* [PATCH] clocksource: add an "menu" item to Kconfig
@ 2013-08-07 10:56 Sebastian Andrzej Siewior
2013-08-07 16:21 ` John Stultz
0 siblings, 1 reply; 7+ messages in thread
From: Sebastian Andrzej Siewior @ 2013-08-07 10:56 UTC (permalink / raw)
To: John Stultz; +Cc: Thomas Gleixner, linux-kernel, Sebastian Andrzej Siewior
This patch adds a menu item to clocksource' Kconfig file. Without it the
changeable items look somehow lost in the main drivers' menu
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
drivers/clocksource/Kconfig | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/clocksource/Kconfig b/drivers/clocksource/Kconfig
index d0e207b..95660d8 100644
--- a/drivers/clocksource/Kconfig
+++ b/drivers/clocksource/Kconfig
@@ -1,3 +1,5 @@
+menu "Clocksource Framework"
+
config CLKSRC_OF
bool
@@ -123,3 +125,4 @@ config VF_PIT_TIMER
bool
help
Support for Period Interrupt Timer on Freescale Vybrid Family SoCs.
+endmenu
--
1.8.3.2
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH] clocksource: add an "menu" item to Kconfig
2013-08-07 10:56 [PATCH] clocksource: add an "menu" item to Kconfig Sebastian Andrzej Siewior
@ 2013-08-07 16:21 ` John Stultz
2013-08-07 16:30 ` Sebastian Andrzej Siewior
0 siblings, 1 reply; 7+ messages in thread
From: John Stultz @ 2013-08-07 16:21 UTC (permalink / raw)
To: Sebastian Andrzej Siewior
Cc: Thomas Gleixner, linux-kernel, Linus Walleij, Mattias Wallin,
Fabio Baltieri
On 08/07/2013 03:56 AM, Sebastian Andrzej Siewior wrote:
> This patch adds a menu item to clocksource' Kconfig file. Without it the
> changeable items look somehow lost in the main drivers' menu
So I'd like to push back on this a bit and ask why we need user-visible
clocksource config options?
I'm much prefer the platform options select the needed clocksources, and
not leave it up to the user (who really doesn't want to go digging
around to find the clocksource sub-menu under drivers, and scan a list
of 100 options that aren't possible on their hardware).
For an example of what I consider bad behavior here, see the RTC driver
directory - at a certain point they just become hashes: "Oh no.. you
selected rtc-387aae6fd, on your hardware, you need rtc-adfb8e5".
It really just seems like it forces more unnecessary options on the user
and more ways to build a kernel that won't boot.
I do get you're trying to fix up the bad situation, caused by
user-prompts slipping into the Kconfig.
But I might prefer a patch closer to whats below.
thanks
-john
diff --git a/drivers/clocksource/Kconfig b/drivers/clocksource/Kconfig
index b7b9b04..b3050db 100644
--- a/drivers/clocksource/Kconfig
+++ b/drivers/clocksource/Kconfig
@@ -57,14 +57,14 @@ config CLKSRC_NOMADIK_MTU_SCHED_CLOCK
Use the Multi Timer Unit as the sched_clock.
config CLKSRC_DBX500_PRCMU
- bool "Clocksource PRCMU Timer"
+ bool
depends on UX500_SOC_DB8500
default y
help
Use the always on PRCMU Timer as clocksource
config CLKSRC_DBX500_PRCMU_SCHED_CLOCK
- bool "Clocksource PRCMU Timer sched_clock"
+ bool
depends on (CLKSRC_DBX500_PRCMU && !CLKSRC_NOMADIK_MTU_SCHED_CLOCK)
default y
help
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH] clocksource: add an "menu" item to Kconfig
2013-08-07 16:21 ` John Stultz
@ 2013-08-07 16:30 ` Sebastian Andrzej Siewior
2013-08-07 16:58 ` John Stultz
0 siblings, 1 reply; 7+ messages in thread
From: Sebastian Andrzej Siewior @ 2013-08-07 16:30 UTC (permalink / raw)
To: John Stultz
Cc: Thomas Gleixner, linux-kernel, Linus Walleij, Mattias Wallin,
Fabio Baltieri
On 08/07/2013 06:21 PM, John Stultz wrote:
> On 08/07/2013 03:56 AM, Sebastian Andrzej Siewior wrote:
>> This patch adds a menu item to clocksource' Kconfig file. Without it the
>> changeable items look somehow lost in the main drivers' menu
>
> So I'd like to push back on this a bit and ask why we need user-visible
> clocksource config options?
