> -----Original Message----- > From: Uwe Kleine-König [mailto:u.kleine-koenig@xxxxxxxxxxxxxx] > Sent: Friday, November 30, 2018 3:35 PM > > Hello Fabio, > > On Fri, Nov 30, 2018 at 07:16:39PM -0200, Fabio Estevam wrote: > > [Adding Matt] > > > > On Fri, Nov 30, 2018 at 6:53 PM Uwe Kleine-König > > <u.kleine-koenig@xxxxxxxxxxxxxx> wrote: > > > > > > The status quo is: > > > - on i.MX25 the overcurrent polarity is always explicitly configured as > > > active high which matches the reset default. > > > - on i.MX6 and i.MX7 the overcurrent polarity is active high after > > > reset. usbmisc_imx6q_init() and usbmisc_imx7d_init() keep the current > > > state (probably as setup by the bootloader or still the reset > > > default) unless the polarity is explicitly configured as active high > > > which then is configured. (So if the pin is active low and the > > > bootloader didn't set this up the configuration is wrong.) > > > > > > To improve the situation always configure the reset default value unless > > > the device tree configures the polarity (and then use this one). > > > > > > Note that as the reset default is active high on all platforms there is > > > no need to check for the presence of "over-current-active-high". In the > > > absence of "over-current-active-low" it doesn't matter if active high is > > > configured because that's the default or because it is configured > > > explicitly. > > > > > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@xxxxxxxxxxxxxx> > > > > It seems that Matt is also interested in adding the > > 'over-current-active-low' property: > > https://patchwork.kernel.org/patch/10701311/ > > Looking at Matt's patch I noticed that I applied my patch on an old > kernel version (4.14) :-| > > But I think Matt's patch is broken because in usbmisc_imx6q_init() in > the last else branch MX6_BM_OVER_CUR_POLARITY isn't cleared. Other > than > that they don't look that different semantically. > > I will respin my patch on Monday. > Uwe, You are correct that I missed the clearing of MX6_BM_OVER_CUR_POLARITY. Also I don't see any reason why that last else that should clear MX6_BM_OVER_CUR_POLARITY and MX6_BM_OVER_CUR_DIS shouldn't also be in the usbmisc_imx7d_init function too. I just resubmitted my patch based on 4.20-rc4 with the mentioned changes. Matt Starr
