Hello Miguel,

Le Thursday 2 September 2010 22:34:17, Miguel Gaio a écrit :
> Hi Florian,
> 
> On Thu, 2010-09-02 at 21:57 +0200, Florian Fainelli wrote:
> > Hello Miguel,
> > 
> > Le Thursday 2 September 2010 12:16:19, Miguel GAIO a écrit :
> > > Hi Florian,
> > > 
> > > On Thu, 2010-09-02 at 10:40 +0200, Florian Fainelli wrote:
> > > > Hi Miguel,
> > > > 
> > > > On Thursday 02 September 2010 10:36:06 Miguel GAIO wrote:
> > > > > Hi all,
> > > > > 
> > > > > I have started the rewrite for the Broadcom BCM5325 switch.
> > > > > It was test on neufbox4 board.
> > > > 
> > > > This is great work! I will give it a try on a couple of BCM6348/38
> > > > boards and report to you how it goes. Might also be worth testing on
> > > > BCM47xx and see if it works.
> > > 
> > > Thanks for you tests. A bug was introduce in previous
> > > 0001-Switch-add-support-for-Broadcom-BCM5325E-phy.patch.
> > > Please use attached update.
> > 
> > I did not get it probing the switch, but I assume this is due to my wrong
> > PHY mask. So far, I set it to 0xffffffff and tried with the same mask as
> > the NB4 board, either without success.
> > 
> > The old kmod-switch was able to detect the switch at PHY address 0, so I
> > am not sure what is wrong here. The switch is a BCM5325E.
> 
> ID 0 through 4 are switch ports PHYs.
> As I did not connect enet1 to any phys, the switch probe was perform via
> pseudo PHY (phy address 1Eh), assume Pseudo PHY dev_id is 0x00000000.

Ok, now I understand why you chose such a phy_mask.

> 
> If it's not case on your board, I have to rewrite probe part.
> 
> 
> Note: probe message is not print on console (log level 7). You should
> check via dmesg or via /sys/bus/mdio_bus/.
> Or use command
> # swicth dev bcm5325e show

On my board, it looks like the pseudo phy is not connected to the MDIO bus, 
because MDIO reg2 returns 0000 and MDIO reg3 returns 0xffff for PHY id 0-4 and 
0x1e. This is kind of expected because I thought the old kmod-switch worked as 
well, but it does not. Most likely the switch is SPI connected instead.

I will try on a BCM47xx board, which has more chances to be working.

> 
> > > > > It's not complete and does not support others Robo switches,
> > > > > comments are welcome.
> > > > > I can help for BCM5325 and BCM5395 support.
> > > > > 
> > > > > Kind regards,
> > > > > 
> > > > > Miguel
> > > > 
> > > > --
> > > > Florian
> > > 
> > > Miguel

Attachment: signature.asc
Description: This is a digitally signed message part.

_______________________________________________
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Reply via email to