linux-kernel - Re: [PATCH v2 02/34] component: Introduce the aggregate bus

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Thu, 7 Oct 2021 16:11:40 -0400
From:   Stephen Boyd <swboyd@...omium.org>
To:     Saravana Kannan <saravanak@...gle.com>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-kernel@...r.kernel.org, linux-arm-msm@...r.kernel.org,
        dri-devel@...ts.freedesktop.org, freedreno@...ts.freedesktop.org,
        Daniel Vetter <daniel.vetter@...ll.ch>,
        "Rafael J. Wysocki" <rafael@...nel.org>,
        Rob Clark <robdclark@...il.com>,
        Russell King <rmk+kernel@....linux.org.uk>
Subject: Re: [PATCH v2 02/34] component: Introduce the aggregate bus_type

Quoting Stephen Boyd (2021-10-07 11:40:07)
> Quoting Saravana Kannan (2021-10-06 20:07:11)
> > On Wed, Oct 6, 2021 at 12:38 PM Stephen Boyd <swboyd@...omium.org> wrote:
> > > diff --git a/drivers/base/component.c b/drivers/base/component.c
> > > index 0a41bbe14981..d99e99cabb99 100644
> > > --- a/drivers/base/component.c
> > > +++ b/drivers/base/component.c
> [...]
> > > +                       continue;
> > > +
> > > +               /* Matches put in component_del() */
> > > +               get_device(&adev->dev);
> > > +               c->link = device_link_add(&adev->dev, c->dev,
> > > +                                         DL_FLAG_STATELESS | DL_FLAG_PM_RUNTIME);
> >
> > Remove the STATELESS flag and you'll get a bunch of other stuff done for free:
>
> I tried that and it didn't work for me. The aggregate device never
> probed and I was left with no display. Let me see if I can reproduce it
> with logging to provide more details.

This patch fixes it (whitespace damaged sorry).

----8<----
diff --git a/drivers/base/component.c b/drivers/base/component.c
index 65042c9f8a42..43cac9ed70b7 100644
--- a/drivers/base/component.c
+++ b/drivers/base/component.c
@@ -202,7 +202,7 @@ static int find_components(struct aggregate_device *adev)
 		/* Matches put in component_del() */
 		get_device(&adev->dev);
 		c->link = device_link_add(&adev->dev, c->dev,
-					  DL_FLAG_STATELESS | DL_FLAG_PM_RUNTIME);
+					  DL_FLAG_PM_RUNTIME);
 		c->adev = adev;
 	}

@@ -749,7 +749,9 @@ static int __component_add(struct device *dev,
const struct component_ops *ops,
 	mutex_unlock(&component_mutex);

 	/* Try to bind */
-	return bus_rescan_devices(&aggregate_bus_type);
+	bus_rescan_devices(&aggregate_bus_type);
+
+	return 0;
 }

 /**


The important part is ignoring the return value of bus_rescan_devices().
It's a cycle problem. The last component is probing and calling
component_add() in its probe function. The call to component_add() is
trying to probe the aggregate device now that all components are added.
But when it tries to probe the aggregate device it sees that a supplier,
which is this component calling compnent_add(), hasn't been probed yet,
so it returns -EPROBE_DEFER. That is passed up to the component and it
defers probe.

I don't think the component device cares at all about the aggregate
device being able to probe or not. We should be able to ignore the
return value of bus_rescan_devices() in component_add(). I'll add a
comment to the code here so it's more obvious.