[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250821-sm8650-power-sequence-fix-v3-1-645816ba3826@quicinc.com>
Date: Thu, 21 Aug 2025 16:24:38 +0530
From: Dikshita Agarwal <quic_dikshita@...cinc.com>
To: Vikash Garodia <quic_vgarodia@...cinc.com>,
Dikshita Agarwal
<quic_dikshita@...cinc.com>,
Abhinav Kumar <abhinav.kumar@...ux.dev>,
"Bryan
O'Donoghue" <bryan.odonoghue@...aro.org>,
Mauro Carvalho Chehab
<mchehab@...nel.org>
CC: <linux-media@...r.kernel.org>, <linux-arm-msm@...r.kernel.org>,
<linux-kernel@...r.kernel.org>,
Neil Armstrong <neil.armstrong@...aro.org>
Subject: [PATCH v3] media: iris: vpu3x: Add MNoC low power handshake during
hardware power-off
Add the missing write to AON_WRAPPER_MVP_NOC_LPI_CONTROL before
reading the LPI status register. Introduce a handshake loop to ensure
MNoC enters low power mode reliably during VPU3 hardware power-off with
timeout handling.
Tested-by: Neil Armstrong <neil.armstrong@...aro.org> # on SM8650-QRD
Tested-by: Neil Armstrong <neil.armstrong@...aro.org> # on SM8650-HDK
Signed-off-by: Dikshita Agarwal <quic_dikshita@...cinc.com>
---
Changes in v3:
- Fixed the loop to capture the success of the last power-on command (Bryan)
- Link to v2: https://lore.kernel.org/r/20250813-sm8650-power-sequence-fix-v2-1-9ed0fc2c45cb@quicinc.com
Changes in v2:
- Restructured loop for readability (Jorge)
- Used defines for bits (Konrad, Jorge)
- Used udelay for short waits (Konrad)
- Link to v1: https://lore.kernel.org/r/20250812-sm8650-power-sequence-fix-v1-1-a51e7f99c56c@quicinc.com
---
drivers/media/platform/qcom/iris/iris_vpu3x.c | 32 +++++++++++++++++++++++++--
1 file changed, 30 insertions(+), 2 deletions(-)
diff --git a/drivers/media/platform/qcom/iris/iris_vpu3x.c b/drivers/media/platform/qcom/iris/iris_vpu3x.c
index 9b7c9a1495ee2f51c60b1142b2ed4680ff798f0a..bfc52eb04ed0e1c88efe74a8d27bb95e8a0ca331 100644
--- a/drivers/media/platform/qcom/iris/iris_vpu3x.c
+++ b/drivers/media/platform/qcom/iris/iris_vpu3x.c
@@ -19,6 +19,9 @@
#define WRAPPER_IRIS_CPU_NOC_LPI_CONTROL (WRAPPER_BASE_OFFS + 0x5C)
#define REQ_POWER_DOWN_PREP BIT(0)
#define WRAPPER_IRIS_CPU_NOC_LPI_STATUS (WRAPPER_BASE_OFFS + 0x60)
+#define NOC_LPI_STATUS_DONE BIT(0) /* Indicates the NOC handshake is complete */
+#define NOC_LPI_STATUS_DENY BIT(1) /* Indicates the NOC handshake is denied */
+#define NOC_LPI_STATUS_ACTIVE BIT(2) /* Indicates the NOC is active */
#define WRAPPER_CORE_CLOCK_CONFIG (WRAPPER_BASE_OFFS + 0x88)
#define CORE_CLK_RUN 0x0
@@ -109,7 +112,9 @@ static void iris_vpu3_power_off_hardware(struct iris_core *core)
static void iris_vpu33_power_off_hardware(struct iris_core *core)
{
+ bool handshake_done = false, handshake_busy = false;
u32 reg_val = 0, value, i;
+ u32 count = 0;
int ret;
if (iris_vpu3x_hw_power_collapsed(core))
@@ -128,13 +133,36 @@ static void iris_vpu33_power_off_hardware(struct iris_core *core)
goto disable_power;
}
+ /* Retry up to 1000 times as recommended by hardware documentation */
+ do {
+ /* set MNoC to low power */
+ writel(REQ_POWER_DOWN_PREP, core->reg_base + AON_WRAPPER_MVP_NOC_LPI_CONTROL);
+
+ udelay(15);
+
+ value = readl(core->reg_base + AON_WRAPPER_MVP_NOC_LPI_STATUS);
+
+ handshake_done = value & NOC_LPI_STATUS_DONE;
+ handshake_busy = value & (NOC_LPI_STATUS_DENY | NOC_LPI_STATUS_ACTIVE);
+
+ if (handshake_done || !handshake_busy)
+ break;
+
+ writel(0, core->reg_base + AON_WRAPPER_MVP_NOC_LPI_CONTROL);
+
+ udelay(15);
+
+ } while (++count < 1000);
+
+ if (!handshake_done && handshake_busy)
+ dev_err(core->dev, "LPI handshake timeout\n");
+
ret = readl_poll_timeout(core->reg_base + AON_WRAPPER_MVP_NOC_LPI_STATUS,
reg_val, reg_val & BIT(0), 200, 2000);
if (ret)
goto disable_power;
- /* set MNoC to low power, set PD_NOC_QREQ (bit 0) */
- writel(BIT(0), core->reg_base + AON_WRAPPER_MVP_NOC_LPI_CONTROL);
+ writel(0, core->reg_base + AON_WRAPPER_MVP_NOC_LPI_CONTROL);
writel(CORE_BRIDGE_SW_RESET | CORE_BRIDGE_HW_RESET_DISABLE,
core->reg_base + CPU_CS_AHB_BRIDGE_SYNC_RESET);
---
base-commit: d968e50b5c26642754492dea23cbd3592bde62d8
change-id: 20250812-sm8650-power-sequence-fix-ba9a92098233
Best regards,
--
Dikshita Agarwal <quic_dikshita@...cinc.com>
Powered by blists - more mailing lists