Re: [dpdk-dev] [PATCH 1/2] virtio: initialize iopl when device is initialized

[David Marchand <david.marchand@6wind.com>]

So, a little summary.

On Fri, Mar 6, 2015 at 5:20 PM, Stephen Hemminger <
stephen@networkplumber.org> wrote:

> > > The virtio driver needs to use in/out instructions therefore it must
> initialize
> > > using iopl(2) system call. The problem is that virtio initialization
> happens very
> > > early, and any application that uses daemon() or calls eal_init later
> in another
> > > context will fail.
>

Part of this is wrong.
The manual tells that this should be inherited.
I added some fork and daemon to test programs of mine which confirmed the
manual seems to be right.

The real problem is that iopl is not called at "constructor" time for it to
be inherited by irq thread.


> The issue is that virtio has no place it can do iopl() and have the IRQ
> thread
> work. It only shows up on real code where application is daemon, not in a
> toy
> demo or test application.
>

Contrary to what you asserted, the problem does happen when using testpmd
(if this is the toy you are talking about).
So, I would say your real world application is no better than testpmd.


I will post the right fixes.

-- 
David Marchand