[OpenWrt-Devel] Booting D-Link DNS-313, kernel requirements for OpenWRT?

[Linus Walleij]

On Wed, Feb 14, 2018 at 5:20 PM, Hans Ulli Kroll
<ulli.kroll at googlemail.com> wrote:
> On Tue, 13 Feb 2018, Linus Walleij wrote:
>> On Sun, Feb 11, 2018 at 11:56 PM, Linus Walleij
>> <linus.walleij at linaro.org> wrote:
>> > On Mon, Feb 5, 2018 at 11:38 PM, Linus Walleij <linus.walleij at linaro.org> wrote:
>> >> On Tue, Jan 16, 2018 at 11:22 PM, Linus Walleij <linus.walleij at linaro.org> wrote:
>> >>
>
> [...]
>
>>
>> So this is libubox returning 4 EINTR (interrupted system
>> call) from the uloop_process.
>>
>> waitpid() returns this while monitoring the process.
>>
>> I wonder why...
>>
>> I traced into the kernel and even checked that the
>> binfmt_script parses and tries to execute the script
>> using the right /bin/sh parser.
>>
>> It would be nice to be able to run strace here I guess.
>> kernel syscall debugging?
>>
>
> I will check this issue my DLINK box and maybe on my IB 4220 if I have
> time.

I dug around in it and didn't get smarter.

Patched the syscall tracer to print to console (since I
didn't have a working console that was the best option)
and what happens is askfirst and other compiled binaries
work fine but as soon as I try to run any busybox binary
(such as the shell) it dies with EINTR.

I don't know if busybox is the whole problem, there may be
something wrong with how it gets compiled or something.

> FWIW:
> For the fotg210 OTG controller (*)
>
> I have a (almost) full understanding which parts are changed from the
> Farday IP.
> For obvious speed reason they done the "usb gadget" work in the
> boodloader. I can achieve about 30Mib/s in device mode, which is
> impressive !

Wow!

Overall the Gemini silicon is pretty impressive. Once I get
OpenWRT up on this I will try to benchmark it a bit.

> For tests it should be possible to (re)use the vbus pin as id pin, so I
> can trigger the "internal" id pin in the syscon registers.
>
> The otg controller driver need a complete rework, already started.
>
> (*)
> As with my first approach the HCD driver is upstream, so also the the
> gadget driver.
> Adding the otg driver without breaking existing hardware is a
> challenging/fun/whatever work.

Hm :/

I will keep my fingers off it but I'm interested in any working
patches you have so I can test it.

My tree is pushed out and right now I am just trying to boot
OpenWRT, then the DSA switch in the DIR-685 needs a working
upstream driver, next driver (and last?) would be a driver for the
crypto block. But that is kind of optional so I wait with it
until the rest is working.

Yours,
Linus Walleij
_______________________________________________
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel