| commit | f4eb11b6390d5287cca28a2d8af206c47d5f6624 | |
| author | Dan Carpenter <dan.carpenter@oracle.com> | |
| Fri, 7 Nov 2014 08:03:03 +0000 (7 11:03 +0300) | ||
| committer | Dan Carpenter <dan.carpenter@oracle.com> | |
| Fri, 7 Nov 2014 08:03:03 +0000 (7 11:03 +0300) | ||
| tree | 2dea53259af70708249a5c81b4b4c1853c584667 | treesnapshot (tar.gz zip) |
| parent | fa29d8aa153f2d7a4f28af4c36fef9da2c73b22a | commitdiff |
strip_expr: strip out known true false conditionals
This changes: "(1) ? foo : bar" to just "foo".
If I recall correctly this is for handling cpu_to_le32() and friends. For
those functions we want to pretend the bytes swapping didn't happen. That
makes it easier to remember the value. It's ok to record a fake value for
the byte swapped data because hopefully the kernel swaps it back before
using it for anything.
Smatch always says that __builtin_constant_p() is false.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
This changes: "(1) ? foo : bar" to just "foo".
If I recall correctly this is for handling cpu_to_le32() and friends. For
those functions we want to pretend the bytes swapping didn't happen. That
makes it easier to remember the value. It's ok to record a fake value for
the byte swapped data because hopefully the kernel swaps it back before
using it for anything.
Smatch always says that __builtin_constant_p() is false.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
| smatch_helper.c | diffblobblamehistory |