* Tony Lindgren <t...@atomide.com> [181113 18:07]: > Hi > > * Belisko Marek <marek.beli...@gmail.com> [180703 18:34]: > > Hi Tony, > > > > On Tue, Jul 3, 2018 at 10:45 AM Tony Lindgren <t...@atomide.com> wrote: > > > > > > * Belisko Marek <marek.beli...@gmail.com> [180620 09:40]: > > > > Hello, > > > > > > > > I'm trying to fix warning (for omap5 board) produced by recent change > > > > to avoid using IRQ_TYPE_NONE like: > > > > [ 1.818666] WARNING: CPU: 1 PID: 778 at > > > > drivers/irqchip/irq-gic.c:1016 gic_irq_domain_translate+0x78/0x100 > > > > [ 1.828839] Modules linked in: > > > > > > > > I did look to other commit which did update and without deep knowledge > > > > I just simply do this small change: > > > > diff --git a/arch/arm/boot/dts/omap5-board-common.dtsi > > > > b/arch/arm/boot/dts/omap5-board-common.dtsi > > > > index 218892b..ab2df8c 100644 > > > > --- a/arch/arm/boot/dts/omap5-board-common.dtsi > > > > +++ b/arch/arm/boot/dts/omap5-board-common.dtsi > > > > @@ -393,7 +393,7 @@ > > > > > > > > palmas: palmas@48 { > > > > compatible = "ti,palmas"; > > > > - interrupts = <GIC_SPI 7 IRQ_TYPE_NONE>; /* IRQ_SYS_1N */ > > > > + interrupts = <GIC_SPI 7 IRQ_TYPE_HIGH>; /* IRQ_SYS_1N */ > > > > reg = <0x48>; > > > > interrupt-controller; > > > > #interrupt-cells = <2>; > > > > > > > > and it looks board boots fine. Only issue is that gpadc driver is not > > > > working (at least not getting interrupts at all ADC fails with > > > > timeout). I did look to gpadc driver and driver is not using > > > > interrupts defined in dts but request interrupt directly from palmas > > > > mfd module. Any ideas what needs to be changed to have gpadc again > > > > working with mentioned patch? > > > > > > Can you try with IRQF_TRIGGER_HIGH added also to the flags to > > > regmap_add_irq_chip() in drivers/mfd/palmas.c? > > Nope issue is till present also after this change like: > > diff --git a/arch/arm/boot/dts/omap5-board-common.dtsi > > b/arch/arm/boot/dts/omap5-board-common.dtsi > > index 218892b..6912769 100644 > > --- a/arch/arm/boot/dts/omap5-board-common.dtsi > > +++ b/arch/arm/boot/dts/omap5-board-common.dtsi > > @@ -393,7 +393,7 @@ > > > > palmas: palmas@48 { > > compatible = "ti,palmas"; > > - interrupts = <GIC_SPI 7 IRQ_TYPE_NONE>; /* IRQ_SYS_1N */ > > + interrupts = <GIC_SPI 7 IRQ_TYPE_LEVEL_HIGH>; /* IRQ_SYS_1N > > */ > > reg = <0x48>; > > interrupt-controller; > > #interrupt-cells = <2>; > > @@ -432,9 +432,9 @@ > > > > gpadc: gpadc { > > compatible = "ti,palmas-gpadc"; > > - interrupts = <18 0 > > - 16 0 > > - 17 0>; > > + interrupts = <18 IRQ_TYPE_LEVEL_HIGH > > + 16 IRQ_TYPE_LEVEL_HIGH > > + 17 IRQ_TYPE_LEVEL_HIGH>; > > #io-channel-cells = <1>; > > ti,channel0-current-microamp = <5>; > > ti,channel3-current-microamp = <10>; > > diff --git a/drivers/mfd/palmas.c b/drivers/mfd/palmas.c > > index 663a239..15d23db 100644 > > --- a/drivers/mfd/palmas.c > > +++ b/drivers/mfd/palmas.c > > @@ -601,7 +601,7 @@ static int palmas_i2c_probe(struct i2c_client *i2c, > > regmap_write(palmas->regmap[slave], addr, reg); > > > > ret = regmap_add_irq_chip(palmas->regmap[slave], palmas->irq, > > - IRQF_ONESHOT | pdata->irq_flags, 0, > > + IRQF_ONESHOT | IRQF_TRIGGER_HIGH | > > pdata->irq_flags, 0, > > driver_data->irq_chip, &palmas->irq_data); > > if (ret < 0) > > goto err_i2c; > > Looks like the IRQ_TYPE_NONE issue still is there for omap5 and > should be fixed with IRQ_TYPE_HIGH.
Looks like the gpadc interrupts get fixed for IRQ_TYPE_LEVEL_HIGH if reconfiguring of PALMAS_POLARITY_CTRL_INT_POLARITY is disabled in drivers/mfd/palmas.c. The test being just: modprobe palmas-gpadc cat /sys/bus/iio/devices/iio:device0/* > No idea about why palmas interrupts would stop working though, > Peter, do you have any ideas on this one? Still no idea why though, it seems tegra is inverting the interrupt externally because of earlier patches for adding "ti,irq-externally-inverted" property that never got added. So I'm guessing the PALMAS_POLARITY_CTRL_INT_POLARITY is wrongly configured on IRQ_TYPE_LEVEL_HIGH while it should be done only for IRQ_TYPE_LEVEL_LOW instead? So adding Laxman to Cc also. Regards, Tony