Re: [PATCH v5 16/21] clk: tegra: pll: Add Set_default logic
From: Rhyland Klein
Date: Wed May 20 2015 - 13:21:12 EST
On 5/13/2015 8:04 PM, Benson Leung wrote:
> On Tue, May 12, 2015 at 10:23 AM, Rhyland Klein <rklein@xxxxxxxxxx> wrote:
>> From: Bill Huang <bilhuang@xxxxxxxxxx>
>>
>> Add logic which (if specified for a pll) can verify that a PLL is set
>> to the proper default value and if not can set it. This can be
>> specified per PLL as each will have different default values.
>>
>> Based on original work by Aleksandr Frid <afrid@xxxxxxxxxx>
>>
>> Signed-off-by: Bill Huang <bilhuang@xxxxxxxxxx>
>> ---
>> v2:
>> - Remove MACRO for PLL_MISC_CHECK_DEFAULT as suggested, and instead
>> the tegra210 driver will include an inline version of this function.
>>
>> drivers/clk/tegra/clk-pll.c | 46 ++++++++++++++++++++++++++++++++-----------
>> drivers/clk/tegra/clk.h | 2 ++
>> 2 files changed, 37 insertions(+), 11 deletions(-)
>>
>> diff --git a/drivers/clk/tegra/clk-pll.c b/drivers/clk/tegra/clk-pll.c
>> index 626466665dde..36aa2a95fac0 100644
>> --- a/drivers/clk/tegra/clk-pll.c
>> +++ b/drivers/clk/tegra/clk-pll.c
>> @@ -652,13 +652,26 @@ static int _program_pll(struct clk_hw *hw, struct tegra_clk_pll_freq_table *cfg,
>> unsigned long rate)
>> {
>> struct tegra_clk_pll *pll = to_clk_pll(hw);
>> + struct tegra_clk_pll_freq_table old_cfg;
>> int state, ret = 0;
>>
>> state = clk_pll_is_enabled(hw);
>>
>> + _get_pll_mnp(pll, &old_cfg);
>> +
>> + if (state && pll->params->defaults_set && pll->params->dyn_ramp &&
>> + (cfg->m == old_cfg.m) && (cfg->p == old_cfg.p)) {
>> + ret = pll->params->dyn_ramp(pll, cfg);
>> + if (!ret)
>> + return 0;
>> + }
>> +
>
> This block above depends on defaults_set, but seems to be some
> additional dyn_ramp logic that is different from what the commit
> message describes of this patch, namely, setting defaults if
> defaults_set is not true.
>
> Could you perhaps reverse the order of this patch [16/21] and the
> preceding one [15/21] clk: tegra: pll: Add dyn_ramp callback, and move
> this chunk of code into the dyn_ramp patch?
Will do.
>
>> if (state)
>> _clk_pll_disable(hw);
>>
>> + if (!pll->params->defaults_set && pll->params->set_defaults)
>> + pll->params->set_defaults(pll);
>> +
>> _update_pll_mnp(pll, cfg);
>>
>> if (pll->params->flags & TEGRA_PLL_HAS_CPCON)
>> @@ -1517,6 +1530,9 @@ static struct clk *_tegra_clk_register_pll(struct tegra_clk_pll *pll,
>> if (!pll->params->calc_rate)
>> pll->params->calc_rate = _calc_rate;
>>
>> + if (pll->params->set_defaults)
>> + pll->params->set_defaults(pll);
>> +
>> /* Data in .init is copied by clk_register(), so stack variable OK */
>> pll->hw.init = &init;
>>
>> @@ -1636,7 +1652,6 @@ struct clk *tegra_clk_register_pllxc(const char *name, const char *parent_name,
>> struct tegra_clk_pll *pll;
>> struct clk *clk, *parent;
>> unsigned long parent_rate;
>> - int err;
>> u32 val, val_iddq;
>>
>> parent = __clk_lookup(parent_name);
>> @@ -1657,18 +1672,27 @@ struct clk *tegra_clk_register_pllxc(const char *name, const char *parent_name,
>> pll_params->vco_min = pll_params->adjust_vco(pll_params,
>> parent_rate);
>>
>> - err = _setup_dynamic_ramp(pll_params, clk_base, parent_rate);
>> - if (err)
>> - return ERR_PTR(err);
>> + /*
>> + * If the pll has a set_defaults callback, it will take care of
>> + * configuring dynamic ramping and setting IDDQ in that path.
>> + */
>> + if (!pll_params->set_defaults) {
>> + int err;
>>
>> - val = readl_relaxed(clk_base + pll_params->base_reg);
>> - val_iddq = readl_relaxed(clk_base + pll_params->iddq_reg);
>> + err = _setup_dynamic_ramp(pll_params, clk_base, parent_rate);
>> + if (err)
>> + return ERR_PTR(err);
>>
>> - if (val & PLL_BASE_ENABLE)
>> - WARN_ON(val_iddq & BIT(pll_params->iddq_bit_idx));
>> - else {
>> - val_iddq |= BIT(pll_params->iddq_bit_idx);
>> - writel_relaxed(val_iddq, clk_base + pll_params->iddq_reg);
>> + val = readl_relaxed(clk_base + pll_params->base_reg);
>> + val_iddq = readl_relaxed(clk_base + pll_params->iddq_reg);
>> +
>> + if (val & PLL_BASE_ENABLE)
>> + WARN_ON(val_iddq & BIT(pll_params->iddq_bit_idx));
>> + else {
>> + val_iddq |= BIT(pll_params->iddq_bit_idx);
>> + writel_relaxed(val_iddq,
>> + clk_base + pll_params->iddq_reg);
>> + }
>> }
>>
>> pll = _tegra_init_pll(clk_base, pmc, pll_params, lock);
>> diff --git a/drivers/clk/tegra/clk.h b/drivers/clk/tegra/clk.h
>> index a61d388364b3..38b4c95cfb2f 100644
>> --- a/drivers/clk/tegra/clk.h
>> +++ b/drivers/clk/tegra/clk.h
>> @@ -261,6 +261,7 @@ struct tegra_clk_pll_params {
>> int stepb_shift;
>> int lock_delay;
>> int max_p;
>> + bool defaults_set;
>> struct pdiv_map *pdiv_tohw;
>> struct div_nmp *div_nmp;
>> struct tegra_clk_pll_freq_table *freq_table;
>> @@ -275,6 +276,7 @@ struct tegra_clk_pll_params {
>> unsigned long parent_rate);
>> int (*dyn_ramp)(struct tegra_clk_pll *pll,
>> struct tegra_clk_pll_freq_table *cfg);
>> + void (*set_defaults)(struct tegra_clk_pll *pll);
>
> Kerneldoc for the two members, please. This is especially useful since
> the two members have such similar names, but they mean different
> things.
>
Yah I thought I got all the new members, but apparently I got distracted
part way through the series and missed some. I'll double check for v6
that ALL members are kerneldoc'd.
-rhyland
--
nvpublic
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/