Hello,
On Mon, Apr 24, 2023 at 03:16:34AM +0530, Naresh Kamboju wrote:
> Following build errors noticed on arm64 and arm while building stable-rc 5.4.
>
> Reported-by: Linux Kernel Functional Testing <lkft@xxxxxxxxxx>
>
> Build error:
> -----------
> drivers/pwm/pwm-meson.c: In function 'meson_pwm_apply':
> drivers/pwm/pwm-meson.c:313:25: error: assignment of member
> 'polarity' in read-only object
> 313 | state->polarity = PWM_POLARITY_NORMAL;
> | ^
> make[3]: *** [scripts/Makefile.build:262: drivers/pwm/pwm-meson.o] Error 1
> make[3]: Target '__build' not remade because of errors.
> make[2]: *** [scripts/Makefile.build:497: drivers/pwm] Error 2
>
>
> Suspected commit:
> ------------
> pwm: meson: Explicitly set .polarity in .get_state()
> commit 8caa81eb950cb2e9d2d6959b37d853162d197f57 upstream.
Thanks for the report. I sent a fix to that patch in reply to the patch
being announced to be picked for 5.4.y, but failed to Cc you. I just
replied to the mail, so it didn't even go to stable@xxxxxxxxxxxxxxx.
For reference, here comes the fixed patch again, see below.
Best regards
Uwe
------->8--------
From: Uwe Kleine-König <u.kleine-koenig@xxxxxxxxxxxxxx>
Date: Wed, 22 Mar 2023 22:45:44 +0100
Subject: [PATCH] pwm: meson: Explicitly set .polarity in .get_state()
[ Upstream commit 8caa81eb950cb2e9d2d6959b37d853162d197f57 ]
The driver only supports normal polarity. Complete the implementation of
.get_state() by setting .polarity accordingly.
This fixes a regression that was possible since commit c73a3107624d
("pwm: Handle .get_state() failures") which stopped to zero-initialize
the state passed to the .get_state() callback. This was reported at
https://forum.odroid.com/viewtopic.php?f=177&t=46360 . While this was an
unintended side effect, the real issue is the driver's callback not
setting the polarity.
There is a complicating fact, that the .apply() callback fakes support
for inversed polarity. This is not (and cannot) be matched by
.get_state(). As fixing this isn't easy, only point it out in a comment
to prevent authors of other drivers from copying that approach.
Fixes: c375bcbaabdb ("pwm: meson: Read the full hardware state in meson_pwm_get_state()")
Reported-by: Munehisa Kamata <kamatam@xxxxxxxxxx>
Acked-by: Martin Blumenstingl <martin.blumenstingl@xxxxxxxxxxxxxx>
Link: https://lore.kernel.org/r/20230310191405.2606296-1-u.kleine-koenig@xxxxxxxxxxxxxx
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@xxxxxxxxxxxxxx>
Signed-off-by: Thierry Reding <thierry.reding@xxxxxxxxx>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@xxxxxxxxxxxxxx>
---
drivers/pwm/pwm-meson.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/drivers/pwm/pwm-meson.c b/drivers/pwm/pwm-meson.c
index 6245bbdb6e6c..3c81858a8261 100644
--- a/drivers/pwm/pwm-meson.c
+++ b/drivers/pwm/pwm-meson.c
@@ -168,6 +168,12 @@ static int meson_pwm_calc(struct meson_pwm *meson, struct pwm_device *pwm,
duty = state->duty_cycle;
period = state->period;
+ /*
+ * Note this is wrong. The result is an output wave that isn't really
+ * inverted and so is wrongly identified by .get_state as normal.
+ * Fixing this needs some care however as some machines might rely on
+ * this.
+ */
if (state->polarity == PWM_POLARITY_INVERSED)
duty = period - duty;
@@ -366,6 +372,7 @@ static void meson_pwm_get_state(struct pwm_chip *chip, struct pwm_device *pwm,
state->period = 0;
state->duty_cycle = 0;
}
+ state->polarity = PWM_POLARITY_NORMAL;
}
static const struct pwm_ops meson_pwm_ops = {
--
2.39.2
--
Pengutronix e.K. | Uwe Kleine-König |
Industrial Linux Solutions | https://www.pengutronix.de/ |
Attachment:
signature.asc
Description: PGP signature
