Just some small fixes for meson pwm.
Changelog:
v1->v2:
pwm: meson: Drop unneeded check in .get_state()
- update commit message
- drop Fixes tag. It's not actually a bug
previous version: [1]
pwm: meson: Add check for error from clk_round_rate()
- update commit message
- change unsigned long fin_freq; to long fin_freq;
previous version: [2]
[1] https://lore.kernel.org/lkml/[email protected]/T/#m6d1d5cf100dc5b77f85e2a7d01c0b0097b6a04da
[2] https://lore.kernel.org/lkml/[email protected]/T/#ma84e2d490bf79bda12f8393773c1af37b48d5473
George Stark (2):
pwm: meson: Drop unneeded check in .get_state()
pwm: meson: Add check for error from clk_round_rate()
drivers/pwm/pwm-meson.c | 12 +++++-------
1 file changed, 5 insertions(+), 7 deletions(-)
--
2.25.1
clk_round_rate() can return not only zero if requested frequency can not
be provided but also negative error code so add check for it too.
Also change type of variable holding clk_round_rate() result from
unsigned long to long. It's safe due to clk_round_rate() returns long.
Fixes: 329db102a26d ("pwm: meson: make full use of common clock framework")
Signed-off-by: Dmitry Rokosov <[email protected]>
Signed-off-by: George Stark <[email protected]>
---
drivers/pwm/pwm-meson.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/drivers/pwm/pwm-meson.c b/drivers/pwm/pwm-meson.c
index f4d70da621ec..d4228ac1e7ea 100644
--- a/drivers/pwm/pwm-meson.c
+++ b/drivers/pwm/pwm-meson.c
@@ -148,7 +148,7 @@ static int meson_pwm_calc(struct pwm_chip *chip, struct pwm_device *pwm,
struct meson_pwm *meson = to_meson_pwm(chip);
struct meson_pwm_channel *channel = &meson->channels[pwm->hwpwm];
unsigned int cnt, duty_cnt;
- unsigned long fin_freq;
+ long fin_freq;
u64 duty, period, freq;
duty = state->duty_cycle;
@@ -168,9 +168,10 @@ static int meson_pwm_calc(struct pwm_chip *chip, struct pwm_device *pwm,
freq = ULONG_MAX;
fin_freq = clk_round_rate(channel->clk, freq);
- if (fin_freq == 0) {
- dev_err(pwmchip_parent(chip), "invalid source clock frequency\n");
- return -EINVAL;
+ if (fin_freq <= 0) {
+ dev_err(pwmchip_parent(chip),
+ "invalid source clock frequency %llu\n", freq);
+ return fin_freq ? fin_freq : -EINVAL;
}
dev_dbg(pwmchip_parent(chip), "fin_freq: %lu Hz\n", fin_freq);
--
2.25.1
On Wed, Apr 24, 2024 at 04:24:08PM +0300, George Stark wrote:
> clk_round_rate() can return not only zero if requested frequency can not
> be provided but also negative error code so add check for it too.
>
> Also change type of variable holding clk_round_rate() result from
> unsigned long to long. It's safe due to clk_round_rate() returns long.
>
> Fixes: 329db102a26d ("pwm: meson: make full use of common clock framework")
> Signed-off-by: Dmitry Rokosov <[email protected]>
> Signed-off-by: George Stark <[email protected]>
> ---
> drivers/pwm/pwm-meson.c | 9 +++++----
> 1 file changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/pwm/pwm-meson.c b/drivers/pwm/pwm-meson.c
> index f4d70da621ec..d4228ac1e7ea 100644
> --- a/drivers/pwm/pwm-meson.c
> +++ b/drivers/pwm/pwm-meson.c
> @@ -148,7 +148,7 @@ static int meson_pwm_calc(struct pwm_chip *chip, struct pwm_device *pwm,
> struct meson_pwm *meson = to_meson_pwm(chip);
> struct meson_pwm_channel *channel = &meson->channels[pwm->hwpwm];
> unsigned int cnt, duty_cnt;
> - unsigned long fin_freq;
> + long fin_freq;
> u64 duty, period, freq;
>
> duty = state->duty_cycle;
> @@ -168,9 +168,10 @@ static int meson_pwm_calc(struct pwm_chip *chip, struct pwm_device *pwm,
> freq = ULONG_MAX;
>
> fin_freq = clk_round_rate(channel->clk, freq);
> - if (fin_freq == 0) {
> - dev_err(pwmchip_parent(chip), "invalid source clock frequency\n");
> - return -EINVAL;
> + if (fin_freq <= 0) {
> + dev_err(pwmchip_parent(chip),
> + "invalid source clock frequency %llu\n", freq);
> + return fin_freq ? fin_freq : -EINVAL;
> }
>
> dev_dbg(pwmchip_parent(chip), "fin_freq: %lu Hz\n", fin_freq);
^
With fin_freq now being unsigned, this ------------' should be a 'd'.
If you want to still extend your fixes series: A bit further below there
is
cnt = div_u64(fin_freq * period, NSEC_PER_SEC);
. The multiplication there might overflow. This should better use
mul_u64_u64_div_u64() (or one of it's variants).
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-K?nig |
Industrial Linux Solutions | https://www.pengutronix.de/ |