RE: [PATCH] serial: qcom_geni_serial: Setup serial port after Deep sleep

[Mehul Raninga (Temp) (QUIC)]

Hello Andersson,
Thanks for the review. Kindly find my reply inline below

> -----Original Message-----
> From: Bjorn Andersson <andersson@xxxxxxxxxx>
> Sent: Tuesday, May 30, 2023 8:37 PM
> To: Mehul Raninga (Temp) (QUIC) <quic_mraninga@xxxxxxxxxxx>
> Cc: agross@xxxxxxxxxx; konrad.dybcio@xxxxxxxxxx; gregkh@xxxxxxxxxxxxxxxxxxx;
> jirislaby@xxxxxxxxxx; linux-arm-msm@xxxxxxxxxxxxxxx; linux-
> serial@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; Viken Dadhaniya (QUIC)
> <quic_vdadhani@xxxxxxxxxxx>; Visweswara Tanuku (QUIC)
> <quic_vtanuku@xxxxxxxxxxx>; Vijaya Krishna Nivarthi (Temp) (QUIC)
> <quic_vnivarth@xxxxxxxxxxx>
> Subject: Re: [PATCH] serial: qcom_geni_serial: Setup serial port after Deep sleep
> 
> WARNING: This email originated from outside of Qualcomm. Please be wary of
> any links or attachments, and do not enable macros.
> 
> On Tue, May 30, 2023 at 04:45:57PM +0530, Mehul Raninga wrote:
> > While exiting deep sleep, serial port loses its configuration hence it
> > prints garbage characters on console.
> 
> Presumably it lost its configuration sometime after suspend, rather than while
> resuming the system?

I will reword commit message in V2 as below:
Serial port lost its configuration sometime after suspend hence it prints garbage characters on console after resuming.
> 
> >
> > Set serial port configuration while resume from deep sleep.
> >
> 
> What happens if you do this unconditionally?

pm_suspend_via_firmware returns true indicating system is resuming from deepsleep. In case we are not resuming from deepsleep, this serial port setup is not required.
> 
> > Signed-off-by: Mehul Raninga <quic_mraninga@xxxxxxxxxxx>
> > ---
> >  drivers/tty/serial/qcom_geni_serial.c | 3 +++
> >  1 file changed, 3 insertions(+)
> >
> > diff --git a/drivers/tty/serial/qcom_geni_serial.c
> > b/drivers/tty/serial/qcom_geni_serial.c
> > index 8582479f0211..c04b8fec30ba 100644
> > --- a/drivers/tty/serial/qcom_geni_serial.c
> > +++ b/drivers/tty/serial/qcom_geni_serial.c
> > @@ -20,6 +20,7 @@
> >  #include <linux/serial.h>
> >  #include <linux/serial_core.h>
> >  #include <linux/slab.h>
> > +#include <linux/suspend.h>
> >  #include <linux/tty.h>
> >  #include <linux/tty_flip.h>
> >  #include <dt-bindings/interconnect/qcom,icc.h>
> > @@ -1737,6 +1738,8 @@ static int qcom_geni_serial_sys_resume(struct
> device *dev)
> >       if (uart_console(uport)) {
> >               geni_icc_set_tag(&port->se, QCOM_ICC_TAG_ALWAYS);
> >               geni_icc_set_bw(&port->se);
> > +             if (pm_suspend_via_firmware())
> 
> I'm not familiar with this api, but aren't all our systems implementing firmware-
> assisted suspend?

Not all the platform supports deep sleep hence to differentiate if resume is from deep sleep suspend or normal suspend, this api is required.
> 
> Regards,
> Bjorn
> 
> > +                     qcom_geni_serial_port_setup(uport);
> >       }
> >       return ret;
> >  }
> > --
> > 2.17.1
> >