On Saturday, 17 of November 2007, Franck Bui-Huu wrote: > Rafael J. Wysocki wrote: > > On Saturday, 17 of November 2007, Franck Bui-Huu wrote: > >> Rafael J. Wysocki wrote: > >>> However, using PF_NOFREEZE to prevent this from happening doesn't seem to be > >>> a good idea. > >>> > >> Indeed but... > >> > >>> I'd probably use wait_event_freezable() (defined in > >>> include/linux/freezer.h) for that. > >> ...I would just revert this bits from now to make sure this driver > >> work again for v2.6.24. > > > > I'd prefer not to. > > > > The PF_NOFREEZE was not present in 2.6.23 already and I wouldn't like to > > reintroduce it now. > > > > Why do you think that using wait_event_freezable() would not work, BTW? > > > > I've never claimed this. I just said it may be safer to revert the > changes for v2.6.24 and improve the current code for next releases. > > >>> It tries to send them fake signals and waits for them to freeze. If > >>> they don't freeze within the timeout, it fails and clears their > >>> TIF_FREEZE bits. > >> But send_fake_signal() seems to wake up task in INTERRUPTIBLE state > >> only. Looking at signal_wake_up(), it basically do: > >> > >> wake_up_state(t, TASK_INTERRUPTIBLE); > >> > >> What am I missing ? > > > > Nothing. :-) > > > > I didn't remember the change that made the freezer use TASK_INTERRUPTIBLE > > explicitly in there (should have looked at the current code before replying). > > > > ok so now we agreed on this point, can we assert that a user > land thread waiting for an event in an UNINTERRUPTIBLE state > will prevent a suspend to happen ? Yes. Rafael _______________________________________________ linux-pm mailing list linux-pm@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linux-foundation.org/mailman/listinfo/linux-pm
