sh: clock framework update, fix count and kill off kref
PATCH 1/1 sh: clock framework update, fix count and kill off kref
From: Magnus Damm <[email protected]>
This patch updates the clock framework use count code. With this patch the enable() and disable() callbacks only get called when counting from and to zero. While at it the kref stuff gets replaced with an int.
---
arch/sh/include/asm/clock.h | 4 +- arch/sh/kernel/cpu/clock.c | 69 +++++++++++++++++++++++-------------------- 2 files changed, 39 insertions(+), 34 deletions(-)
arch/sh/include/asm/clock.h +2 -2
@@ -1,7 +1,6 @@#ifndef __ASM_SH_CLOCK_H#define __ASM_SH_CLOCK_H -#include <linux/kref.h> #include <linux/list.h> #include <linux/seq_file.h> #include <linux/clk.h>
@@ -28,7 +27,7 @@ struct clk {struct clk *parent;struct clk_ops *ops; -struct kref kref; +int usecount;unsigned long rate;unsigned long flags;
@@ -37,6 +36,7 @@ struct clk {#define CLK_ALWAYS_ENABLED (1 << 0)#define CLK_RATE_PROPAGATES (1 << 1) +#define CLK_NEEDS_INIT (1 << 2)/* Should be defined by processor-specific code */ void arch_init_clk_ops (struct clk_ops **,int type);
arch/sh/kernel/cpu/clock.c +37 -32
137 lines not on this page · raw message
Paul Mundt
AppliedQuoted 4 lines · Magnus Damm
This patch updates the clock framework use count code. With this patch the
enable()anddisable()callbacks only get called when counting from and to zero. While at it the kref stuff gets replaced with an int.
Quoted 3 lines · Magnus Damm
This patch enables the TMU clocksource on sh7722. To prioritize TMU over CMT we also adjust the CMT clock source rating.
Quoted 3 lines · Magnus Damm
This patch adds TMU platform data for sh7723. Both clockevent and clocksource support is enabled. While at it, adjust the CMT clocksource rating to prioritize the TMU.
Applied, thanks.
Hi Paul
Few days ago I had a discussion with Magnus on the current status of the clock framework
and both we agree to change how the .init call back should work.
The design I have in mind is:
- the .init callback should be called during the clk_registration and if the init fails the clock framework rejects the clock.
- Any operations with (CLK_NEEDS_INIT) has to be managed internally to the clock-SOC implementation via archflag (or a void* private_data)
to hive any
implementation issue at clock framework level.
Magnus, Do you want add anything?
Paul, Do you have some issue on this design?
Regards
Francesco
Paul Mundt ha scritto:
Quoted 7 lines · Magnus Damm
Quoted 6 lines
This patch updates the clock framework use count code. With this patch the
enable()anddisable()callbacks only get called when counting from and to zero. While at it the kref stuff gets replaced with an int.
Quoted 9 lines · Francesco VIRLINZI
The design I have in mind is:
- the .init callback should be called during theclk_registrationand if the init fails the clock framework rejects the clock.
- Any operations with (CLK_NEEDS_INIT) has to be managed internally to the clock-SOC implementation via archflag (or a void*private_data) to hive any
implementation issue at clock framework level.Magnus, Do you want add anything?
Not really, except I have to test it and make sure it will work. =) There may be historical reasons why the code is written as it is, so let's try to only call init() from clock registration and see how that goes. Do you have a patch? =)
/ magnus
Hi Magnus
Quoted 5 lines
Not really, except I have to test it and make sure it will work. =) There may be historical reasons why the code is written as it is, so let's try to only call
init()from clock registration and see how that goes. Do you have a patch? =)
I will post it asap, just the time to rebase my repository. Regards
Francesco
Quoted 7 lines
/ magnus
--
To unsubscribe from this list: send the line "unsubscribe linux-sh" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html