On 03/14/2016 02:42 AM, Jean Delvare wrote: > Hi Peter, > > On Fri, 11 Mar 2016 06:43:59 -0800, Peter Hurley wrote: >> On 03/11/2016 06:31 AM, Jean Delvare wrote: >>> Le Thursday 10 March 2016 à 14:19 -0800, Peter Hurley a écrit : >>>> On 03/08/2016 05:30 AM, Jean Delvare wrote: >>>>> Kconfig option SERIAL_8250_RT288X seems to be only relevant on a >>>>> limited number of platforms, so do not present it on other >>>>> architectures, unless build-testing. >>>>> >>>>> Signed-off-by: Jean Delvare <jdelvare@xxxxxxx> >>>>> Acked-by: John Crispin <blogic@xxxxxxxxxxx> >>>>> Cc: Mans Rullgard <mans@xxxxxxxxx> >>>>> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> >>>>> Cc: Jiri Slaby <jslaby@xxxxxxxx> >>>>> --- >>>>> Changes since v1: >>>>> * Add ARCH_TANGO to the list of platforms, as Måns Rullgård tells me >>>>> it has this hardware too. >>>>> >>>>> drivers/tty/serial/8250/Kconfig | 1 + >>>>> 1 file changed, 1 insertion(+) >>>>> >>>>> --- linux-4.5-rc7.orig/drivers/tty/serial/8250/Kconfig 2016-03-08 09:18:45.610936903 +0100 >>>>> +++ linux-4.5-rc7/drivers/tty/serial/8250/Kconfig 2016-03-08 14:26:14.565737750 +0100 >>>>> @@ -295,6 +295,7 @@ config SERIAL_8250_EM >>>>> config SERIAL_8250_RT288X >>>>> bool "Ralink RT288x/RT305x/RT3662/RT3883 serial port support" >>>>> depends on SERIAL_8250 >>>>> + depends on MIPS || ARCH_TANGO || COMPILE_TEST >>>> >>>> This seems pretty speculative. >>> >>> This is based on the Kconfig plus information I was given meanwhile. >>> This may be incremental but not speculative. >> >> Well, that's not the only way to configure a serial port. >> It can be done from userspace with TIOCSSERIAL ioctl. > > I don't follow you. Aren't we talking about hardware-specific > characteristics? That can't be changed by software at run-time. Yes, it can. ioctl(TIOCSSERIAL) can change the base port address, irq, iotype (including to UPIO_AU which is the port type relevant to this discussion). >>>> I'm concerned that patch v1 was only for MIPS, and now is already >>>> needed on at least one ARM platform. >>> >>> This is still a short list, and this is why I think it should be added. >> >> So what happens when it turns out that this breaks some >> platform and we find out 5 or even 10 versions from now? > > We can always revisit if/when the list gets too long. This isn't carved > in stone. -stable doesn't take build patches. How are you going to make this list longer? > Please understand that my goal is to make "make oldconfig" easier on > kernel version updates (in particular for, but not limited to, > distribution kernel maintainers.) I believe this important goal is > worth some efforts, including the need to update (or revert) a given > hardware dependency later if it was incomplete or wrong. The lack of > documentation is an issue and my work is also an invitation to > contributors to document hardware targets better and/or set hardware > dependencies themselves. > >>>> Also this won't prevent this type of port at all, but rather >>>> will only break its i/o path. >>> >>> Not related to my patch. But I did indeed noticed that PORT_RT2880 is >>> defined regardless of option SERIAL_8250_RT288X, and I agree this is >>> awkward. >> >> And notice how the HUB6 i/o accessors aren't optioned either. > > Yes, same thing. But again unrelated with my patch. I am not > maintaining that subsystem, the complexity/correctness balance is up to > the maintainers really, I don't even have an opinion on it. Sorry, this logic makes no sense to me. But no big deal because I'm just going to remove the Kconfig option so that it doesn't confuse maintainers. >>>> Simpler to just remove the option, as I already wrote. >>> >>> I disagreed back then, and still do. >> >> Sorry, I never saw any reply. Are you sure you sent one? > > I thought I had, but I looked again and no I didn't. Please accept my > apology. No worries. Regards, Peter Hurley -- To unsubscribe from this list: send the line "unsubscribe linux-serial" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
