Inserts a set_current_state(TASK_INTERRUPTIBLE) before the
schedule_timeout() call. Without this change, after the first iteration of the
loop, schedule_timeout() will not only return immediately, but the loop will
break, as the conditional will no longer be satisfied. In fact, this conditional
makes little sense given the workings of schedule_timeout. The timeout variable
is ignored, as well, and I'm fairly certain that it should be included in the
loop conditional. That way, if the timeout expires before a signal hits, -ETIME
will be returned by fdc_interrupt_wait() instead of -EINTR.
Signed-off-by: Nishanth Aravamudan <[email protected]>
Signed-off-by: Domen Puncer <[email protected]>
---
kj-domen/drivers/char/ftape/lowlevel/fdc-io.c | 3 ++-
1 files changed, 2 insertions(+), 1 deletion(-)
diff -puN drivers/char/ftape/lowlevel/fdc-io.c~set_current_state-drivers_char_ftape_lowlevel_fdc-io drivers/char/ftape/lowlevel/fdc-io.c
--- kj/drivers/char/ftape/lowlevel/fdc-io.c~set_current_state-drivers_char_ftape_lowlevel_fdc-io 2005-03-05 16:11:11.000000000 +0100
+++ kj-domen/drivers/char/ftape/lowlevel/fdc-io.c 2005-03-05 16:11:11.000000000 +0100
@@ -387,7 +387,8 @@ int fdc_interrupt_wait(unsigned int time
set_current_state(TASK_INTERRUPTIBLE);
add_wait_queue(&ftape_wait_intr, &wait);
- while (!ft_interrupt_seen && (current->state == TASK_INTERRUPTIBLE)) {
+ while (!ft_interrupt_seen && timeout) {
+ set_current_state(TASK_INTERRUPTIBLE);
timeout = schedule_timeout(timeout);
}
_