linux-kernel - Re: [PATCH] irqchip/irq-mst: Support polarity configuration

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAFr9PXkK+L0kpEfJsQMVa0R9P3jMDJFHQKCR8jaCuL=Um6G=gQ@mail.gmail.com>
Date:   Mon, 8 Mar 2021 22:06:05 +0900
From:   Daniel Palmer <daniel@...f.com>
To:     Mark-PK Tsai (蔡沛剛) 
        <Mark-PK.Tsai@...iatek.com>
Cc:     Marc Zyngier <maz@...nel.org>, Daniel Palmer <daniel@...ngy.jp>,
        Thomas Gleixner <tglx@...utronix.de>,
        Jason Cooper <jason@...edaemon.net>,
        Matthias Brugger <matthias.bgg@...il.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
        "linux-mediatek@...ts.infradead.org" 
        <linux-mediatek@...ts.infradead.org>,
        YJ Chiang (江英杰) <yj.chiang@...iatek.com>
Subject: Re: [PATCH] irqchip/irq-mst: Support polarity configuration

Hi Mark-PK,

On Mon, 8 Mar 2021 at 11:42, Mark-PK Tsai (蔡沛剛)
<Mark-PK.Tsai@...iatek.com> wrote:
> > It seems like the polarity of interrupts is always the same between the MStar intc and the GIC? Low level interrupts are handled in the mstar intc and become high level interrupts to the GIC?
> > I think for the Mstar MSC313(e) and SigmaStar chips all of the internal interrupts are high level so I never noticed this behaviour.
> > I can't remember seeing anything that handled this in the MStar kernel code I looked at.
> > Is this specific to a certain chip or does it apply for everything with this intc?
>
> I suppose Mstar SoCs also need this patch which depends on what kind of interrupt source the HW designer wire to this intc.
> If an interrupt source is active low, we need to set the corresponding bit to reverse the polarity to meet GIC SPI requirement as Marc mentioned.

That makes sense. Looking back at the original MStar driver again they
did handle reversing the polarity but they passed the original
polarity through to the GIC too.
So maybe it never worked properly. :)

I think I can test this with GPIOs that are wired into the mstar intc
on the MStar/SigmaStar chips. I'll get back to you with how that works
out.

Thanks,

Daniel



>
> > The register values being lost if the chip goes into suspend to memory makes sense for the MStar chips too I think as everything that is not in the "pmsleep" register group seems to be lost.
>
> There are mask and eoi bits I did not handle here.
> That's because kernel will handle the mask and eoi status when system going to suspend/resume in suspend_device_irqs/ resume_device_irqs.
> And all the irqs of Mstar intc are masked by default when the IP powered on.
>
>
> Best regards,
> Mark-PK Tsai
>
>
> -----Original Message-----
> From: Marc Zyngier [mailto:maz@...nel.org]
> Sent: Sunday, March 7, 2021 2:28 AM
> To: Daniel Palmer <daniel@...f.com>
> Cc: Mark-PK Tsai (蔡沛剛) <Mark-PK.Tsai@...iatek.com>; Daniel Palmer <daniel@...ngy.jp>; Thomas Gleixner <tglx@...utronix.de>; Jason Cooper <jason@...edaemon.net>; Matthias Brugger <matthias.bgg@...il.com>; Linux Kernel Mailing List <linux-kernel@...r.kernel.org>; linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>; linux-mediatek@...ts.infradead.org; YJ Chiang (江英杰) <yj.chiang@...iatek.com>
> Subject: Re: [PATCH] irqchip/irq-mst: Support polarity configuration
>
> On Sat, 06 Mar 2021 17:06:51 +0000,
> Daniel Palmer <daniel@...f.com> wrote:
> >
> > Hi Mark-PK,
> >
> > I'm trying to understand the logic behind the changes.
> > It seems like the polarity of interrupts is always the same between
> > the MStar intc and the GIC? Low level interrupts are handled in the
> > mstar intc and become high level interrupts to the GIC?
>
> That's because the GIC only supports level-high input interrupts when they are level triggered (and rising edge when edge triggered).
>
> Thanks,
>
>         M.
>
> --
> Without deviation from the norm, progress is not possible.