On 21/11/13 10:56, Daniel Schwierzeck wrote:
2013/11/20 Matti Laakso<malaa...@elisanet.fi>:
On 19.11.2013 21:10, John Crispin wrote:
On 14/11/13 08:59, Matti Laakso wrote:
At the moment the MII/RMII setting in the lantiq .dts-files has no
effect, since wrong registers are written. Once this patch is applied,
it is necessary to make sure that the MII/RMII setting in the .dts-file
is correct.
i have not merged this yet. i am wondering if we should rename th
property in all relevant dts files and print a warning during boot if
the real option is not set.... that way there is a migration path that
wont break lots of boards for people
That might be a good idea. For the Danube routers I was able to compile a
list of switch chips and connection types:
* AR8216, MII
arv4519pw
arv7518pw
arv752dpw22
* RTL8306, reversed MII
arv752dpw
arv452cqw
arv4518pw
* ADM6996, reversed MII
arv4510pw/arv7510pw
sx76x
easy50712
acmp252
arv4520pw
which leaves out just arv4525pw and arv7525pw. Based on this list only
arv4519pw and arv7518pw have currently an incorrect setting in the
.dts-file.
But then there are the AR9-based routers which use the same driver, and
which I'm not familiar with. Based on the new U-Boot sources at least
p2601hnfx and h201l use reversed MII, but in the current driver all
AR9-routers are forced to use MII regardless of .dts-file?
ignore the U-Boot sources for AR9 for now because the AR9 ethernet
support is still in development and is not working yet. But the MII
settings should be the same as on Danube boards. That depends only on
the assembled switch chip.
Hi,
the link between the etop and gbit switch is mii and then the gbit
switch has 2 ports that can be R/Mii depending on how they are wired.
... correct me if i am wrong ...
John
_______________________________________________
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
