On Thu, 7 Mar 2024 11:07:17 +1100 Stephen Rothwell wrote: > Hi all, > > After merging the ext3 tree, today's linux-next build (x86_64 > allmodconfig) failed like this: > > In file included from include/linux/sysctl.h:27, > from include/linux/fanotify.h:5, > from fs/notify/fanotify/fanotify.c:2: > fs/notify/fanotify/fanotify.c: In function 'fanotify_get_response': > fs/notify/fanotify/fanotify.c:233:48: error: suggest parentheses > around arithmetic in operand of '|' [-Werror=parentheses] 233 | > TASK_KILLABLE|TASK_FREEZABLE); | > ^ include/linux/wait.h:283:11: > note: in definition of macro '___wait_is_interruptible' 283 | > (state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL))) | ^~~~~ > include/linux/wait.h:935:9: note: in expansion of macro > '___wait_event' 935 | ___wait_event(wq, condition, state, 0, > 0, schedule()) | ^~~~~~~~~~~~~ > include/linux/wait.h:958:25: note: in expansion of macro > '__wait_event_state' 958 | __ret = > __wait_event_state(wq_head, condition, state); \ | > ^~~~~~~~~~~~~~~~~~ > fs/notify/fanotify/fanotify.c:231:15: note: in expansion of macro > 'wait_event_state' 231 | ret = > wait_event_state(group->fanotify_data.access_waitq, | > ^~~~~~~~~~~~~~~~ cc1: all warnings being treated as errors > > Caused by commit > > 3440e7e55ced ("fanotify: allow freeze when waiting response for > permission events") > > Though, I guess, you could argue that the ___wait_is_interruptible > macro should parenthesise the use of its "state" argument. > > I have used the ext3 tree from next-20240306 for today. > Sorry I missed this warning. And agreed! I can add parentheses on the call side, but it may be more reasonable to add them in the macro. Hey Peter, Could you please take a look at the patch attached at your convenience? Thanks! -- Thanks, Winston