carl9170: fix spurious restart due to high latency
authorChristian Lamparter <chunkeey@googlemail.com>
Fri, 29 Oct 2010 21:26:13 +0000 (23:26 +0200)
committerJohn W. Linville <linville@tuxdriver.com>
Mon, 15 Nov 2010 18:25:32 +0000 (13:25 -0500)
RX Stress tests of unidirectional bulk traffic with
bitrates of up to 220Mbit/s have revealed that the
fatal-event recovery logic [which was solely triggered
by an out-of-rx-buffer situation] is too aggressive.

The new method now "pings" the device and then
decides - based on the response - whenever
a restart is needed or not.

Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

No differences found