Re: Deferred if_start

To: Ryota Ozaki <ozaki-r%netbsd.org@localhost>
Subject: Re: Deferred if_start
From: Joerg Sonnenberger <joerg%bec.de@localhost>
Date: Fri, 2 Dec 2016 10:32:45 +0100

On Fri, Dec 02, 2016 at 04:07:24PM +0900, Ryota Ozaki wrote:
> We need it for MP-ification of bpf because some drivers call
> if_start (or equivalent, say drvX_start) from hardware
> interrupt, which means that bpf_mtap is also called from
> hardware interrupt. That makes MP-ification hard.

What is the reason for bpf_mtap to fail in hardware interrupt context?
I really don't like this requirement change.

Joerg

Follow-Ups:
- Re: Deferred if_start
  - From: Ryota Ozaki

References:
- Deferred if_start
  - From: Ryota Ozaki

Prev by Date: Deferred if_start
Next by Date: Re: Deferred if_start
Previous by Thread: Deferred if_start
Next by Thread: Re: Deferred if_start
Indexes:

Home | Main Index | Thread Index | Old Index