| Message ID | 20170302173114.28508-4-alexandre.belloni@free-electrons.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hei hei, there's a typo in the subject you may want to fix in a v2: watchodg → watchdog Greets Alex
Hello,
On Thu, 2 Mar 2017 18:31:13 +0100, Alexandre Belloni wrote:
> + irq = irq_of_parse_and_map(pdev->dev.of_node, 0);
Any reason to use irq_of_parse_and_map() over the more conventional
platform_get_irq() ?
Best regards,
Thomas
On 03/03/2017 at 09:03:25 +0100, Thomas Petazzoni wrote: > On Thu, 2 Mar 2017 18:31:13 +0100, Alexandre Belloni wrote: > > > + irq = irq_of_parse_and_map(pdev->dev.of_node, 0); > > Any reason to use irq_of_parse_and_map() over the more conventional > platform_get_irq() ? > No particular reason but I'm just removing the if (pdev->dev.of_node)
On 03/03/2017 01:29 AM, Alexandre Belloni wrote: > On 03/03/2017 at 09:03:25 +0100, Thomas Petazzoni wrote: >> On Thu, 2 Mar 2017 18:31:13 +0100, Alexandre Belloni wrote: >> >>> + irq = irq_of_parse_and_map(pdev->dev.of_node, 0); >> >> Any reason to use irq_of_parse_and_map() over the more conventional >> platform_get_irq() ? >> > > No particular reason but I'm just removing the if (pdev->dev.of_node) > A function call change would (should) be a separate patch. Guenter
On 03/02/2017 09:31 AM, Alexandre Belloni wrote: > Because the only way to use the driver is to have a device tree enabling > it, pdev->dev.of_node will never be NULL. Remove the unnecessary check. > > Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> > --- > drivers/watchdog/sama5d4_wdt.c | 14 ++++++-------- > 1 file changed, 6 insertions(+), 8 deletions(-) > > diff --git a/drivers/watchdog/sama5d4_wdt.c b/drivers/watchdog/sama5d4_wdt.c > index 362fd229786d..d710014f3b7d 100644 > --- a/drivers/watchdog/sama5d4_wdt.c > +++ b/drivers/watchdog/sama5d4_wdt.c > @@ -228,15 +228,13 @@ static int sama5d4_wdt_probe(struct platform_device *pdev) > > wdt->reg_base = regs; > > - if (pdev->dev.of_node) { > - irq = irq_of_parse_and_map(pdev->dev.of_node, 0); > - if (!irq) > - dev_warn(&pdev->dev, "failed to get IRQ from DT\n"); > + irq = irq_of_parse_and_map(pdev->dev.of_node, 0); > + if (!irq) > + dev_warn(&pdev->dev, "failed to get IRQ from DT\n"); > > - ret = of_sama5d4_wdt_init(pdev->dev.of_node, wdt); > - if (ret) > - return ret; > - } > + ret = of_sama5d4_wdt_init(pdev->dev.of_node, wdt); > + if (ret) > + return ret; > > if ((wdt->mr & AT91_WDT_WDFIEN) && irq) { > ret = devm_request_irq(&pdev->dev, irq, sama5d4_wdt_irq_handler, >
diff --git a/drivers/watchdog/sama5d4_wdt.c b/drivers/watchdog/sama5d4_wdt.c index 362fd229786d..d710014f3b7d 100644 --- a/drivers/watchdog/sama5d4_wdt.c +++ b/drivers/watchdog/sama5d4_wdt.c @@ -228,15 +228,13 @@ static int sama5d4_wdt_probe(struct platform_device *pdev) wdt->reg_base = regs; - if (pdev->dev.of_node) { - irq = irq_of_parse_and_map(pdev->dev.of_node, 0); - if (!irq) - dev_warn(&pdev->dev, "failed to get IRQ from DT\n"); + irq = irq_of_parse_and_map(pdev->dev.of_node, 0); + if (!irq) + dev_warn(&pdev->dev, "failed to get IRQ from DT\n"); - ret = of_sama5d4_wdt_init(pdev->dev.of_node, wdt); - if (ret) - return ret; - } + ret = of_sama5d4_wdt_init(pdev->dev.of_node, wdt); + if (ret) + return ret; if ((wdt->mr & AT91_WDT_WDFIEN) && irq) { ret = devm_request_irq(&pdev->dev, irq, sama5d4_wdt_irq_handler,
Because the only way to use the driver is to have a device tree enabling it, pdev->dev.of_node will never be NULL. Remove the unnecessary check. Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com> --- drivers/watchdog/sama5d4_wdt.c | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-)