netdev - Inphi IN112525 PHY

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-next>] [day] [month] [year] [list]

Message-ID: <HE1PR0402MB35781CC11AAC5DB372D194CFEED00@HE1PR0402MB3578.eurprd04.prod.outlook.com>
Date:   Thu, 15 Mar 2018 13:19:27 +0000
From:   Vicenţiu Galanopulo 
        <vicentiu.galanopulo@....com>
To:     "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Inphi IN112525 PHY

Hi Everyone,

We have a new board which is to be released this year, which has a Inphi IN112525 PHY. We are currently developing under kernel 4.9 and using the xgmac_mdio driver for the MDIO bus. The long term plan is to have a Inphi kernel PHY driver upstream, but right now we want to make it work for 4.9.

Our problem is that:
The Inphi PHY has a vendor specific address space for accessing the C45 MDIO registers -  MMD30 address space - starting from 0x1E.
I noticed that during the probing of the MDIO bus, in phy_device.c, get_phy_c45_ids(), it first searches for non-zero Devices In package, with the dev_addr = 1.
After that, device zero (dev_addr = 0)   is probed ("Device zero is reserved for 802.3 c45 complied PHYs[.]") as a special case for Cortina CS4315/CS4340 PHY.
Since the dev_addr for our Inphi PHY is 0x1E, it never gets found and the PHY driver doesn't get registered.
My colleagues have proposed a solution in which another property could be added in the PHY device tree node, that would contain the device address (dev_addr). 
This would imply some minor changes in phy_device.c , in get_phy_c45_ids(), but the benefit of this would be that it should work for other PHY vendors which decide to have a non-standard MDIO register set address.
I have a patch ready for this, but we would like to know if this is right way to go about this.

Please share your thoughts on this.

All the best,
Vicentiu