Re: 9.1: forcible coredump -> panic

[Jason Thorpe <thorpej%me.com@localhost>]

> On Apr 9, 2026, at 4:27 PM, Mouse <mouse%rodents-montreal.org@localhost> wrote:
> 
> That certainly would qualify as the stupid thing I was doing which I
> figured must exist!
> 
> As for "dispatch[ing] to another kernel thread"...can a soft-int create
> a kernel thread, or do I need to precreate the thread and have it in my
> back pocket so I just need to wake it up when the callout fires?

You either need to have the thread already created and waiting, or you need to have set up a work queue job that you can fire off.

In general, the soft-int threads are allowed to block to acquire a mutex, but they’re not allowed to block to allocate memory.

> Might this be a good use case for workqueue(9)?  The manpage makes it
> look as though that would manage the thread(s) and queue(s) and such
> for me, which feels like a significant benefit in this case.

Yes.

-- thorpej