| Message ID | 1373924770-22599-1-git-send-email-gsi@denx.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, Jul 15, 2013 at 11:46:01PM +0200, Gerhard Sittig wrote: > with device tree based clock lookup, the MCLK name no longer > depends on the PSC index > > Signed-off-by: Gerhard Sittig <gsi@denx.de> > --- > drivers/tty/serial/mpc52xx_uart.c | 8 ++------ > 1 file changed, 2 insertions(+), 6 deletions(-) > > diff --git a/drivers/tty/serial/mpc52xx_uart.c b/drivers/tty/serial/mpc52xx_uart.c > index 53c1093..221fb89 100644 > --- a/drivers/tty/serial/mpc52xx_uart.c > +++ b/drivers/tty/serial/mpc52xx_uart.c > @@ -619,21 +619,17 @@ static irqreturn_t mpc512x_psc_handle_irq(struct uart_port *port) > static int mpc512x_psc_clock(struct uart_port *port, int enable) > { > struct clk *psc_clk; > - int psc_num; > - char clk_name[10]; > > if (uart_console(port)) > return 0; > > - psc_num = (port->mapbase & 0xf00) >> 8; > - snprintf(clk_name, sizeof(clk_name), "psc%d_mclk", psc_num); > - psc_clk = clk_get(port->dev, clk_name); > + psc_clk = clk_get(port->dev, "mclk"); Same comment applies here as Mark made to the spi driver. Sascha
On Mon, Jul 15, 2013 at 23:54 +0200, Sascha Hauer wrote: > > On Mon, Jul 15, 2013 at 11:46:01PM +0200, Gerhard Sittig wrote: > > with device tree based clock lookup, the MCLK name no longer > > depends on the PSC index > > > > Signed-off-by: Gerhard Sittig <gsi@denx.de> > > --- > > drivers/tty/serial/mpc52xx_uart.c | 8 ++------ > > 1 file changed, 2 insertions(+), 6 deletions(-) > > > > diff --git a/drivers/tty/serial/mpc52xx_uart.c b/drivers/tty/serial/mpc52xx_uart.c > > index 53c1093..221fb89 100644 > > --- a/drivers/tty/serial/mpc52xx_uart.c > > +++ b/drivers/tty/serial/mpc52xx_uart.c > > @@ -619,21 +619,17 @@ static irqreturn_t mpc512x_psc_handle_irq(struct uart_port *port) > > static int mpc512x_psc_clock(struct uart_port *port, int enable) > > { > > struct clk *psc_clk; > > - int psc_num; > > - char clk_name[10]; > > > > if (uart_console(port)) > > return 0; > > > > - psc_num = (port->mapbase & 0xf00) >> 8; > > - snprintf(clk_name, sizeof(clk_name), "psc%d_mclk", psc_num); > > - psc_clk = clk_get(port->dev, clk_name); > > + psc_clk = clk_get(port->dev, "mclk"); > > Same comment applies here as Mark made to the spi driver. So I'd like to respond in the same way as I did for the SPI driver. :) The scope of this series is the introduction of support for the common clock framework. Addressing other (legacy non-fatal and previously accepted) issues as they get identified in bypassing would be the scope of a separate patch or series. I'm not questioning the need to fix other additionally identified issues. I'm just asking whether they shall be in the scope of this very series. virtually yours Gerhard Sittig
diff --git a/drivers/tty/serial/mpc52xx_uart.c b/drivers/tty/serial/mpc52xx_uart.c index 53c1093..221fb89 100644 --- a/drivers/tty/serial/mpc52xx_uart.c +++ b/drivers/tty/serial/mpc52xx_uart.c @@ -619,21 +619,17 @@ static irqreturn_t mpc512x_psc_handle_irq(struct uart_port *port) static int mpc512x_psc_clock(struct uart_port *port, int enable) { struct clk *psc_clk; - int psc_num; - char clk_name[10]; if (uart_console(port)) return 0; - psc_num = (port->mapbase & 0xf00) >> 8; - snprintf(clk_name, sizeof(clk_name), "psc%d_mclk", psc_num); - psc_clk = clk_get(port->dev, clk_name); + psc_clk = clk_get(port->dev, "mclk"); if (IS_ERR(psc_clk)) { dev_err(port->dev, "Failed to get PSC clock entry!\n"); return -ENODEV; } - dev_dbg(port->dev, "%s %sable\n", clk_name, enable ? "en" : "dis"); + dev_dbg(port->dev, "mclk %sable\n", enable ? "en" : "dis"); if (enable) clk_prepare_enable(psc_clk);
with device tree based clock lookup, the MCLK name no longer depends on the PSC index Signed-off-by: Gerhard Sittig <gsi@denx.de> --- drivers/tty/serial/mpc52xx_uart.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-)