Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Received-SPF: Pass (protection.outlook.com: domain of nvidia.com designates
 216.228.112.36 as permitted sender) receiver=protection.outlook.com;
 client-ip=216.228.112.36; helo=mail.nvidia.com;
From:   Om Prakash Singh <omp@nvidia.com>
To:     <vidyas@nvidia.com>, <lorenzo.pieralisi@arm.com>,
        <bhelgaas@google.com>, <thierry.reding@gmail.com>,
        <jonathanh@nvidia.com>
CC:     <linux-tegra@vger.kernel.org>, <linux-pci@vger.kernel.org>,
        <linux-kernel@vger.kernel.org>, <kthota@nvidia.com>,
        <mmaddireddy@nvidia.com>, Om Prakash Singh <omp@nvidia.com>
Subject: [RESEND PATCH V1 3/5] PCI: tegra: Disable interrupts before entering L2
Date:   Thu, 27 May 2021 17:22:44 +0530
Message-ID: <20210527115246.20509-4-omp@nvidia.com>
In-Reply-To: <20210527115246.20509-1-omp@nvidia.com>
References: <20210527115246.20509-1-omp@nvidia.com>
MIME-Version: 1.0
Content-Type: text/plain
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 27 May 2021 11:53:06.7142
 (UTC)
X-MS-Exchange-CrossTenant-Network-Message-Id: 83425fd2-10ff-46e1-6503-08d92105fdeb
X-MS-Exchange-CrossTenant-Id: 43083d15-7273-40c1-b7db-39efd9ccc17a
X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=43083d15-7273-40c1-b7db-39efd9ccc17a;Ip=[216.228.112.36];Helo=[mail.nvidia.com]
X-MS-Exchange-CrossTenant-AuthSource: DM6NAM11FT066.eop-nam11.prod.protection.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Anonymous
X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem
X-MS-Exchange-Transport-CrossTenantHeadersStamped: DM5PR1201MB0156
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

Tegra194 implements suspend_noirq() hook and during the system suspend, the link
is taken to L2 state after PME_Turn_off handshake and if it doesn't go into L2,
PERST# is asserted. It is observed that with some of the endpoints (Ex:- Marvell
SATA controller), the link doesn't go into L2 state and asserting PERST# results
in Surprise Link Down error and the corresponding AER interrupt is also raised.
Since the system is in noirq phase, this interrupt is not served. Both PME and
AER interrupts are served by the same wire interrupt in Tegra194, and since the
PCIe sub-system enables wake capability for PME interrupt, having a pending AER
interrupt is treated as PME wake interrupt by the system and prevents the system
going into the suspend state. To address this issue, the interrupts are disabled
before taking the link into L2 state as the interrupts are not expected anyway
from the controller afterward.

Signed-off-by: Om Prakash Singh <omp@nvidia.com>
---
 drivers/pci/controller/dwc/pcie-tegra194.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/pci/controller/dwc/pcie-tegra194.c b/drivers/pci/controller/dwc/pcie-tegra194.c
index 58fc2615014d..ae62fdc840e6 100644
--- a/drivers/pci/controller/dwc/pcie-tegra194.c
+++ b/drivers/pci/controller/dwc/pcie-tegra194.c
@@ -1593,6 +1593,16 @@ static void tegra_pcie_dw_pme_turnoff(struct tegra_pcie_dw *pcie)
 		return;
 	}
 
+	/*
+	 * PCIe controller exits from L2 only if reset is applied, so
+	 * controller doesn't handle interrupts. But in cases where
+	 * L2 entry fails, PERST# is asserted which can trigger surprise
+	 * link down AER. However this function call happens in
+	 * suspend_noirq(), so AER interrupt will not be processed.
+	 * Disable all interrupts to avoid such a scenario.
+	 */
+	appl_writel(pcie, 0x0, APPL_INTR_EN_L0_0);
+
 	if (tegra_pcie_try_link_l2(pcie)) {
 		dev_info(pcie->dev, "Link didn't transition to L2 state\n");
 		/*
-- 
2.17.1