Hi,

> My plan was to put the bcm4716 related stuff into the patch directory
> for kernel 2.6.37 and stay at 2.6.36 as default kernel for trunk until
> the new stuff is somehow stable. I think we should get this really work
> soon as it hardware is out for 1.5 years now.
> 
> Hauke

It would be great but must to get a rid of those 4 SSB_BUG_ON(1) in pmu
r5/r6 code I left in patch for them. Otherwise its too much dangerous to
let it get built for anything else besides 4716. Might get back to that
in few days as soon as done with fixed code for AI bus core.
Unfortunately I put too much messup to ssb core abstraction layer with
changes made. I was plain wrong in few assumptions being blinded with
mips-only part of the bus code base. Thanks to Michael Buesch I got
another understanding of things how they are done currently and
therefore how could they be done to avoid such a messup and provide
clean way of these two buses coexistance on signle code base.
I'm still sure SB and AI are both much more SSB rather than SSB SB and
lets say BCM AI code wise. It could have been changed if only more
technical implementations' background is awailable for AI but
unfortunately no .pdf's from Broadcom. 

Now back to subject. Think we could leave it as it is for few days -
might get some reports from bcm4716 owners. I will publish fixed
decision in few evenings after done with code changes and tests for SB
code on non-mipses. Its really long time to get it finally supported
after 1,5 years since release but think it can wait 1ne more week
still ;)

Have nice day,
George


_______________________________________________
openwrt-devel mailing list
[email protected]
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Reply via email to