| commit | b267c9cd1516dd349e3891f22198dd076723d22e | |
| author | Rasmus Villemoes <rv@rasmusvillemoes.dk> | |
| Wed, 11 Feb 2015 12:46:22 +0000 (11 13:46 +0100) | ||
| committer | Dan Carpenter <dan.carpenter@oracle.com> | |
| Thu, 12 Feb 2015 20:43:12 +0000 (12 23:43 +0300) | ||
| tree | ac437453febfcc0af93a080271478e48d75b245d | treesnapshot (tar.gz zip) |
| parent | 4a08be60b341329d17800e9c32761cfde54f103e | commitdiff |
check_kernel_printf.c: Handle 'fmt ? : "default"'
It is not completely unlikely that someone could make a call like
sprintf(buf, fmt ? : "some default");. In that case, ->cond_true is
NULL, and we should therefore use ->conditional instead. That is very
unlikely to be a string literal which we can actually check (why would
the ?: then be used), but it might come from some macro expansion, and
it's certainly better than segfaulting.
It is not completely unlikely that someone could make a call like
sprintf(buf, fmt ? : "some default");. In that case, ->cond_true is
NULL, and we should therefore use ->conditional instead. That is very
unlikely to be a string literal which we can actually check (why would
the ?: then be used), but it might come from some macro expansion, and
it's certainly better than segfaulting.
| check_kernel_printf.c | diffblobblamehistory |