[Bug 231080] callout structure corruption and panic

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=231080

--- Comment #23 from Mark Johnston <markj at FreeBSD.org> ---
(In reply to Sean Bruno from comment #22)
Thanks.  That suggests that the callout was scheduled after the llentry was
freed (or that I'm completely off-base), so let's use a bigger hammer.  Can you
try the following instead of my patch?

1. compile the kernel with "options DEBUG_MEMGUARD"
2. set vm.memguard.desc="512" after boot

I'm reading the llentry code and am concerned that the change of the if_adata
read lock to epoch_enter() breaks synchronization of updates to llentry state. 
In particular, IF_AFDATA_RLOCK and _WLOCK aren't mutually exclusive anymore,
but I suspect that we have some code that doesn't handle that quite correctly. 
However, I don't see a bug yet.

-- 
You are receiving this mail because:
You are the assignee for the bug.