On Tue, Mar 19, 2019 at 12:52:12PM +0530, Yash Shah wrote: > On Tue, Mar 19, 2019 at 3:16 AM Uwe Kleine-König > <u.kleine-koenig@xxxxxxxxxxxxxx> wrote: > > On Mon, Mar 18, 2019 at 05:17:14PM +0530, Yash Shah wrote: > > > + val = PWM_SIFIVE_PWMCFG_EN_ALWAYS | > > > + FIELD_PREP(PWM_SIFIVE_PWMCFG_SCALE, scale); > > > + writel(val, pwm->regs + PWM_SIFIVE_PWMCFG); > > > > Starting with this write the new period length might be active with the > > previous duty cycle. Is this worth a comment? I think the window where > > this can actually happen should be made as small as possible, so it > > would be great to first calculate both register values and then write > > them in two consecutive writels. > > The comment for this scenario has already been mentioned under the > limitation on top of this driver. > Anyway, I will try to implement your suggestion of consecutive writes Yeah, the comment at the top is for general information about the shortcomings. The comment here would be to say: The problem occurs *here*. Best regards Uwe -- Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | http://www.pengutronix.de/ |
