_sym, because it gets passed to
smatch_extra mod expression. So the hack says we don't use it but that's
wrong. Then check_rosenberg.c hooks into smatch_extra and looks up the
type of the symbol which doesn't work since it's an expression pointer.
And it cause a crash.
I think this fix will work 99 times out of a hundred and it's probably
harmless the remaining time.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
| smatch_return_to_param.c | diffblobblamehistory |