Re: [PATCH] irq_poll: Use raise_softirq_irqoff() in cpu_dead notifier

[Christoph Hellwig]

On Thu, Sep 30, 2021 at 12:37:54PM +0200, Sebastian Andrzej Siewior wrote:
> __raise_softirq_irqoff() adds a bit to the pending sofirq mask and this
> is it. The softirq won't be handled in a deterministic way but randomly
> when an interrupt fires and handles softirq in its irq_exit() routine or
> if something randomly checks and handles pending softirqs in the call
> chain before the CPU goes idle.
> 
> Add a local_bh_disable/enable() around the IRQ-off section which will
> handle pending softirqs.

This patch leaves me extremely confused, and it would even more if I was
just reading the code.  local_irq_disable is supposed to disable BHs
as well, so the code looks pretty much nonsensical to me.  But
apparently that isn't the point if I follow your commit message as you
don't care about an extra level of BH disabling but want to force a
side-effect of the re-enabling?  Why not directly call the helper
to schedule the softirq then?