[PATCH] clk: qcom: gcc-kaanapali: Enable FORCE_MEM_CORE_ON for UFS AXI PHY clock

[Joonhoe Kim]

As on SM8650, SM8750 and Eliza, the UFS AXI PHY clock needs
FORCE_MEM_CORE_ON for the UFS controller in MCQ mode. Without it the MCQ
configuration is lost over system suspend, and with MCQ enabled every
resume from s2idle ends in an SMMU fault and a host reset.

The UFS PHY ICE core clock already has the bit set; set it for the UFS
PHY AXI clock as well.

Fixes: d1919c375f21 ("clk: qcom: Add support for Global clock controller on Kaanapali")
Assisted-by: LLM
Signed-off-by: Joonhoe Kim <26rote@xxxxxxxxx>
---
Tested on a Lenovo Legion Tab Y700 Gen 5 with MCQ enabled in its DT:
the MCQ registers keep their values over s2idle, 7 of 7 resumes
without a UFS error. Kaanapali was noted as the missing one in the
Eliza thread [1].

[1] https://lore.kernel.org/all/a0b6030f-be42-4f5d-a59c-db78912ff0de@xxxxxxxxxxxxxxxx/

 drivers/clk/qcom/gcc-kaanapali.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/clk/qcom/gcc-kaanapali.c b/drivers/clk/qcom/gcc-kaanapali.c
index c9473fb79ccc..6834f43b72ca 100644
--- a/drivers/clk/qcom/gcc-kaanapali.c
+++ b/drivers/clk/qcom/gcc-kaanapali.c
@@ -3484,6 +3484,7 @@ static void clk_kaanapali_regs_configure(struct device *dev, struct regmap *regm
 
 	/* FORCE_MEM_CORE_ON for ufs phy ice core clocks */
 	qcom_branch_set_force_mem_core(regmap, gcc_ufs_phy_ice_core_clk, true);
+	qcom_branch_set_force_mem_core(regmap, gcc_ufs_phy_axi_clk, true);
 }
 
 static const struct qcom_cc_driver_data gcc_kaanapali_driver_data = {
-- 
2.43.0