[PATCH v4 1/7] OMAP2+: clockdomain: Add an api to read idle mode
Paul Walmsley
paul at pwsan.com
Fri Jul 8 20:06:08 EDT 2011
Hi
On Fri, 1 Jul 2011, Benoit Cousson wrote:
> From: Rajendra Nayak <rnayak at ti.com>
>
> Add a clockdomain api to check if hardware supervised
> idle transitions are enabled on a clockdomain.
>
> Thanks to Todd Poynor <toddpoynor at google.com> for a
> better function name suggestion.
>
> Signed-off-by: Rajendra Nayak <rnayak at ti.com>
> Cc: Paul Walmsley <paul at pwsan.com>
> Cc: Todd Poynor <toddpoynor at google.com>
I've been looking at this patch and the subsequent one. Two thoughts
occurred to me.
The first is that clkdm_allows_idle() shouldn't need to touch the CM
hardware at all, since all clockdomain auto-idle mode changes should go
through the clockdomain code. So we can just keep an in-memory flag in
the struct clockdomain and avoid the slow CM register access.
The second is that the name 'clkdm_allows_idle', while a better name
than the original, still doesn't seem right. Naïvely, I'd expect a
function with that name to reveal whether the clockdomain supports
auto-idle mode, not to test its current state. So this function has been
renamed to 'clkdm_in_hwsup'. This name seems slightly uglier, but more
accurate.
Updated patch below, queued for 3.1 at git://git.pwsan.com/linux-2.6 in
the 'clockdomain_a_3.1' branch.
- Paul
From: Paul Walmsley <paul at pwsan.com>
Date: Thu, 7 Jul 2011 22:55:08 -0600
Subject: [PATCH] OMAP2+: clockdomain: add clkdm_in_hwsup()
Add a new function, clkdm_in_hwsup(), that returns true if a clockdomain
is configured for hardware-supervised idle. It does not actually read the
hardware; rather, it checks an internal flag in the struct clockdomain, which
is changed when the clockdomain is switched in and out of hardware-supervised
idle. This should be safe, since all changes to the idle mode should
pass through the clockdomain code.
Based on a set of patches by Rajendra Nayak <rnayak at ti.com> which do
the same thing by checking the hardware bits. This approach should be
faster and more compact.
Signed-off-by: Paul Walmsley <paul at pwsan.com>
Cc: Rajendra Nayak <rnayak at ti.com>
Cc: Todd Poynor <toddpoynor at google.com>
Cc: Benoît Cousson <b-cousson at ti.com>
---
arch/arm/mach-omap2/clockdomain.c | 30 ++++++++++++++++++++++++++++--
arch/arm/mach-omap2/clockdomain.h | 6 ++++++
2 files changed, 34 insertions(+), 2 deletions(-)
diff --git a/arch/arm/mach-omap2/clockdomain.c b/arch/arm/mach-omap2/clockdomain.c
index 5a57de5..239b558 100644
--- a/arch/arm/mach-omap2/clockdomain.c
+++ b/arch/arm/mach-omap2/clockdomain.c
@@ -1,8 +1,8 @@
/*
* OMAP2/3/4 clockdomain framework functions
*
- * Copyright (C) 2008-2010 Texas Instruments, Inc.
- * Copyright (C) 2008-2010 Nokia Corporation
+ * Copyright (C) 2008-2011 Texas Instruments, Inc.
