lkml.org 
[lkml]   [2015]   [Mar]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 6/8] amd64_edac: enforce synchronous probe
On Wed, Mar 18, 2015 at 10:45:44AM -0700, Dmitry Torokhov wrote:
> On Wed, Mar 18, 2015 at 12:56:18PM -0400, Tejun Heo wrote:
> > On Fri, Jan 16, 2015 at 03:33:15PM -0800, Dmitry Torokhov wrote:
> > > From: Luis R. Rodriguez <mcgrof@suse.com>
> > >
> > > While testing asynchronous PCI probe on this driver I noticed it failed
> > > so enforce just synchronouse probe for now. Asynchronous probe is not
> > > used by default and requires userepace intervention. Patches for its
> > > support will be merged later.
> > >
> > > The reason async probe fails is that the init call for this driver
> > > relies on probe to have finished for at least one device. This needs to
> > > be addressed before enabling async probe.
> >
> > I'm still kinda uncomfortable with this both white and black list
> > behavior. If we're gonna do this, let's please drop the debug options
> > and build proper blacklists; otherwise, this will never be complete
> > and we're gonna left with the in-between situation forever.
>
> Without the debug options how can we do that? I will definitely not be
> able to go through all the in-tree drivers myself and see if they can be
> asynchronously probed or not. The most I can do is to try enabling the
> option on our side and fixing the drivers/subsystems that fail with
> asynchronous probing. This will be iterative process for some time and
> then we'll drop the debug option and flip the flag to do asynchronous
> probing by default.

By the way, at that point I think we should be able to remove the
FORCE_SYNCHRONOUS option (and maybe PREFER_ASYNCHRONOUS as well?).

Thanks.

--
Dmitry


\
 
 \ /
  Last update: 2015-03-18 19:01    [W:0.185 / U:0.068 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site