linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: mturquette@ti.com (Mike Turquette)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH 4/6] ARM: tegra: Add clk_tegra structure and helper functions
Date: Thu, 28 Jun 2012 11:28:00 -0700	[thread overview]
Message-ID: <20120628182759.GA28424@gmail.com> (raw)
In-Reply-To: <1340879846-12900-5-git-send-email-pgaikwad@nvidia.com>

On 20120628-16:07, Prashant Gaikwad wrote:
> Add Tegra platform specific clock structure clk_tegra and
> some helper functions for generic clock framework.
> 
> struct clk_tegra is the single strcture used for all types of
> clocks. reset and cfg_ex ops moved to clk_tegra from clk_ops.
> 
> Signed-off-by: Prashant Gaikwad <pgaikwad@nvidia.com>

Hi Prashant,

I'm happy to see Tegra getting ported over to the common clk framework.

<snip>
> +void tegra_clk_add(struct clk *clk)
> +{
> +	struct clk_tegra *c = to_clk_tegra(__clk_get_hw(clk));
> +
> +	mutex_lock(&clock_list_lock);
> +	list_add(&c->node, &clocks);
> +	mutex_unlock(&clock_list_lock);
> +}
> +
> +struct clk *tegra_get_clock_by_name(const char *name)
> +{
> +	struct clk_tegra *c;
> +	struct clk *ret = NULL;
> +	mutex_lock(&clock_list_lock);
> +	list_for_each_entry(c, &clocks, node) {
> +		if (strcmp(__clk_get_name(c->hw.clk), name) == 0) {
> +			ret = c->hw.clk;
> +			break;
> +		}
> +	}
> +	mutex_unlock(&clock_list_lock);
> +	return ret;
> +}
> +

Are you planning to continue using an internal list of struct clk_tegra?
OMAP had a similar list for the legacy clock framework but that was
since removed and now lookups are done solely through clkdev.

<snip>
> +void tegra_periph_reset_deassert(struct clk *c)
> +{
> +	struct clk_tegra *clk = to_clk_tegra(__clk_get_hw(c));
> +	BUG_ON(!clk->reset);
> +	clk->reset(__clk_get_hw(c), false);
> +}
> +EXPORT_SYMBOL(tegra_periph_reset_deassert);
> +
> +void tegra_periph_reset_assert(struct clk *c)
> +{
> +	struct clk_tegra *clk = to_clk_tegra(__clk_get_hw(c));
> +	BUG_ON(!clk->reset);
> +	clk->reset(__clk_get_hw(c), true);
> +}
> +EXPORT_SYMBOL(tegra_periph_reset_assert);
> +
> +/* Several extended clock configuration bits (e.g., clock routing, clock
> + * phase control) are included in PLL and peripheral clock source
> + * registers. */
> +int tegra_clk_cfg_ex(struct clk *c, enum tegra_clk_ex_param p, u32 setting)
> +{
> +	int ret = 0;
> +	struct clk_tegra *clk = to_clk_tegra(__clk_get_hw(c));
> +
> +	if (!clk->clk_cfg_ex) {
> +		ret = -ENOSYS;
> +		goto out;
> +	}
> +	ret = clk->clk_cfg_ex(__clk_get_hw(c), p, setting);
> +
> +out:
> +	return ret;
> +}

We had some discussions in the past on your clock reset and external
line request operations which you've had to put into struct clk_tegra.

Do you need to expose those ops to code in drivers/*?  I consider that a
reasonable litmus test to start considering if something should be moved
into the generic clk.h api.

<snip>
> +enum clk_state {
> +	UNINITIALIZED = 0,
> +	ON,
> +	OFF,
> +};

How is clk_state used?  Might it be helpful to you if that was in the
generic code in drivers/clk/clk.c?

<snip>
> +struct clk_tegra {
> +	/* node for master clocks list */
> +	struct list_head	node;	/* node for list of all clocks */
> +	struct clk_lookup	lookup;
> +	struct clk_hw		hw;
> +
> +	bool			set;
> +	unsigned long		fixed_rate;
> +	unsigned long		max_rate;
> +	unsigned long		min_rate;
> +	u32			flags;
> +	const char		*name;
> +
> +	enum clk_state		state;
> +	u32			div;
> +	u32			mul;
> +
> +	u32				reg;
> +	u32				reg_shift;
> +
> +	struct list_head		shared_bus_list;
> +
> +	union {
> +		struct {
> +			unsigned int			clk_num;
> +		} periph;
> +		struct {
> +			unsigned long			input_min;
> +			unsigned long			input_max;
> +			unsigned long			cf_min;
> +			unsigned long			cf_max;
> +			unsigned long			vco_min;
> +			unsigned long			vco_max;
> +			const struct clk_pll_freq_table	*freq_table;
> +			int				lock_delay;
> +			unsigned long			fixed_rate;
> +		} pll;
> +		struct {
> +			u32				sel;
> +			u32				reg_mask;
> +		} mux;
> +		struct {
> +			struct clk			*main;
> +			struct clk			*backup;
> +		} cpu;
> +		struct {
> +			struct list_head		node;
> +			bool				enabled;
> +			unsigned long			rate;
> +		} shared_bus_user;
> +	} u;
> +
> +	void (*reset)(struct clk_hw *, bool);
> +	int (*clk_cfg_ex)(struct clk_hw *, enum tegra_clk_ex_param, u32);
> +};

I know this is an initial effort but I hope in time that struct
clk_tegra can be broken up into separate clk types.  It would be even
better if some of the Tegra clocks code be converted over to the common
clk types in drivers/clk/clk-*.c

Regards,
Mike

  reply	other threads:[~2012-06-28 18:28 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-06-28 10:37 [PATCH 0/6] Port Tegra to generic clk framework Prashant Gaikwad
2012-06-28 10:37 ` [PATCH 1/6] ARM: tegra30: Separate out clk ops and clk data Prashant Gaikwad
2012-06-28 10:37 ` [PATCH 2/6] ARM: tegra20: " Prashant Gaikwad
2012-06-28 17:37   ` Stephen Warren
2012-06-29  2:06     ` Prashant G
2012-06-28 10:37 ` [PATCH 3/6] ARM: tegra: Rename tegra20 clock file Prashant Gaikwad
2012-06-28 10:37 ` [PATCH 4/6] ARM: tegra: Add clk_tegra structure and helper functions Prashant Gaikwad
2012-06-28 18:28   ` Mike Turquette [this message]
2012-06-29  8:34     ` Prashant Gaikwad
2012-07-03  0:09       ` Turquette, Mike
2012-07-03  1:31         ` Stephen Warren
2012-07-03  1:49           ` Turquette, Mike
2012-06-28 10:37 ` [PATCH 5/6] ARM: tegra: Port tegra to generic clock framework Prashant Gaikwad
2012-06-28 17:29   ` Prashant G
2012-06-28 10:37 ` [PATCH 6/6] ARM: tegra: Remove duplicate code Prashant Gaikwad
2012-06-28 17:39 ` [PATCH 0/6] Port Tegra to generic clk framework Stephen Warren
2012-06-29 17:24 ` Stephen Warren
2012-07-02 18:12   ` Stephen Warren
2012-07-03 11:29     ` Prashant Gaikwad

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20120628182759.GA28424@gmail.com \
    --to=mturquette@ti.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).