Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 22 Oct 2013 18:18:25 -0700
From:      Adrian Chadd <adrian@freebsd.org>
To:        claudiu vasadi <claudiu.vasadi@gmail.com>
Cc:        "freebsd-hackers@freebsd.org" <freebsd-hackers@freebsd.org>, "freebsd-wireless@freebsd.org" <freebsd-wireless@freebsd.org>
Subject:   Re: 9.2-STABLE r255918 with GENERIC and iwn - core dump
Message-ID:  <CAJ-Vmommb%2BtNy5EbvDyxk6sAGRFVardQ9Sebjtyi0aR9%2BV0umA@mail.gmail.com>
In-Reply-To: <CAM-i3ijHDVG5XQ0LjfWHT%2BcyECja7VQRR9wXV71wEOeHr4eVDw@mail.gmail.com>
References:  <CAM-i3iiem_3-tv90k0NWeJjocx77RhT%2BCeZPHZRHZS3_AsgZkQ@mail.gmail.com> <CAJ-VmomY16hJSJgj5jAq9ywxCsY67AmmesPSSgsb1OekTGikww@mail.gmail.com> <CAM-i3ijHDVG5XQ0LjfWHT%2BcyECja7VQRR9wXV71wEOeHr4eVDw@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Grr, it's slightly more complicated than that.

That whole timer mess is actually used for two things:

* if the management transmit succeeds - it acts as a short-interval (a few
seconds) timer to ensure that the probe request ends up providing some
response that transitions to auth; otherwise it aborts and transitions back
to the SCAN state;
* if the management transmit fails - it immediately transitions back to
SCAN.

So, the "correct" fix involves correctly locking things and turning that
timer into a "if I fail to transition into a run state, then just go back
to scanning" timeout.

It sucks and it's going to take an evening of deep thought to figure out
the correct solution to.

Thanks for reminding me about this mess!




-adrian



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-Vmommb%2BtNy5EbvDyxk6sAGRFVardQ9Sebjtyi0aR9%2BV0umA>