On Tue, Sep 01, 2020 at 06:11:54PM +0200, Oleg Nesterov wrote:
On 08/31, Christian Brauner wrote:
--- a/kernel/exit.c +++ b/kernel/exit.c @@ -934,6 +934,7 @@ struct wait_opts { wait_queue_entry_t child_wait; int notask_error;
- int eagain_error;
}; static int eligible_pid(struct wait_opts *wo, struct task_struct *p) @@ -1461,6 +1462,8 @@ static long do_wait(struct wait_opts *wo) notask: retval = wo->notask_error;
- if (!retval)
retval = wo->eagain_error;I must have missed something but I don't understand why do we need the new ->eagain_error and the change in do_wait().
@@ -1544,6 +1551,11 @@ static long kernel_waitid(int which, pid_t upid, struct waitid_info *infop, wo.wo_flags = options; wo.wo_info = infop; wo.wo_rusage = ru;
- wo.eagain_error = 0;
- if (f_flags & O_NONBLOCK) {
wo.wo_flags |= WNOHANG;wo.eagain_error = -EAGAIN;- } ret = do_wait(&wo);
Can't kernel_waitid() simply do
if (f_flags & O_NONBLOCK) wo.wo_flags |= WNOHANG; ret = do_wait(); if (!ret & (f_flags & O_NONBLOCK)) ret = -EAGAIN;
?
Heh, indeed, that's even a smaller patch. Will change to that!
Thanks for the review, Oleg! Christia