Re: Easily reproducible stable/13 kernel crash

On 2021-Jun-11, at 00:02, Herbert J. Skuhra <herbert at gojira.at> wrote:

> On Fri, 11 Jun 2021 08:00:56 +0200, "Herbert J. Skuhra" wrote:
>> 
>> On Fri, 11 Jun 2021 07:42:27 +0200, Matthew Grooms wrote:
>>> 
>>> Hey all,
>>> 
>>> If I build an up-to-date stable/13 arm64 kernel and type sysctl -a on
>>> a rpi4 system, it reboots every time. I've blown away the source and
>>> object tree multiple times and still get the same result.
>>> 
>>> FreeBSD generic 13.0-STABLE FreeBSD 13.0-STABLE #1
>>> stable/13-n245932-04c4bd7f7b5: Fri Jun 11 00:09:11 CDT 2021
>>> mgrooms@x.x.x:/var/rpi4/build/usr/src/arm64.aarch64/sys/GENERIC arm64
>>> 
>>> Has anyone else been testing stable builds recently? I'll try to build
>>> a debug kernel and see if I can dig up more info.
>> 
>> I can confirm this issue: Rasperry Pi 2 (armv7), 3 and 4 (both aarch64).
>> When booting kernel.old (from May 17th; e0f2b8aaf1ed) 'sysctl -a'
>> doesn't crash the system.
> 
> The kernel from the May 27th snapshot (stable/13-n245691-024a9aa7010)
> is OK, the one from June 3rd (13-n245852-4775325dd66) isn't.

I tried the stable/13-n245765-bec0d2c9c841 based bectl environment
that I happen to have in place on a RPi4B:

# sysctl net.inet.tcp.lro
net.inet.tcp.lro.sackwakeups: 0
net.inet.tcp.lro.lockcnt: 0
net.inet.tcp.lro.single: 0
net.inet.tcp.lro.compressed: 0
net.inet.tcp.lro.wokeup: 0
net.inet.tcp.lro.fullqueue: 0
net.inet.tcp.lro.entries: 8
net.inet.tcp.lro.hold_lock: 0

# sysctl -a > /dev/null
#

No crash. That might narrow the range a bit unless some other
context-specific issue is involved.

===
Mark Millard
marklmi at yahoo.com
( dsl-only.net went
away in early 2018-Mar)