| Message ID | 1412845852-26157-1-git-send-email-jingchang.lu@freescale.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 10/09/2014 02:10 AM, Jingchang Lu wrote: > Mandatorily disabling the uart clock will cause register access > hung on "no_console_suspend". This patch add condition check on > it and only disable the clock with console_suspend_enabled true. It would be useful if the email/patch subject line said "V2", and there was a changelog below. > Signed-off-by: Joseph Lo <josephl@nvidia.com> I don't think he did; he just suggested a small fix. > Signed-off-by: Jingchang Lu <jingchang.lu@freescale.com> > --- > drivers/tty/serial/of_serial.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/tty/serial/of_serial.c b/drivers/tty/serial/of_serial.c ...
>-----Original Message----- >From: Stephen Warren [mailto:swarren@wwwdotorg.org] >Sent: Friday, October 10, 2014 1:05 PM >To: Lu Jingchang-B35083; gregkh@linuxfoundation.org >Cc: arnd@arndb.de; Joseph Lo; linux-kernel@vger.kernel.org; linux- >serial@vger.kernel.org; linux-arm-kernel@lists.infradead.org >Subject: Re: [PATCH] serial: of-serial: fix up PM ops on >no_console_suspend > >On 10/09/2014 02:10 AM, Jingchang Lu wrote: >> Mandatorily disabling the uart clock will cause register access hung >> on "no_console_suspend". This patch add condition check on it and only >> disable the clock with console_suspend_enabled true. > >It would be useful if the email/patch subject line said "V2", and there >was a changelog below. This is not a v2, patch "serial: of-serial: add PM suspend/resume support (commit 2dea53bf57783f243c892e99c10c6921e956aa7e)" has been merged into linux-next tree, this is a fixup on that PM ops. Thanks. Best Regards, Jingchang
On 10/10/2014 02:16 AM, Jingchang Lu wrote: >> From: Stephen Warren [mailto:swarren@wwwdotorg.org] >> On 10/09/2014 02:10 AM, Jingchang Lu wrote: >>> Mandatorily disabling the uart clock will cause register access hung >>> on "no_console_suspend". This patch add condition check on it and only >>> disable the clock with console_suspend_enabled true. >> >> It would be useful if the email/patch subject line said "V2", and there >> was a changelog below. > This is not a v2, patch "serial: of-serial: add PM suspend/resume support > (commit 2dea53bf57783f243c892e99c10c6921e956aa7e)" has been merged into > linux-next tree, this is a fixup on that PM ops. Then please say so in the commit log. Something like, Fix commit 2dea53bf57783f243c892e99c10c6921e956aa7e, "serial: of-serial: add PM suspend/resume support", which disables the uart clock on suspend, but also causes a hardware hang on register access if no_console_suspend command line option is used. Regards, Peter Hurley
Hi Jingchang, On 10/09/2014 05:10 AM, Jingchang Lu wrote: > Mandatorily disabling the uart clock will cause register access > hung on "no_console_suspend". This patch add condition check on > it and only disable the clock with console_suspend_enabled true. > > Signed-off-by: Joseph Lo <josephl@nvidia.com> > Signed-off-by: Jingchang Lu <jingchang.lu@freescale.com> > --- > drivers/tty/serial/of_serial.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/tty/serial/of_serial.c b/drivers/tty/serial/of_serial.c > index 8bc2563..a765399 100644 > --- a/drivers/tty/serial/of_serial.c > +++ b/drivers/tty/serial/of_serial.c > @@ -246,7 +246,7 @@ static int of_serial_suspend(struct device *dev) > struct of_serial_info *info = dev_get_drvdata(dev); > > serial8250_suspend_port(info->line); > - if (info->clk) > + if (info->clk && console_suspend_enabled) > clk_disable_unprepare(info->clk); This needs to check that _this_ port is the console port, like: if (info->clk && uart_console(port) && and console_suspend_enabled) clk_disable_unprepare(info->clk); which means you need something like: struct uart_8250_port *uart = serial8250_get_port(info->line); struct uart_port *port = &uart->port; Also, there's another problem with your original patch: not