netdev - Re: [PATCH V2] net: phy: tja11xx: Add IRQ support to the driver

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: <19f9e596-5b51-8c76-396e-572d3e8da463@denx.de>
Date:   Thu, 30 May 2019 01:46:49 +0200
From:   Marek Vasut <marex@...x.de>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     Heiner Kallweit <hkallweit1@...il.com>, netdev@...r.kernel.org,
        Florian Fainelli <f.fainelli@...il.com>,
        Guenter Roeck <linux@...ck-us.net>,
        Jean Delvare <jdelvare@...e.com>, linux-hwmon@...r.kernel.org
Subject: Re: [PATCH V2] net: phy: tja11xx: Add IRQ support to the driver

On 5/30/19 1:29 AM, Andrew Lunn wrote:
> On Tue, May 28, 2019 at 11:33:33PM +0200, Marek Vasut wrote:
>> On 5/28/19 11:22 PM, Andrew Lunn wrote:
>>>> The link detection on the TJA1100 (not TJA1101) seems unstable at best,
>>>> so I better use all the interrupt sources to nudge the PHY subsystem and
>>>> have it check the link change.
>>>
>>> Then it sounds like you should just ignore interrupts and stay will
>>> polling for the TJA1100.
>>
>> Polling for the link status change is slow(er) than the IRQ driven
>> operation, so I would much rather use the interrupts.
> 
> I agree about the speed, but it seems like interrupts on this PHY are
> not so reliable. Polling always works. But unfortunately, you cannot
> have both interrupts and polling to fix up problems when interrupts
> fail. Your call, do you think interrupts really do work?

It works fine for me this way. And mind you, it's only the TJA1100
that's flaky, the TJA1101 is better.

> If you say that tja1101 works as expected, then please just use the
> link up/down bits for it.

I still don't know which bits really trigger link status changes, so I'd
like to play it safe and just trigger on all of them.

-- 
Best regards,
Marek Vasut