| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <OS0PR01MB5922C5E362A582509A814C1586CF9@OS0PR01MB5922.jpnprd01.prod.outlook.com>
Date: Mon, 16 May 2022 07:20:47 +0000
From: Biju Das <biju.das.jz@...renesas.com>
To: Marc Zyngier <maz@...nel.org>
CC: "Lad, Prabhakar" <prabhakar.csengg@...il.com>,
Geert Uytterhoeven <geert@...ux-m68k.org>,
Prabhakar Mahadev Lad <prabhakar.mahadev-lad.rj@...renesas.com>,
Linus Walleij <linus.walleij@...aro.org>,
Thomas Gleixner <tglx@...utronix.de>,
Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Bartosz Golaszewski <brgl@...ev.pl>,
Philipp Zabel <p.zabel@...gutronix.de>,
"open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Linux-Renesas <linux-renesas-soc@...r.kernel.org>,
"open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS"
<devicetree@...r.kernel.org>,
Phil Edworthy <phil.edworthy@...esas.com>
Subject: RE: [PATCH v3 5/5] pinctrl: renesas: pinctrl-rzg2l: Add IRQ domain to
handle GPIO interrupt
Hi Marc,
Thanks for the feedback.
> Subject: Re: [PATCH v3 5/5] pinctrl: renesas: pinctrl-rzg2l: Add IRQ domain
> to handle GPIO interrupt
>
> On Sun, 15 May 2022 06:13:22 +0100,
> Biju Das <biju.das.jz@...renesas.com> wrote:
> >
> > Hi Prabhakar,
> >
> > Thanks for the example.
> >
> > > Subject: Re: [PATCH v3 5/5] pinctrl: renesas: pinctrl-rzg2l: Add IRQ
> > > domain to handle GPIO interrupt
> > >
> > > > But "offset" is a number from the GPIO offset space (0-122), while
> > >
> > > The "offset" reported by kernel is 120-511:
> > >
> > > root@...rc-rzg2l:~# cat /sys/kernel/debug/gpio
> > > gpiochip0: GPIOs 120-511, parent: platform/11030000.pinctrl,
> > > 11030000.pinctrl:
> > > gpio-120 (P0_0 )
> > > gpio-121 (P0_1 )
> > > gpio-122 (P0_2 )
> > > gpio-123 (P0_3 )
> > > gpio-124 (P0_4 )
> > > .....
> > > gpio-507 (P48_3 )
> > > gpio-508 (P48_4 )
> > > gpio-509 (P48_5 )
> > > gpio-510 (P48_6 )
> > > gpio-511 (P48_7 )
> > >
> > > > irq_find_mapping() expects a number from the domain's IRQ space,
> > > > which is only 0-31?
> > > >
> > > Nope, let me demonstrate with an example, I have configured the gpio
> > > pins as GPIO keys in DTS:
> > >
> > > + keyboard {
> > > + compatible = "gpio-keys";
> > > + status = "okay";
> > > +
> > > + key-1 {
> > > + gpios = <&pinctrl RZG2L_GPIO(43, 0)
> > > GPIO_ACTIVE_HIGH>;
> > > + linux,code = <KEY_1>;
> > > + linux,input-type = <EV_KEY>;
> > > + wakeup-source;
> > > + label = "SW1";
> > > + };
> > > +
> > > + key-2 {
> > > + gpios = <&pinctrl RZG2L_GPIO(41, 0)
> > > GPIO_ACTIVE_HIGH>;
> > > + linux,code = <KEY_2>;
> > > + linux,input-type = <EV_KEY>;
> > > + wakeup-source;
> > > + label = "SW2";
> > > + };
> > > +
> > > + key-3 {
> > > + gpios = <&pinctrl RZG2L_GPIO(43, 1)
> > > GPIO_ACTIVE_HIGH>;
> > > + linux,code = <KEY_3>;
> > > + linux,input-type = <EV_KEY>;
> > > + wakeup-source;
> > > + label = "SW3";
> > > + };
> > > + };
> > >
> > > root@...rc-rzg2l:~# cat /proc/interrupts | grep SW
> > > root@...rc-rzg2l:~# root@...rc-rzg2l:~# insmod gpio_keys.ko [
> > > 925.002720] input: keyboard as
> > > /devices/platform/keyboard/input/input3
> > > root@...rc-rzg2l:~# cat /proc/interrupts | grep SW
> > > 82: 0 0 11030000.pinctrl 344 Edge SW1
> > > 83: 0 0 11030000.pinctrl 328 Edge SW2
> > > 84: 0 0 11030000.pinctrl 345 Edge SW3
> > > root@...rc-rzg2l:~#
> > >
> > > In here 82/83/84 are virq and 344/328/345 are hwirq, which can be
> > > confirmed from sysfs file:
> >
> > From your example, Looks like
> >
> > I believe from interrupt statistics point of view, cat
> > /proc/interrupts should report actual gpioint number (0->122)
> > corresponding to pin index for SW1, SW2 and SW3 ??
>
> No. There is no need for such userspace-visible behaviour. Userspace has no
> business tracking those. The required information is in debugfs, and that
> more than enough.
Ok, So far I used cat /proc/interrupts for debugging, since I don't need to enable DEBUG config for
Enabling Debugfs for irq. This Debugfs irq is new info to me.
Our hardware manual has below info for usb-phy irq
2H0_OBINT 126(InterruptID) SPI 94 IRQ 94 Level
cat /proc/interrupts matches with GICV3 Interrupt ID/ type in the HW manual
113: 0 0 GICv3 126 Level 11c50200.usb-phy
Debugfs is also showing similar info like hwirq and interrupt type. But I don't know which field corresponds to number
of interrupts?
root@...rc-rzg2l:~# cat /sys/kernel/debug/irq/irqs/113
handler: handle_fasteoi_irq
device: (null)
status: 0x00000104
istate: 0x00000000
ddepth: 0
wdepth: 0
dstate: 0x13402204
IRQ_TYPE_LEVEL_HIGH
IRQD_LEVEL
IRQD_ACTIVATED
IRQD_IRQ_STARTED
IRQD_SINGLE_TARGET
IRQD_DEFAULT_TRIGGER_SET
IRQD_HANDLE_ENFORCE_IRQCTX
node: 0
affinity: 0-1
effectiv: 0
domain: :soc:interrupt-controller@...00000-1
hwirq: 0x7e
chip: GICv3
flags: 0x15
IRQCHIP_SET_TYPE_MASKED
IRQCHIP_MASK_ON_SUSPEND
IRQCHIP_SKIP_SET_WAKE
Now coming to current case,
Currently GPIO INT 0-122(123 interrupts) corresponding to 120-511(291 interrupts) with same invalid lines.
>From a debugging point, If user has put same irq name for gpioints(cat /proc/interrupts case), then how do we distinguish these interrupts??
(using hwirq??)
For using Debugfs, Do we need to first execute cat /proc/interrupts to get virq and from there we need to use virq to get statistics, right?
Cheers,
Biju
Powered by blists - more mailing lists