[PATCH] clk: qcom: gcc-sm8750: Use retention for USB power domains

[PATCH] clk: qcom: gcc-sm8750: Use retention for USB power domains

[Vladyslav Tymoshchuk]

The USB controller does not expect to lose its state in system suspend.
The controller node is a wakeup source, so in host mode dwc3 keeps the
core initialized across suspend, while genpd still powers the USB GDSCs
off. With a USB stick on the port, the xHCI comes back from s2idle with
an SMMU fault and is given up:

  arm-smmu 15000000.iommu: Unhandled context fault: fsr=0x402, iova=0x669fd3200, fsynr=0x370011, cbfrsynra=0x40, cb=5
  xhci-hcd xhci-hcd.3.auto: WARNING: Host System Error
  xhci-hcd xhci-hcd.3.auto: xHCI host not responding to stop endpoint command
  xhci-hcd xhci-hcd.3.auto: xHCI host controller not responding, assume dead
  xhci-hcd xhci-hcd.3.auto: HC died; cleaning up

After that the port is dead in device mode as well, until a reboot.

Keep the controller and PHY GDSCs in retention, as SM8650 and X1E80100
do.

Fixes: 3267c774f3ff ("clk: qcom: Add support for GCC on SM8750")
Signed-off-by: Vladyslav Tymoshchuk <tvlad744@gmail.com>
---
Tested on OnePlus Pad 3 (SM8750; the board DT is not upstream yet),
v7.3-rc5 plus board patches, 150 s of s2idle ended by an RTC alarm,
with a USB stick on the Type-C port in host mode:

                                    without    with this patch
  stick left in during suspend      1/1 died   2/2 fine
  stick unplugged during suspend    2/2 died   3/3 fine

Two more 25 s cycles with the stick left in were fine with the patch.

Without the patch and with power/wakeup of the controller set to
"disabled" (one boot, two 25 s cycles) the first resume still hit the
SMMU fault and "Host System Error" and lost the stick; the second one
ended in "xHC error in resume, USBSTS 0x101d, Reinit" and the stick
came back.

Device mode (NCM gadget) suspends and resumes with the patch as it
did without it. I did not measure the suspend current.

I used an AI assistant (Claude) for the analysis and to draft this
patch and its description; I reviewed the result and ran the tests on
my device.

 drivers/clk/qcom/gcc-sm8750.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/clk/qcom/gcc-sm8750.c b/drivers/clk/qcom/gcc-sm8750.c
index 80b8ec119c55..080f2fde53a1 100644
--- a/drivers/clk/qcom/gcc-sm8750.c
+++ b/drivers/clk/qcom/gcc-sm8750.c
@@ -2940,7 +2940,7 @@ static struct gdsc gcc_usb30_prim_gdsc = {
 	.pd = {
 		.name = "gcc_usb30_prim_gdsc",
 	},
-	.pwrsts = PWRSTS_OFF_ON,
+	.pwrsts = PWRSTS_RET_ON,
 	.flags = POLL_CFG_GDSCR | RETAIN_FF_ENABLE,
 };
 
@@ -2952,7 +2952,7 @@ static struct gdsc gcc_usb3_phy_gdsc = {
 	.pd = {
 		.name = "gcc_usb3_phy_gdsc",
 	},
-	.pwrsts = PWRSTS_OFF_ON,
+	.pwrsts = PWRSTS_RET_ON,
 	.flags = POLL_CFG_GDSCR | RETAIN_FF_ENABLE,
 };
 

base-commit: 14825c9204bad7f4e0bd3a528cbf4a7e24e757d0
-- 
2.43.0