netdev - Re: [PATCH] net: phy: at803x: disable delay only for RGMII mode

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]

Date:   Thu, 14 Feb 2019 15:22:28 +0200
From:   Peter Ujfalusi <peter.ujfalusi@...com>
To:     Niklas Cassel <niklas.cassel@...aro.org>
CC:     Marc Gonzalez <marc.w.gonzalez@...e.fr>,
        Andrew Lunn <andrew@...n.ch>,
        Florian Fainelli <f.fainelli@...il.com>,
        Vinod Koul <vkoul@...nel.org>,
        David S Miller <davem@...emloft.net>,
        <linux-arm-msm@...r.kernel.org>,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        <netdev@...r.kernel.org>, "Nori, Sekhar" <nsekhar@...com>
Subject: Re: [PATCH] net: phy: at803x: disable delay only for RGMII mode

Hi Niklas,

On 14/02/2019 14.39, Niklas Cassel wrote:
>>> So, I've rebased your old patch, see attachment.
>>> I suggest that Peter test it on am335x-evm.
>>
>> with the patch + s/rgmii-txid/rgmii-id in the am335x-evmsk.dts ethernet
>> is working.
>> I don't have am335x-evm to test, but it has the same PHY as evmsk.
>>
> 
> Florian's concern was that this PHY driver looked at "phy-mode" from the
> perspective of the MAC rather than the PHY.
> However, if s/rgmii-txid/rgmii-id is the correct fix for am335x-evm,
> then this means that this PHY driver was just broken.
> 
> If the driver had misinterpreted the perspective, then the correct
> fix for am335x-evm would have been s/rgmii-txid/rgmii-rxid.

Not sure if I got this right, but:
rgmii-id/txid/rxid is the delay mode between PHY and MAC, right?
on the PHY node it is from the PHY perspective, right?

The errata I have mentioned for am335x say:
"The reset state of RGMII1_IDMODE (bit 4) and RGMII2_IDMODE (bit 5) in
the GMII_SEL register enables internal delay mode on the transmit clock
of the respective RGMII port. The AM335x device does not support
internal delay mode, so RGMII1_IDMODE and RGMII2_IDMODE must be set to 1b."

If the delay mode on the transmit clock is not working on the am335x,
then this translate that the rxid needs to be enabled on the PHY side?

But then why it worked when only the txid was enabled and rxid was not
on the PHY side, and why it works if both txid and rxid is enabled?

Just tried w/ your patch and setting rgmii-rxid for am335x-evmsk and
ethernet is not working, it only works w/ rgmii-id (so both tx and rx
delay is enabled on the PHY side?)

> So considering that this driver seems to be really broken
> (rather then just inverted perspective),
> perhaps we can merge the patch I attached in my previous email after all?
> (Together with a s/rgmii-txid/rgmii-id in the am335x-evmsk.dts.)

at the same time am335x-evm.dts needs to have the same change and most
likely other boards which uses the same PHY needs to be checked?

PS: sorry for my lack of knowledge on the networking stuff...

- Péter

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki