linux-kernel - Re: [PATCH 04/13] irqchip: Add driver for Loongson-3 I/O interrupt controller

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: <45615c1a-7af8-3496-5369-4b2f174a76e7@kernel.org>
Date:   Wed, 28 Aug 2019 17:02:17 +0100
From:   Marc Zyngier <maz@...nel.org>
To:     Jiaxun Yang <jiaxun.yang@...goat.com>
Cc:     linux-mips@...r.kernel.org, chenhc@...ote.com,
        paul.burton@...s.com, tglx@...utronix.de, jason@...edaemon.net,
        linux-kernel@...r.kernel.org, robh+dt@...nel.org,
        mark.rutland@....co, devicetree@...r.kernel.org
Subject: Re: [PATCH 04/13] irqchip: Add driver for Loongson-3 I/O interrupt
 controller

On 28/08/2019 16:31, Jiaxun Yang wrote:
> 
> On 2019/8/28 下午2:59, Marc Zyngier wrote:
>> On Wed, 28 Aug 2019 08:27:05 +0800
>> Jiaxun Yang <jiaxun.yang@...goat.com> wrote:
>>
>>> On 2019/8/28 上午12:45, Marc Zyngier wrote:
>>>> On 27/08/2019 09:52, Jiaxun Yang wrote:
>>>>> +	chained_irq_enter(chip, desc);
>>>>> +
>>>>> +	pending = readl(priv->intc_base + LS3_REG_INTC_EN_STATUS) &
>>>>> +		readl(priv->intc_base + LS3_REG_INTC_STATUS);
>>>> Reading the enabled status from the HW on each interrupt? I'm sure
>>>> that's pretty cheap...
>>> Seems expensive but to deal with a buggy hardware... That's worthy.
>> How broken is it? You very much seem to rely on the HW being correct
>> here, since you trust it exclusively. I'd expect the enable mask to be
>> a SW construct if you didn't blindly trust it
> Hi Marc
> 
> Thanks for your answering.
> 
> The vendor code did this and said there is a HW issue. I just don't have 
> the guts to remove this check.
> Seems like sometimes masked interrupt may get ISR set wrongly.

And that would just as well avoided by a SW managed mask.

>> And if this is truly the right way to do it, please document the
>> various problems with the controller so that we don't break it at a
>> later time.
> Thanks, will do.
>>
>> Then how comes this comes from the irqchip's DT node? This should be
>> part of the endpoint's interrupt specifier.
> 
> In theory it should be, However if we set different interrupt 
> lines/cores on that controller, interrupts may get lost. It means we can 
> only have single parent core/interrupt.
> 
> So I'd prefer just set them uniformly by controller's dt-binding to 
> prevent confusing.

And I disagree. You can document the restriction, and even maybe enforce
it by validating the DT one way or another. But we're not putting what
ends up being a routing table in the irqchip binding.

	M.
-- 
Jazz is not dead, it just smells funny...