Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To:     linux-kernel@vger.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        stable@vger.kernel.org,
        Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>,
        Ranjani Sridharan <ranjani.sridharan@linux.intel.com>,
        Rander Wang <rander.wang@intel.com>,
        Bard Liao <yung-chuan.liao@linux.intel.com>,
        Vinod Koul <vkoul@kernel.org>, Sasha Levin <sashal@kernel.org>
Subject: [PATCH 5.15 039/247] soundwire: intel: prevent pm_runtime resume prior to system suspend
Date:   Mon, 13 Jun 2022 12:09:01 +0200
Message-Id: <20220613094924.131527944@linuxfoundation.org>
In-Reply-To: <20220613094922.843438024@linuxfoundation.org>
References: <20220613094922.843438024@linuxfoundation.org>
User-Agent: quilt/0.66
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>

[ Upstream commit 6d9f2dadba698114fed97b224578c5338a36b0d9 ]

commit e38f9ff63e6d ("ACPI: scan: Do not add device IDs from _CID if _HID is not valid")
exposes a race condition on a TGL RVP device leading to a timeout.

The detailed analysis shows the RT711 codec driver scheduling a jack
detection workqueue while attaching during a spurious pm_runtime
resume, and the work function happens to be scheduled after the
manager device is suspended.

The direct link between this ACPI patch and a spurious pm_runtime
resume is not obvious; the most likely explanation is that a change in
the ACPI device linked list management modifies the order in which the
pm_runtime device status is checked and exposes a race condition that
was probably present for a very long time, but was not identified.

We already have a check in the .prepare stage, where we will resume to
full power from specific clock-stop modes. In all other cases, we
don't need to resume to full power by default. Adding the
SMART_SUSPEND flag prevents the spurious resume from happening.

BugLink: https://github.com/thesofproject/linux/issues/3459
Fixes: 029bfd1cd53cd ("soundwire: intel: conditionally exit clock stop mode on system suspend")
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Link: https://lore.kernel.org/r/20220420023241.14335-2-yung-chuan.liao@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/soundwire/intel.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/soundwire/intel.c b/drivers/soundwire/intel.c
index f72d36654ac2..38e7f1a2bb97 100644
--- a/drivers/soundwire/intel.c
+++ b/drivers/soundwire/intel.c
@@ -1298,6 +1298,9 @@ static int intel_link_probe(struct auxiliary_device *auxdev,
 	/* use generic bandwidth allocation algorithm */
 	sdw->cdns.bus.compute_params = sdw_compute_params;
 
+	/* avoid resuming from pm_runtime suspend if it's not required */
+	dev_pm_set_driver_flags(dev, DPM_FLAG_SMART_SUSPEND);
+
 	ret = sdw_bus_master_add(bus, dev, dev->fwnode);
 	if (ret) {
 		dev_err(dev, "sdw_bus_master_add fail: %d\n", ret);
-- 
2.35.1