Re: [PATCH v2 03/22] rv: Let the reactors take care of buffers

From: Nam Cao
Date: Tue Apr 15 2025 - 05:53:55 EST


On Tue, Apr 15, 2025 at 11:32:44AM +0200, Petr Mladek wrote:
> On Fri 2025-04-11 09:37:19, Nam Cao wrote:
> > -static void rv_printk_reaction(char *msg)
> > +static void rv_printk_reaction(const char *msg, ...)
>
> I wonder whether "make W=1 kernel/trace/rv/reactor_printk.o" would
> start complaining about that this function is a candidate for
> ‘gnu_printf’ format attribute.

I checked, it does. Thanks for pointing it out.

> I am not sure. Maybe it is enough that this function is later assigned to
> the .react callback in struct rv_reactor.
>
> I wanted to tried it myself. But I was not able to compile the
> code in linux-next. I got something like:
>
> ./include/linux/rv.h: In function ‘rv_ltl_valid_state’:
> ./include/linux/rv.h:55:43: error: ‘struct ltl_monitor’ has no member named ‘states’
> 55 | for (int i = 0; i < ARRAY_SIZE(mon->states); ++i) {
> | ^~

This is a problem with the series. For now, you could "fix" this error with
CONFIG_RV_MON_RTAPP=y. I will fix it up properly in the next version.
> ...
>
> I am actually not sure against which tree I should apply this patchset.
> It did apply on linux-next after skipping the 1st patch.
> But it does not compile there.

linux-next is supposed to be fine. It doesn't build due to a problem
introduced by this series.

> The __printf statement might be missing also in the other two
> reactors (panic, nop).

Yes. Will be fixed.

Thanks for the review!
Nam