| Message ID | 1420551550-27512-1-git-send-email-geert+renesas@glider.be (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi Geert, Thank you for the patch. On Tuesday 06 January 2015 14:39:10 Geert Uytterhoeven wrote: > The sh73a0 INTC can't mask interrupts properly most likely due to a > hardware bug. Set the .control_parent flag to delegate masking to the > parent interrupt controller, like was already done for irqpin1. > > Without this, accessing the three-axis digital accelerometer ADXL345 > on kzm9g through /dev/input/event1 causes an interrupt storm, which > requires a power-cycle to recover from. > > This was inspired by a patch for arch/arm/boot/dts/sh73a0.dtsi from > Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>. This looks correct to me, but shouldn't we concentrate on getting rid of legacy code instead ? > Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> > Fixes: 341eb5465f67437a ("ARM: shmobile: INTC External IRQ pin driver on > sh73a0") --- > arch/arm/mach-shmobile/setup-sh73a0.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/arch/arm/mach-shmobile/setup-sh73a0.c > b/arch/arm/mach-shmobile/setup-sh73a0.c index > 354cab111bf17ea6..613ce65184a6648c 100644 > --- a/arch/arm/mach-shmobile/setup-sh73a0.c > +++ b/arch/arm/mach-shmobile/setup-sh73a0.c > @@ -595,6 +595,7 @@ static struct platform_device ipmmu_device = { > > static struct renesas_intc_irqpin_config irqpin0_platform_data = { > .irq_base = irq_pin(0), /* IRQ0 -> IRQ7 */ > + .control_parent = true, > }; > > static struct resource irqpin0_resources[] = { > @@ -656,6 +657,7 @@ static struct platform_device irqpin1_device = { > > static struct renesas_intc_irqpin_config irqpin2_platform_data = { > .irq_base = irq_pin(16), /* IRQ16 -> IRQ23 */ > + .control_parent = true, > }; > > static struct resource irqpin2_resources[] = { > @@ -686,6 +688,7 @@ static struct platform_device irqpin2_device = { > > static struct renesas_intc_irqpin_config irqpin3_platform_data = { > .irq_base = irq_pin(24), /* IRQ24 -> IRQ31 */ > + .control_parent = true, > }; > > static struct resource irqpin3_resources[] = {
On Tue, Jan 06, 2015 at 04:14:09PM +0200, Laurent Pinchart wrote: > Hi Geert, > > Thank you for the patch. > > On Tuesday 06 January 2015 14:39:10 Geert Uytterhoeven wrote: > > The sh73a0 INTC can't mask interrupts properly most likely due to a > > hardware bug. Set the .control_parent flag to delegate masking to the > > parent interrupt controller, like was already done for irqpin1. > > > > Without this, accessing the three-axis digital accelerometer ADXL345 > > on kzm9g through /dev/input/event1 causes an interrupt storm, which > > requires a power-cycle to recover from. > > > > This was inspired by a patch for arch/arm/boot/dts/sh73a0.dtsi from > > Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>. > > This looks correct to me, but shouldn't we concentrate on getting rid of > legacy code instead ? Probably, but as we now have this patch I'm inclined to apply it. > > Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> > > Fixes: 341eb5465f67437a ("ARM: shmobile: INTC External IRQ pin driver on > > sh73a0") --- > > arch/arm/mach-shmobile/setup-sh73a0.c | 3 +++ > > 1 file changed, 3 insertions(+) > > > > diff --git a/arch/arm/mach-shmobile/setup-sh73a0.c > > b/arch/arm/mach-shmobile/setup-sh73a0.c index > > 354cab111bf17ea6..613ce65184a6648c 100644 > > --- a/arch/arm/mach-shmobile/setup-sh73a0.c > > +++ b/arch/arm/mach-shmobile/setup-sh73a0.c > > @@ -595,6 +595,7 @@ static struct platform_device ipmmu_device = { > > > > static struct renesas_intc_irqpin_config irqpin0_platform_data = { > > .irq_base = irq_pin(0), /* IRQ0 -> IRQ7 */ > > + .control_parent = true, > > }; > > > > static struct resource irqpin0_resources[] = { > > @@ -656,6 +657,7 @@ static struct platform_device irqpin1_device = { > > > > static struct renesas_intc_irqpin_config irqpin2_platform_data = { > > .irq_base = irq_pin(16), /* IRQ16 -> IRQ23 */ > > + .control_parent = true, > > }; > > > > static struct resource irqpin2_resources[] = { > > @@ -686,6 +688,7 @@ static struct platform_device irqpin2_device = { > > > > static struct renesas_intc_irqpin_config irqpin3_platform_data = { > > .irq_base = irq_pin(24), /* IRQ24 -> IRQ31 */ > > + .control_parent = true, > > }; > > > > static struct resource irqpin3_resources[] = { > > -- > Regards, > > Laurent Pinchart >
