| Message ID | 20211205111349.51213-1-lars@metafoo.de (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
| Series | remoteproc: ingenic: Request IRQ disabled | expand |
Hi Lars, Le dim., déc. 5 2021 at 12:13:49 +0100, Lars-Peter Clausen <lars@metafoo.de> a écrit : > The ingenic remoteproc driver requests its IRQ and then immediately > disables it. > > The disable is necessary since irq_request() normally enables the > IRQ. But > there is a new flag IRQF_NO_AUTOEN that when specified keeps the IRQ > disabled. Use this new flag rather than calling disable_irq(). > > This slightly reduce the boilerplate code and also avoids a > theoretical > race condition where the IRQ could fire between irq_request() and > disable_irq(). > > Signed-off-by: Lars-Peter Clausen <lars@metafoo.de> Acked-by: Paul Cercueil <paul@crapouillou.net> Cheers, -Paul > --- > drivers/remoteproc/ingenic_rproc.c | 5 ++--- > 1 file changed, 2 insertions(+), 3 deletions(-) > > diff --git a/drivers/remoteproc/ingenic_rproc.c > b/drivers/remoteproc/ingenic_rproc.c > index a356738160a4..9902cce28692 100644 > --- a/drivers/remoteproc/ingenic_rproc.c > +++ b/drivers/remoteproc/ingenic_rproc.c > @@ -218,14 +218,13 @@ static int ingenic_rproc_probe(struct > platform_device *pdev) > if (vpu->irq < 0) > return vpu->irq; > > - ret = devm_request_irq(dev, vpu->irq, vpu_interrupt, 0, "VPU", > rproc); > + ret = devm_request_irq(dev, vpu->irq, vpu_interrupt, IRQF_NO_AUTOEN, > + "VPU", rproc); > if (ret < 0) { > dev_err(dev, "Failed to request IRQ\n"); > return ret; > } > > - disable_irq(vpu->irq); > - > ret = devm_rproc_add(dev, rproc); > if (ret) { > dev_err(dev, "Failed to register remote processor\n"); > -- > 2.30.2 >
On Sun, Dec 05, 2021 at 12:13:49PM +0100, Lars-Peter Clausen wrote: > The ingenic remoteproc driver requests its IRQ and then immediately > disables it. > > The disable is necessary since irq_request() normally enables the IRQ. But > there is a new flag IRQF_NO_AUTOEN that when specified keeps the IRQ > disabled. Use this new flag rather than calling disable_irq(). > > This slightly reduce the boilerplate code and also avoids a theoretical > race condition where the IRQ could fire between irq_request() and > disable_irq(). > > Signed-off-by: Lars-Peter Clausen <lars@metafoo.de> > --- > drivers/remoteproc/ingenic_rproc.c | 5 ++--- > 1 file changed, 2 insertions(+), 3 deletions(-) > > diff --git a/drivers/remoteproc/ingenic_rproc.c b/drivers/remoteproc/ingenic_rproc.c > index a356738160a4..9902cce28692 100644 > --- a/drivers/remoteproc/ingenic_rproc.c > +++ b/drivers/remoteproc/ingenic_rproc.c > @@ -218,14 +218,13 @@ static int ingenic_rproc_probe(struct platform_device *pdev) > if (vpu->irq < 0) > return vpu->irq; > > - ret = devm_request_irq(dev, vpu->irq, vpu_interrupt, 0, "VPU", rproc); > + ret = devm_request_irq(dev, vpu->irq, vpu_interrupt, IRQF_NO_AUTOEN, > + "VPU", rproc); > if (ret < 0) { > dev_err(dev, "Failed to request IRQ\n"); > return ret; > } > > - disable_irq(vpu->irq); > - I have applied this patch. Thanks, Mathieu > ret = devm_rproc_add(dev, rproc); > if (ret) { > dev_err(dev, "Failed to register remote processor\n"); > -- > 2.30.2 >
diff --git a/drivers/remoteproc/ingenic_rproc.c b/drivers/remoteproc/ingenic_rproc.c index a356738160a4..9902cce28692 100644 --- a/drivers/remoteproc/ingenic_rproc.c +++ b/drivers/remoteproc/ingenic_rproc.c @@ -218,14 +218,13 @@ static int ingenic_rproc_probe(struct platform_device *pdev) if (vpu->irq < 0) return vpu->irq; - ret = devm_request_irq(dev, vpu->irq, vpu_interrupt, 0, "VPU", rproc); + ret = devm_request_irq(dev, vpu->irq, vpu_interrupt, IRQF_NO_AUTOEN, + "VPU", rproc); if (ret < 0) { dev_err(dev, "Failed to request IRQ\n"); return ret; } - disable_irq(vpu->irq); - ret = devm_rproc_add(dev, rproc); if (ret) { dev_err(dev, "Failed to register remote processor\n");
The ingenic remoteproc driver requests its IRQ and then immediately disables it. The disable is necessary since irq_request() normally enables the IRQ. But there is a new flag IRQF_NO_AUTOEN that when specified keeps the IRQ disabled. Use this new flag rather than calling disable_irq(). This slightly reduce the boilerplate code and also avoids a theoretical race condition where the IRQ could fire between irq_request() and disable_irq(). Signed-off-by: Lars-Peter Clausen <lars@metafoo.de> --- drivers/remoteproc/ingenic_rproc.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-)