On 2017-02-24 14:31, Arend Van Spriel wrote:
On 24-2-2017 14:10, Rafał Miłecki wrote:
From: Rafał Miłecki <rafal@xxxxxxxxxx>
It avoids some unnecessary work. Most likely there wasn't much sense
in
doing this before bus reset anyway, as reset is supposed to put chip
into default state. In PCIe code (only bus implementing reset) we e.g.
use watchdog to perform a chip "reboot".
Sorry, but removing the fisrt passive call will cause chip lockups for
sure on certain systems. We learned the hard way :-p
OK, acknowledged! ;) I just tested it on my BCM43602 doing warm reboots
(they
were causing problems earlier) and it worked fine.
It also seems I don't use SDK recent enough as my reference :)
Kalle please drop this patch.
