On Thu, 04/20 12:03, Kevin Wolf wrote: > Am 20.04.2017 um 10:32 hat Fam Zheng geschrieben: > > > /var/tmp/patchew-tester-tmp-7vbaovju/src/block/file-posix.c: In function > > > ‘raw_handle_lock_update’: > > > /var/tmp/patchew-tester-tmp-7vbaovju/src/block/file-posix.c:1117:17: > > > error: ‘lock_fd’ may be used uninitialized in this function > > > [-Werror=maybe-uninitialized] > > > qemu_close(lock_fd); > > > ^~~~~~~~~~~~~~~~~~~ > > > cc1: all warnings being treated as errors > > > /var/tmp/patchew-tester-tmp-7vbaovju/src/rules.mak:69: recipe for target > > > 'block/file-posix.o' failed > > > > False positive. lock_fd is initialized in a "if (op == RAW_LT_PREPARE)" > > branch > > above, and we are in a "case RAW_LT_PREPARE" switch branch too. > > Even if it's a false positive, obviously we still need to work around > the build failure.
Sure, easy to do. I'll initialize the variable. Fam