clk: tegra: move _calc_dynamic_ram_rate out of #ifdef
Subsystems (3)
PATCH 1/1 clk: tegra: move _calc_dynamic_ram_rate out of #ifdef
_calc_dynamic_ram_rate is defined inside an #ifdef but called later in the same file outside of that #ifdef, which can cause a build error:
drivers/clk/tegra/clk-pll.c: In function '_tegra_clk_register_pll':drivers/clk/tegra/clk-pll.c:1541:29: error: '_calc_dynamic_ramp_rate' undeclared (first use in this function)
This moves both _calc_dynamic_ram_rate and _pll_fixed_mdiv in front of the #ifdef to make all configurations compile again.
----
The patch that caused it appears to be older, but I only ran into the randconfig bug today for the first time. Apparently the commit is only in linux-next at the moment, not in 4.4-rc1
drivers/clk/tegra/clk-pll.c +34 -34
@@ -967,11 +967,6 @@ const struct clk_ops tegra_clk_plle_ops = { .enable = clk_plle_enable, }; -#if defined(CONFIG_ARCH_TEGRA_114_SOC ) || \ - defined(CONFIG_ARCH_TEGRA_124_SOC ) || \ - defined(CONFIG_ARCH_TEGRA_132_SOC ) || \ - defined(CONFIG_ARCH_TEGRA_210_SOC ) -static int _pll_fixed_mdiv (struct tegra_clk_pll_params *pll_params,unsigned long parent_rate) {
@@ -990,6 +985,40 @@ static int _pll_fixed_mdiv(struct tegra_clk_pll_params *pll_params,return 1 ; } +static int _calc_dynamic_ramp_rate (struct clk_hw *hw, +struct tegra_clk_pll_freq_table *cfg, +unsigned long rate,unsigned long parent_rate) +{ +struct tegra_clk_pll *pll =to_clk_pll (hw); +unsigned int p; +int p_div; + +if (!rate) +return -EINVAL ; + + p =DIV_ROUND_UP (pll->params->vco_min, rate); + cfg->m =_pll_fixed_mdiv (pll->params, parent_rate); + cfg->output_rate = rate * p; + cfg->n = cfg->output_rate * cfg->m / parent_rate; + cfg->input_rate = parent_rate; + + p_div =_p_div_to_hw (hw, p); +if (p_div <0 ) +return p_div; + + cfg->p = p_div; + +if (cfg->n >divn_max (pll) || cfg->output_rate > pll->params->vco_max) +return -EINVAL ; + +return 0 ; +} + +#if defined(CONFIG_ARCH_TEGRA_114_SOC ) || \ + defined(CONFIG_ARCH_TEGRA_124_SOC ) || \ + defined(CONFIG_ARCH_TEGRA_132_SOC ) || \ + defined(CONFIG_ARCH_TEGRA_210_SOC ) +u16 tegra_pll_get_fixed_mdiv (struct clk_hw *hw,unsigned long input_rate) {struct tegra_clk_pll *pll =to_clk_pll (hw);
@@ -1039,35 +1068,6 @@ static int _setup_dynamic_ramp(struct tegra_clk_pll_params *pll_params,return 0 ; } -static int _calc_dynamic_ramp_rate (struct clk_hw *hw, -struct tegra_clk_pll_freq_table *cfg, -unsigned long rate,unsigned long parent_rate) -{ -struct tegra_clk_pll *pll =to_clk_pll (hw); -unsigned int p; -int p_div; - -if (!rate) -return -EINVAL ; - - p =DIV_ROUND_UP (pll->params->vco_min, rate); - cfg->m =_pll_fixed_mdiv (pll->params, parent_rate); - cfg->output_rate = rate * p; - cfg->n = cfg->output_rate * cfg->m / parent_rate; - cfg->input_rate = parent_rate; - - p_div =_p_div_to_hw (hw, p); -if (p_div <0 ) -return p_div; - - cfg->p = p_div; - -if (cfg->n >divn_max (pll) || cfg->output_rate > pll->params->vco_max) -return -EINVAL ; - -return 0 ; -} -static int _pll_ramp_calc_pll (struct clk_hw *hw,struct tegra_clk_pll_freq_table *cfg,unsigned long rate,unsigned long parent_rate)
Quoted 10 lines · Arnd Bergmann
_calc_dynamic_ram_rateis defined inside an #ifdef but called later in the same file outside of that #ifdef, which can cause a build error:
drivers/clk/tegra/clk-pll.c: In function '_tegra_clk_register_pll':drivers/clk/tegra/clk-pll.c:1541:29: error: '_calc_dynamic_ramp_rate' undeclared (first use in this function)This moves both
_calc_dynamic_ram_rateand_pll_fixed_mdivin front of the #ifdef to make all configurations compile again.
Quoted 7 lines
Signed-off-by Arnd Bergmann <[email protected]>Fixes 44c8b9fa432c ("clk: tegra: pll: Fix _pll_ramp_calc_pll logic and _calc_dynamic_ramp_rate")----
The patch that caused it appears to be older, but I only ran into the randconfig bug today for the first time. Apparently the commit is only in linux-next at the moment, not in 4.4-rc1
-rhyland
Signature · 1 line
nvpublic
Quoted 10 lines · Arnd Bergmann
_calc_dynamic_ram_rateis defined inside an #ifdef but called later in the same file outside of that #ifdef, which can cause a build error:
drivers/clk/tegra/clk-pll.c: In function '_tegra_clk_register_pll':drivers/clk/tegra/clk-pll.c:1541:29: error: '_calc_dynamic_ramp_rate' undeclared (first use in this function)This moves both
_calc_dynamic_ram_rateand_pll_fixed_mdivin front of the #ifdef to make all configurations compile again.
Quoted 6 lines
Signed-off-by Arnd Bergmann <[email protected]>Fixes 44c8b9fa432c ("clk: tegra: pll: Fix _pll_ramp_calc_pll logic and _calc_dynamic_ramp_rate")----
The patch that caused it appears to be older, but I only ran into the randconfig bug today for the first time. Apparently the commit is only in linux-next at the moment, not in 4.4-rc1
Hmm... this looks familiar, I thought I had included an equivalent patch in the clock tree. I'll go do some digging, but this looks exactly like what I had done at the time.
Thanks,
Thierry
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20151120/ef3499f5/attachment.sig>