svn commit: r332860 - head/sys/kern

[Warner Losh]

On Mon, Apr 23, 2018 at 9:12 AM, Jonathan T. Looney <jtl at freebsd.org>
wrote:
>
> If we leave the default alone, I agree we should print the assertion
> message (albeit with some rate limit).
>

We should print the first N asserts we hit during a kernel panic core dump,
then stop. I'd suggest N should be 10 or 20.

For me, the only asserts that have value during a panic are the ones that
prevent bad things from happening. The only bad thing that really can
happen is either wedging the machine, so we don't finish the dump
(watchdogs solve that) or writing to a naughty place (defined as not within
the bounds of a swap partition). I'm having a hard time seeing in what
actual value assertions that aren't related to either of these areas buy us.

Warner