Re: [PATCH v6 2/2] EDAC: al-mc-edac: Introduce Amazon's Annapurna Labs Memory Controller EDAC

From: "Shenhar, Talel" <talel@amazon.com>
To: Borislav Petkov <bp@alien8.de>
Cc: <mchehab@kernel.org>, <james.morse@arm.com>,
	<davem@davemloft.net>, <gregkh@linuxfoundation.org>,
	<nicolas.ferre@microchip.com>, <robh+dt@kernel.org>,
	<mark.rutland@arm.com>, <catalin.marinas@arm.com>,
	<will@kernel.org>, <linux-edac@vger.kernel.org>,
	<devicetree@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
	<linux-arm-kernel@lists.infradead.org>, <dwmw@amazon.co.uk>,
	<benh@kernel.crashing.org>, <hhhawa@amazon.com>,
	<ronenk@amazon.com>, <jonnyc@amazon.com>, <hanochu@amazon.com>,
	<eitan@amazon.com>
Subject: Re: [PATCH v6 2/2] EDAC: al-mc-edac: Introduce Amazon's Annapurna Labs Memory Controller EDAC
Date: Tue, 5 May 2020 13:55:14 +0300	[thread overview]
Message-ID: <4d389169-89d1-4512-f6f0-c98477ab3623@amazon.com> (raw)
In-Reply-To: <20200504183716.GJ15046@zn.tnic>


On 5/4/2020 9:37 PM, Borislav Petkov wrote:
>
> On Mon, May 04, 2020 at 01:16:10PM +0300, Shenhar, Talel wrote:
>>>> +     mci = edac_mc_alloc(0, ARRAY_SIZE(layers), layers,
>>>> +                         sizeof(struct al_mc_edac));
>>> You can let that line stick out.
>> I rather avoid having this as a checkpatch warnning... (automation and
>> stuff...)
> checkpatch.pl - while useful - should not be taken to the letter and
> human brain should be applied to sanity check it what it warns about.
>
>> This line break does seems to my eye as too hard to read.
>>
>> Let me know if you feel strongly about it.
> I'm just sayin' - in the end of the day you'll be staring at that code -
> not me - so whatever *you* prefer. :-)
>
> Just don't follow tools blindly.
Thanks, I will leave it that way as it will make my life easier (with 
automatic vim tools and automation) and doesn't really break code 
understanding.
>
>>>> +     if (al_mc->irq_ue <= 0 || al_mc->irq_ce <= 0)
>>> Shouldn't this be && here?
>>>
>>> I mean, you want to poll when neither of the IRQs can be found. But then
>>> if you find one of them and not the other, what do you do? Poll and
>>> interrupt? Is that case even possible?
>> Correct.
>>
>> In case dt defined interrupt line only for one type and not for the other,
>> than the interrupt mode shall be used for one of them while polling mode for
>> the other.
> That warrants a comment above it.
Shall be part of v7.
>
> Thx.
>
> --
> Regards/Gruss,
>      Boris.
>
> https://people.kernel.org/tglx/notes-about-netiquette