Am 2021-02-24 19:19, schrieb Vladimir Oltean:
On Wed, Feb 24, 2021 at 07:08:51PM +0100, Michael Walle wrote:
Am 2021-02-24 18:45, schrieb Vladimir Oltean:
> On Wed, Feb 24, 2021 at 06:29:39PM +0100, Michael Walle wrote:
> > What is the reason to remove that device in the first place? Like is
> > this really a valid scenario? I really don't know when a device is
> > removed and if its remove, will it still be there or is it rather
> > a hot-plug type and rebinding it won't work anyways.
>
> Did you get this to crash under any circumstance other than using the
> 'unbind' command?
Nope, thus I was curious about that comment in dsa_port_stop().
Someone
(Alex, Claudiu maybe?) must have something in mind when writing about
it.
But I couldn't figure out in which case a device is removed.
I'm pretty sure that the checks that are in place now were once written
so that the sandbox tests would pass. If they still do, we should be
fine.
Ah.
You can run the sandbox tests using:
Just if one is trying to follow this thread: you'll also need to
have the following series applied:
https://patchwork.ozlabs.org/project/uboot/list/?series=229778
make sandbox_defconfig NO_SDL=1
make -j 8 NO_SDL=1
./u-boot -d ./arch/sandbox/dts/test.dtb
btw theres a shortcut for this "u-boot -T"
setenv ethact swp0
"setenv ethact lan0" I guess
ping 1.2.3.5
ut dm dsa_probe
ut dm dsa
ut dm
ut dm net_retry
Not more failures than without my patch.
-michael