+ * Copyright (C) 2008-2011 Nokia Corporation
*
* Written by Paul Walmsley and Jouni Högander
* Added OMAP4 specific support by Abhijit Pagare <abhijitpagare at ti.com>
@@ -704,6 +704,8 @@ int clkdm_sleep(struct clockdomain *clkdm)
pr_debug("clockdomain: forcing sleep on %s\n", clkdm->name);
+ clkdm->_flags &= ~_CLKDM_FLAG_HWSUP_ENABLED;
+
return arch_clkdm->clkdm_sleep(clkdm);
}
@@ -732,6 +734,8 @@ int clkdm_wakeup(struct clockdomain *clkdm)
pr_debug("clockdomain: forcing wakeup on %s\n", clkdm->name);
+ clkdm->_flags &= ~_CLKDM_FLAG_HWSUP_ENABLED;
+
return arch_clkdm->clkdm_wakeup(clkdm);
}
@@ -762,6 +766,8 @@ void clkdm_allow_idle(struct clockdomain *clkdm)
pr_debug("clockdomain: enabling automatic idle transitions for %s\n",
clkdm->name);
+ clkdm->_flags |= _CLKDM_FLAG_HWSUP_ENABLED;
+
arch_clkdm->clkdm_allow_idle(clkdm);
pwrdm_clkdm_state_switch(clkdm);
}
@@ -792,9 +798,29 @@ void clkdm_deny_idle(struct clockdomain *clkdm)
pr_debug("clockdomain: disabling automatic idle transitions for %s\n",
clkdm->name);
+ clkdm->_flags &= ~_CLKDM_FLAG_HWSUP_ENABLED;
+
arch_clkdm->clkdm_deny_idle(clkdm);
}
+/**
+ * clkdm_in_hwsup - is clockdomain @clkdm have hardware-supervised idle enabled?
+ * @clkdm: struct clockdomain *
+ *
+ * Returns true if clockdomain @clkdm currently has
+ * hardware-supervised idle enabled, or false if it does not or if
+ * @clkdm is NULL. It is only valid to call this function after
+ * clkdm_init() has been called. This function does not actually read
+ * bits from the hardware; it instead tests an in-memory flag that is
+ * changed whenever the clockdomain code changes the auto-idle mode.
+ */
+bool clkdm_in_hwsup(struct clockdomain *clkdm)
+{
+ if (!clkdm)
+ return false;
+
+ return (clkdm->_flags & _CLKDM_FLAG_HWSUP_ENABLED) ? true : false;
+}
/* Clockdomain-to-clock/hwmod framework interface code */
diff --git a/arch/arm/mach-omap2/clockdomain.h b/arch/arm/mach-omap2/clockdomain.h
index 8e0da64..8782a5c 100644
--- a/arch/arm/mach-omap2/clockdomain.h
+++ b/arch/arm/mach-omap2/clockdomain.h
@@ -83,6 +83,9 @@ struct clkdm_dep {
const struct omap_chip_id omap_chip;
};
+/* Possible flags for struct clockdomain._flags */
+#define _CLKDM_FLAG_HWSUP_ENABLED BIT(0)
+
/**
* struct clockdomain - OMAP clockdomain
* @name: clockdomain name
@@ -90,6 +93,7 @@ struct clkdm_dep {
* @clktrctrl_reg: CLKSTCTRL reg for the given clock domain
* @clktrctrl_mask: CLKTRCTRL/AUTOSTATE field mask in CM_CLKSTCTRL reg
* @flags: Clockdomain capability flags
+ * @_flags: Flags for use only by internal clockdomain code
* @dep_bit: Bit shift of this clockdomain's PM_WKDEP/CM_SLEEPDEP bit
* @prcm_partition: (OMAP4 only) PRCM partition ID for this clkdm's registers
* @cm_inst: (OMAP4 only) CM instance register offset
@@ -114,6 +118,7 @@ struct clockdomain {
} pwrdm;
const u16 clktrctrl_mask;
const u8 flags;
+ u8 _flags;
const u8 dep_bit;
const u8 prcm_partition;
const s16 cm_inst;
@@ -178,6 +183,7 @@ int clkdm_clear_all_sleepdeps(struct clockdomain *clkdm);
void clkdm_allow_idle(struct clockdomain *clkdm);
void clkdm_deny_idle(struct clockdomain *clkdm);
+bool clkdm_in_hwsup(struct clockdomain *clkdm);
int clkdm_wakeup(struct clockdomain *clkdm);
int clkdm_sleep(struct clockdomain *clkdm);
--
1.7.5.4
More information about the linux-arm-kernel
mailing list