There was one visible Option if I remember correctly. I added two and
was wondering once I found them. I do have a PCI card which provides a
clock source and clock events device. So I assumed select <m|y> would
be doable.
> I'm much prefer the platform options select the needed clocksources, and
> not leave it up to the user (who really doesn't want to go digging
> around to find the clocksource sub-menu under drivers, and scan a list
> of 100 options that aren't possible on their hardware).
>
> For an example of what I consider bad behavior here, see the RTC driver
> directory - at a certain point they just become hashes: "Oh no.. you
> selected rtc-387aae6fd, on your hardware, you need rtc-adfb8e5".
>
> It really just seems like it forces more unnecessary options on the user
> and more ways to build a kernel that won't boot.
>
> I do get you're trying to fix up the bad situation, caused by
> user-prompts slipping into the Kconfig.
Yes, that is fine. That PCI device is divided into multiple devices so
once it is enabled I could use "select" for required clock source.
> But I might prefer a patch closer to whats below.
Okay.
>
> thanks
> -john
Sebastian
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] clocksource: add an "menu" item to Kconfig
2013-08-07 16:30 ` Sebastian Andrzej Siewior
@ 2013-08-07 16:58 ` John Stultz
2013-08-12 12:55 ` Daniel Lezcano
0 siblings, 1 reply; 7+ messages in thread
From: John Stultz @ 2013-08-07 16:58 UTC (permalink / raw)
To: Sebastian Andrzej Siewior
Cc: Thomas Gleixner, linux-kernel, Linus Walleij, Daniel Lezcano
On 08/07/2013 09:30 AM, Sebastian Andrzej Siewior wrote:
> On 08/07/2013 06:21 PM, John Stultz wrote:
>> On 08/07/2013 03:56 AM, Sebastian Andrzej Siewior wrote:
>>> This patch adds a menu item to clocksource' Kconfig file. Without it the
>>> changeable items look somehow lost in the main drivers' menu
>> So I'd like to push back on this a bit and ask why we need user-visible
>> clocksource config options?
> There was one visible Option if I remember correctly. I added two and
> was wondering once I found them. I do have a PCI card which provides a
> clock source and clock events device. So I assumed select <m|y> would
> be doable.
PCI card, eh? (Heh. In writing my previous mail, I think I even wrote
"These are clocksources soldered to the motherboard, not pci cards!",
but deleted it as it didn't seem useful. :)
So I think that's actually a first for clocksource drivers.
So, if there are genuinely cross-architecture pluggable clocksources,
that aren't part of a platform definition, I would be ok with making an
exception. But I'd want to make sure that we kept a high-bar for adding
clocksource options, so that folks don't just clutter up the Kconfig menu.
That said, Daniel (now cc'ed - sorry for not cc'ing you earlier!) is now
in charge of the clocksource directory, so my objections and wishes
aren't as important as his.
thanks
-john
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] clocksource: add an "menu" item to Kconfig
2013-08-07 16:58 ` John Stultz
@ 2013-08-12 12:55 ` Daniel Lezcano
2013-08-12 13:04 ` Sebastian Andrzej Siewior
0 siblings, 1 reply; 7+ messages in thread
From: Daniel Lezcano @ 2013-08-12 12:55 UTC (permalink / raw)
To: John Stultz
Cc: Sebastian Andrzej Siewior, Thomas Gleixner, linux-kernel, Linus Walleij
On 08/07/2013 06:58 PM, John Stultz wrote:
> On 08/07/2013 09:30 AM, Sebastian Andrzej Siewior wrote:
>> On 08/07/2013 06:21 PM, John Stultz wrote:
>>> On 08/07/2013 03:56 AM, Sebastian Andrzej Siewior wrote:
>>>> This patch adds a menu item to clocksource' Kconfig file. Without it
>>>> the
>>>> changeable items look somehow lost in the main drivers' menu
>>> So I'd like to push back on this a bit and ask why we need user-visible
>>> clocksource config options?