every of_serial device is an 8250, right? serial8250_suspend_port() on a port that's not registered with 8250 core is probably going to blow up. Which means the body of of_platform_suspend() and of_platform_resume() should probably look like switch (info->type) { #ifdef CONFIG_SERIAL_8250 case PORT_8250 ... PORT_MAX_8250: { /* current body of function */ } #endif default: } Regards, Peter Hurley > > return 0; > @@ -256,7 +256,7 @@ static int of_serial_resume(struct device *dev) > { > struct of_serial_info *info = dev_get_drvdata(dev); > > - if (info->clk) > + if (info->clk && console_suspend_enabled) > clk_prepare_enable(info->clk); > > serial8250_resume_port(info->line); >
>-----Original Message----- >From: Peter Hurley [mailto:peter@hurleysoftware.com] >Sent: Friday, October 10, 2014 7:39 PM >To: Lu Jingchang-B35083; gregkh@linuxfoundation.org >Cc: arnd@arndb.de; linux-serial@vger.kernel.org; linux-arm- >kernel@lists.infradead.org; linux-kernel@vger.kernel.org; Joseph Lo >Subject: Re: [PATCH] serial: of-serial: fix up PM ops on >no_console_suspend > >Hi Jingchang, > >On 10/09/2014 05:10 AM, Jingchang Lu wrote: >> Mandatorily disabling the uart clock will cause register access hung >> on "no_console_suspend". This patch add condition check on it and only >> disable the clock with console_suspend_enabled true. >> >> Signed-off-by: Joseph Lo <josephl@nvidia.com> >> Signed-off-by: Jingchang Lu <jingchang.lu@freescale.com> >> --- >> drivers/tty/serial/of_serial.c | 4 ++-- >> 1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/drivers/tty/serial/of_serial.c >> b/drivers/tty/serial/of_serial.c index 8bc2563..a765399 100644 >> --- a/drivers/tty/serial/of_serial.c >> +++ b/drivers/tty/serial/of_serial.c >> @@ -246,7 +246,7 @@ static int of_serial_suspend(struct device *dev) >> struct of_serial_info *info = dev_get_drvdata(dev); >> >> serial8250_suspend_port(info->line); >> - if (info->clk) >> + if (info->clk && console_suspend_enabled) >> clk_disable_unprepare(info->clk); > >This needs to check that _this_ port is the console port, like: > > if (info->clk && uart_console(port) && and console_suspend_enabled) > clk_disable_unprepare(info->clk); > >which means you need something like: > > struct uart_8250_port *uart = serial8250_get_port(info->line); > struct uart_port *port = &uart->port; > >Also, there's another problem with your original patch: >not every of_serial device is an 8250, right? >serial8250_suspend_port() on a port that's not registered with 8250 core >is probably going to blow up. > >Which means the body of of_platform_suspend() and of_platform_resume() >should probably look like > > switch (info->type) { >#ifdef CONFIG_SERIAL_8250 > case PORT_8250 ... PORT_MAX_8250: > { > /* current body of function */ > } >#endif > default: > } > Yes, this should be considered, then serial8250_get_port(info->line) for uart_console() also should be put in this type switch, Thanks. Best Regards, Jingchang >Regards, >Peter Hurley > > >> >> return 0; >> @@ -256,7 +256,7 @@ static int of_serial_resume(struct device *dev) { >> struct of_serial_info *info = dev_get_drvdata(dev); >> >> - if (info->clk) >> + if (info->clk && console_suspend_enabled) >> clk_prepare_enable(info->clk); >> >> serial8250_resume_port(info->line); >>
diff --git a/drivers/tty/serial/of_serial.c b/drivers/tty/serial/of_serial.c index 8bc2563..a765399 100644 --- a/drivers/tty/serial/of_serial.c +++ b/drivers/tty/serial/of_serial.c @@ -246,7 +246,7 @@ static int of_serial_suspend(struct device *dev) struct of_serial_info *info = dev_get_drvdata(dev); serial8250_suspend_port(info->line); - if (info->clk) + if (info->clk && console_suspend_enabled) clk_disable_unprepare(info->clk); return 0; @@ -256,7 +256,7 @@ static int of_serial_resume(struct device *dev) { struct of_serial_info *info = dev_get_drvdata(dev); - if (info->clk) + if (info->clk && console_suspend_enabled) clk_prepare_enable(info->clk); serial8250_resume_port(info->line);