linux-kernel - [PATCH] PM: domains: Don't attach a device to genpd that corresponds to a provider

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-next>] [day] [month] [year] [list]

Message-Id: <20210709125611.135920-1-ulf.hansson@linaro.org>
Date:   Fri,  9 Jul 2021 14:56:11 +0200
From:   Ulf Hansson <ulf.hansson@...aro.org>
To:     "Rafael J . Wysocki" <rjw@...ysocki.net>, linux-pm@...r.kernel.org
Cc:     Bjorn Andersson <bjorn.andersson@...aro.org>,
        Stephen Boyd <sboyd@...nel.org>,
        Saravana Kannan <saravanak@...gle.com>,
        Dmitry Baryshkov <dmitry.baryshkov@...aro.org>,
        Kevin Hilman <khilman@...nel.org>,
        Geert Uytterhoeven <geert+renesas@...der.be>,
        Ulf Hansson <ulf.hansson@...aro.org>,
        linux-kernel@...r.kernel.org
Subject: [PATCH] PM: domains: Don't attach a device to genpd that corresponds to a provider

According to the common power domain DT bindings, a power domain provider
must have a "#power-domain-cells" property in its OF node. Additionally, if
a provider has a "power-domains" property, it means that it has a parent
domain.

It has turned out that some OF nodes that represents a genpd provider may
also be compatible with a regular platform device. This leads to, during
probe, genpd_dev_pm_attach(), genpd_dev_pm_attach_by_name() and
genpd_dev_pm_attach_by_id() tries to attach the corresponding struct device
to the genpd provider's parent domain, which is wrong. Instead the genpd
provider should only assign a parent domain, through
pm_genpd_add_subdomain() or of_genpd_add_subdomain().

Let's fix this problem in genpd by checking for the "#power-domain-cells"
property in the OF node for the struct device that is about to be attached.

Signed-off-by: Ulf Hansson <ulf.hansson@...aro.org>
---

This turned up when I was reviewing a couple of genpd realted Qcom patches [1],
from Dmitry Baryshkov (on cc).

I haven't been able to run much tests, so help with that would be greatly
appreciated. If someone encounter issues, I am happy to help to fix those.

Kind regards
Ulf Hansson

[1]
https://lkml.org/lkml/2021/7/9/7
https://lkml.org/lkml/2021/7/9/8

---
 drivers/base/power/domain.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c
index a934c679e6ce..21991fb56a19 100644
--- a/drivers/base/power/domain.c
+++ b/drivers/base/power/domain.c
@@ -2699,6 +2699,10 @@ int genpd_dev_pm_attach(struct device *dev)
 	if (!dev->of_node)
 		return 0;

+	/* Don't try to attach a genpd provider. */
+	if (of_find_property(dev->of_node, "#power-domain-cells", NULL))
+		return NULL;
+
 	/*
 	 * Devices with multiple PM domains must be attached separately, as we
 	 * can only attach one PM domain per device.
@@ -2737,6 +2741,10 @@ struct device *genpd_dev_pm_attach_by_id(struct device *dev,
 	if (!dev->of_node)
 		return NULL;

+	/* Don't try to attach a genpd provider. */
+	if (of_find_property(dev->of_node, "#power-domain-cells", NULL))
+		return NULL;
+
 	/* Verify that the index is within a valid range. */
 	num_domains = of_count_phandle_with_args(dev->of_node, "power-domains",
 						 "#power-domain-cells");
-- 
2.25.1