>> There was one visible Option if I remember correctly. I added two and
>> was wondering once I found them. I do have a PCI card which provides a
>> clock source and clock events device. So I assumed select <m|y> would
>> be doable.
>
> PCI card, eh? (Heh. In writing my previous mail, I think I even wrote
> "These are clocksources soldered to the motherboard, not pci cards!",
> but deleted it as it didn't seem useful. :)
>
> So I think that's actually a first for clocksource drivers.
>
> So, if there are genuinely cross-architecture pluggable clocksources,
> that aren't part of a platform definition, I would be ok with making an
> exception. But I'd want to make sure that we kept a high-bar for adding
> clocksource options, so that folks don't just clutter up the Kconfig menu.
>
> That said, Daniel (now cc'ed - sorry for not cc'ing you earlier!) is now
> in charge of the clocksource directory, so my objections and wishes
> aren't as important as his.
Sebastian,
instead of adding a menu item in the Kconfig, is it possible to have
this Kconfig PCI board support to select the right driver ?
Thanks
-- Daniel
--
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] clocksource: add an "menu" item to Kconfig
2013-08-12 12:55 ` Daniel Lezcano
@ 2013-08-12 13:04 ` Sebastian Andrzej Siewior
2013-08-12 13:06 ` Daniel Lezcano
0 siblings, 1 reply; 7+ messages in thread
From: Sebastian Andrzej Siewior @ 2013-08-12 13:04 UTC (permalink / raw)
To: Daniel Lezcano; +Cc: John Stultz, Thomas Gleixner, linux-kernel, Linus Walleij
On 08/12/2013 02:55 PM, Daniel Lezcano wrote:
> Sebastian,
Hi Daniel,
> instead of adding a menu item in the Kconfig, is it possible to have
> this Kconfig PCI board support to select the right driver ?
I decided to go with only with "tristate" so it does not pop up in
Kconfig and will be selected by the MFD/PCI driver. Is this what you
have in mind?
> Thanks
> -- Daniel
Sebastian
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] clocksource: add an "menu" item to Kconfig
2013-08-12 13:04 ` Sebastian Andrzej Siewior
@ 2013-08-12 13:06 ` Daniel Lezcano
0 siblings, 0 replies; 7+ messages in thread
From: Daniel Lezcano @ 2013-08-12 13:06 UTC (permalink / raw)
To: Sebastian Andrzej Siewior
Cc: John Stultz, Thomas Gleixner, linux-kernel, Linus Walleij
On 08/12/2013 03:04 PM, Sebastian Andrzej Siewior wrote:
> On 08/12/2013 02:55 PM, Daniel Lezcano wrote:
>> Sebastian,
>
> Hi Daniel,
>
>> instead of adding a menu item in the Kconfig, is it possible to have
>> this Kconfig PCI board support to select the right driver ?
>
> I decided to go with only with "tristate" so it does not pop up in
> Kconfig and will be selected by the MFD/PCI driver. Is this what you
> have in mind?
Yes, right. The clocksource's Kconfig is kept untouched.
Thanks
-- Daniel
--
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2013-08-12 13:06 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-08-07 10:56 [PATCH] clocksource: add an "menu" item to Kconfig Sebastian Andrzej Siewior
2013-08-07 16:21 ` John Stultz
2013-08-07 16:30 ` Sebastian Andrzej Siewior
2013-08-07 16:58 ` John Stultz
2013-08-12 12:55 ` Daniel Lezcano
2013-08-12 13:04 ` Sebastian Andrzej Siewior
2013-08-12 13:06 ` Daniel Lezcano
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.