netdev - Re: [PATCH] Allow MACB to connect to a higher addresses 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-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20100331055306.GC4837@datarespons.se>
Date:	Wed, 31 Mar 2010 07:53:06 +0200
From:	Anders Darander <anders.darander@...arespons.se>
To:	Grant Likely <grant.likely@...retlab.ca>
Cc:	"David S. Miller" <davem@...emloft.net>,
	Ralf Baechle <ralf@...ux-mips.org>,
	Maxime Bizon <mbizon@...ebox.fr>,
	David Daney <ddaney@...iumnetworks.com>,
	Sekhar Nori <nsekhar@...com>,
	Anton Vorontsov <avorontsov@...mvista.com>,
	Andy Fleming <afleming@...escale.com>, netdev@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] Allow MACB to connect to a higher addresses PHY.

Hi Grant,

Thanks for the feedback!

* Grant Likely <grant.likely@...retlab.ca> [100330 16:42]:
> On Tue, Mar 30, 2010 at 3:58 AM, Anders Darander
> <anders.darander@...il.com> wrote:
> > Using the Atmel MACB together with an integrated switch, can make only port 1
> > work. This is caused by macb_mii_probe trying to attach the MAC to the first
> > PHY, which often is on one of the external ports.
> >
> > E.g. the Micrel KSZ8873 connects to the MAC on port 3, thus phy_addr should be
> > set to 3.

> I understand what you are trying to do, but this is the wrong way to
> go about it.  Hard coding it into Kconfig breaks multiplatform
> kernels.  Besides, systems may have more than one physical MDIO bus.
> This patch would make CONFIG_SWITCHING_PHY_ADDR the only address
> accessible on all MDIO busses.

True, it would at least make all addresses < CONFIG_SWITCHING_PHY_ADDR
non-accessible.

We developed the patch to solve this problem on a 2.6.29-kernel, in
which the loop in question is directly in the macb_mii_probe() in the
macb.c, thus, the original patch is at least a _little_ bit less ugly...

> Nak.

I fully agree on the nak; I completely forgot that the functions in
phy_devices should be used from more interfaces than only the macb.

> The right thing to do is to add a runtime configuration option (ie.
> kernel parameter or platform data) to the mac driver to specify
> exactly which PHY address it is supposed to use.

I'll see if I can get some time to do such a thing. Unfortunately, it'll
probably have to be dealt with on my sparetime (the customer is quite
unlikely to let me work on that).

One question, if I add such an kernel parameter, would it be OK to
implement something like this in the macb-driver:

if parameter == SOME_DEFAULT
	phydev = phy_find_first(bp->mii_bus);
else
	phydev = bp->mii_bus->phy_map[parameter];

I.e. if the parameter is set to some default value, the old behaviour is
retained, otherwise the specified parameter is used for the phy_addr?

I'd appreciate your input on such a design.

Best regards,
Anders Darander

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html