linux-kernel - Re: [PATCH] rpmsg: Make rpmsg sample selectable

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:	Wed, 11 May 2016 12:45:29 -0700
From:	Bjorn Andersson <bjorn.andersson@...aro.org>
To:	Suman Anna <s-anna@...com>
Cc:	Ohad Ben-Cohen <ohad@...ery.com>, linux-remoteproc@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] rpmsg: Make rpmsg sample selectable

On Wed 11 May 08:39 PDT 2016, Suman Anna wrote:

> On 05/06/2016 05:15 PM, Bjorn Andersson wrote:
> > On Fri 06 May 12:25 PDT 2016, Suman Anna wrote:
> > 
> >> Hi Bjorn,
> >>
> >> On 05/06/2016 01:06 PM, Bjorn Andersson wrote:
> >>> rpmsg is not user selectable so the rpmsg sample should not depend on
> >>> it, as this limits the possibility of compiling the sample client
> >>> without any other clients.
> >>>
> >>> Signed-off-by: Bjorn Andersson <bjorn.andersson@...aro.org>
> >>> ---
> >>>  samples/Kconfig | 3 ++-
> >>>  1 file changed, 2 insertions(+), 1 deletion(-)
> >>>
> >>> diff --git a/samples/Kconfig b/samples/Kconfig
> >>> index d54f28c6dc5e..61a251e579ed 100644
> >>> --- a/samples/Kconfig
> >>> +++ b/samples/Kconfig
> >>> @@ -57,7 +57,8 @@ config SAMPLE_KDB
> >>>  
> >>>  config SAMPLE_RPMSG_CLIENT
> >>>  	tristate "Build rpmsg client sample -- loadable modules only"
> >>> -	depends on RPMSG && m
> >>> +	select RPMSG
> >>> +	depends on m
> >>
> >> Well, I believe this has to do with the fact that we also need to have a
> >> remoteproc driver enabled, as you cannot have the sample running by
> >> itself. A remoteproc driver implementing RPMSG would select it, and then
> >> this sample can be enabled. In fact, I follow this for all rpmsg bus
> >> drivers.
> >>
> > 
> > That does sound backwards though. Remoteproc provides means to control
> > the life cycle of your remote processor and establish (among other
> > things) the virtio channels for e.g. rpmsg, but it certainly does not
> > depend on rpmsg.
> 
> True, but the boot of a remote processor that supports virtio devices is
> automatic only when you have the virtio_rpmsg_bus probed because of
> the link between remoteproc_virtio and virtio_rpmsg_bus.
> 

If I read the code correctly we register the virtios found in the
resource table, which will trigger a probe of a virtio driver with the
matching "id".

The probed vdev will acquire the virtqueues through the find_vqs op,
which will upon matching boot the remoteproc.

So as far as I understand this does not only apply to rpmsg, but to any
virtio driver. Not saying that the "depends" is wrong, just trying to
better understand the relationship between these components.

> > 
> > As far as I can see there's 8 other types of virtio services supported
> > and I presume we're supposed to select those based on the device's need.
> > 
> > 
> > But based on this I believe the correct solution is not to fix the
> > sample but to make RPMSG selectable in itself.
> 
> Yeah, agreed.
> 

I'll send a patch for that next time I'm messing around in there.

Regards,
Bjorn