Hi Laurent, On Tue, Jan 6, 2015 at 3:14 PM, Laurent Pinchart <laurent.pinchart@ideasonboard.com> wrote: > On Tuesday 06 January 2015 14:39:10 Geert Uytterhoeven wrote: >> The sh73a0 INTC can't mask interrupts properly most likely due to a >> hardware bug. Set the .control_parent flag to delegate masking to the >> parent interrupt controller, like was already done for irqpin1. >> >> Without this, accessing the three-axis digital accelerometer ADXL345 >> on kzm9g through /dev/input/event1 causes an interrupt storm, which >> requires a power-cycle to recover from. >> >> This was inspired by a patch for arch/arm/boot/dts/sh73a0.dtsi from >> Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>. > > This looks correct to me, but shouldn't we concentrate on getting rid of > legacy code instead ? Sure. But this is a serious bug, which allows to crash the kernel from user space. Worse, a simple reset is not sufficient to recover. You need to power-cycle the board. Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds
On Wed, Jan 07, 2015 at 09:14:19AM +0100, Geert Uytterhoeven wrote: > Hi Laurent, > > On Tue, Jan 6, 2015 at 3:14 PM, Laurent Pinchart > <laurent.pinchart@ideasonboard.com> wrote: > > On Tuesday 06 January 2015 14:39:10 Geert Uytterhoeven wrote: > >> The sh73a0 INTC can't mask interrupts properly most likely due to a > >> hardware bug. Set the .control_parent flag to delegate masking to the > >> parent interrupt controller, like was already done for irqpin1. > >> > >> Without this, accessing the three-axis digital accelerometer ADXL345 > >> on kzm9g through /dev/input/event1 causes an interrupt storm, which > >> requires a power-cycle to recover from. > >> > >> This was inspired by a patch for arch/arm/boot/dts/sh73a0.dtsi from > >> Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>. > > > > This looks correct to me, but shouldn't we concentrate on getting rid of > > legacy code instead ? > > Sure. > > But this is a serious bug, which allows to crash the kernel from user space. > Worse, a simple reset is not sufficient to recover. You need to power-cycle > the board. With that in mind I have queued this up as a fix for v3.19. For (my) reference, 341eb5465f67437a ("ARM: shmobile: INTC External IRQ pin driver on sh73a0"), which this patch fixes, was part of v3.9.
diff --git a/arch/arm/mach-shmobile/setup-sh73a0.c b/arch/arm/mach-shmobile/setup-sh73a0.c index 354cab111bf17ea6..613ce65184a6648c 100644 --- a/arch/arm/mach-shmobile/setup-sh73a0.c +++ b/arch/arm/mach-shmobile/setup-sh73a0.c @@ -595,6 +595,7 @@ static struct platform_device ipmmu_device = { static struct renesas_intc_irqpin_config irqpin0_platform_data = { .irq_base = irq_pin(0), /* IRQ0 -> IRQ7 */ + .control_parent = true, }; static struct resource irqpin0_resources[] = { @@ -656,6 +657,7 @@ static struct platform_device irqpin1_device = { static struct renesas_intc_irqpin_config irqpin2_platform_data = { .irq_base = irq_pin(16), /* IRQ16 -> IRQ23 */ + .control_parent = true, }; static struct resource irqpin2_resources[] = { @@ -686,6 +688,7 @@ static struct platform_device irqpin2_device = { static struct renesas_intc_irqpin_config irqpin3_platform_data = { .irq_base = irq_pin(24), /* IRQ24 -> IRQ31 */ + .control_parent = true, }; static struct resource irqpin3_resources[] = {
The sh73a0 INTC can't mask interrupts properly most likely due to a hardware bug. Set the .control_parent flag to delegate masking to the parent interrupt controller, like was already done for irqpin1. Without this, accessing the three-axis digital accelerometer ADXL345 on kzm9g through /dev/input/event1 causes an interrupt storm, which requires a power-cycle to recover from. This was inspired by a patch for arch/arm/boot/dts/sh73a0.dtsi from Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>. Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Fixes: 341eb5465f67437a ("ARM: shmobile: INTC External IRQ pin driver on sh73a0") --- arch/arm/mach-shmobile/setup-sh73a0.c | 3 +++ 1 file changed, 3 insertions(+)