From: George Moussalem <george.moussalem@outlook.com>

The sleep clock never be disabled. To avoid the kernel trying to disable
it and keep it always on, flag it as critical.

Fixes: e3fdbef1bab8 ("clk: qcom: Add Global Clock controller (GCC) driver for IPQ5018")
Signed-off-by: George Moussalem <george.moussalem@outlook.com>
---
additional context:

This issue occurs on any device referencing the GCC_SLEEP_CLK_SRC clock,
as required by drivers to bring up the Q6 processor and WCSS, and making
a call to devm_clk_get_enabled (or similar) implicitly registers a call
to clk_disable_unprepare which causes below kernel error. This clock
should never be disabled as also evidenced in the downstream driver:
https://git.codelinaro.org/clo/qsdk/oss/kernel/linux-ipq-5.4/-/blob/NHSS.QSDK.12.5/drivers/clk/qcom/gcc-ipq5018.c?ref_type=heads#L1424

[    3.012124] ------------[ cut here ]------------
[    3.012976] gcc_sleep_clk_src status stuck at 'on'
[    3.013022] WARNING: CPU: 0 PID: 1 at drivers/clk/qcom/clk-branch.c:87 clk_branch_toggle+0x168/0x180
[    3.022181] Modules linked in:
[    3.031465] CPU: 0 UID: 0 PID: 1 Comm: swapper/0 Not tainted 6.12.59 #0
[    3.034328] Hardware name: Linksys MR5500 (DT)
[    3.040837] pstate: 604000c5 (nZCv daIF +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[    3.045354] pc : clk_branch_toggle+0x168/0x180
[    3.052208] lr : clk_branch_toggle+0x168/0x180
[    3.056722] sp : ffffffc082393ac0
[    3.061147] x29: ffffffc082393ac0 x28: 0000000000000000 x27: 0000000000000000
[    3.064539] x26: 0000000000000000 x25: 0000000000000000 x24: ffffffc080aac890
[    3.071655] x23: 0000000000000000 x22: 0000000000000000 x21: ffffffc08047c2a8
[    3.078773] x20: ffffffc0822d62c8 x19: 0000000000000000 x18: ffffffc08224b878
[    3.085892] x17: ffffffbf9dd04000 x16: ffffffc082378000 x15: 00000000000000b7
[    3.093010] x14: 00000000000000b7 x13: 00000000ffffffea x12: ffffffc0822a3820
[    3.100128] x11: ffffffc08224b878 x10: ffffffc0822a3878 x9 : 0000000000000001
[    3.107247] x8 : 0000000000000001 x7 : 0000000000017fe8 x6 : c0000000ffffefff
[    3.114364] x5 : 0000000000057fa8 x4 : 0000000000000000 x3 : ffffffc0823938a0
[    3.121482] x2 : ffffffc08224b7a0 x1 : ffffffc08224b7a0 x0 : 0000000000000026
[    3.128603] Call trace:
[    3.135713]  clk_branch_toggle+0x168/0x180
[    3.137973]  clk_branch2_disable+0x1c/0x30
[    3.142137]  clk_core_disable+0x60/0xac
[    3.146218]  clk_disable+0x30/0x4c
[    3.149950]  clk_disable_unprepare+0x18/0x30
[    3.153423]  devm_clk_release+0x24/0x40
[    3.157850]  devres_release_all+0xb0/0x120
[    3.161409]  device_unbind_cleanup+0x18/0x60
[    3.165577]  really_probe+0x210/0x2bc
[    3.170003]  __driver_probe_device+0x78/0x118
[    3.173562]  driver_probe_device+0x40/0xf0
[    3.177902]  __driver_attach+0x90/0x160
[    3.181895]  bus_for_each_dev+0x64/0xb8
[    3.185627]  driver_attach+0x24/0x3c
[    3.189447]  bus_add_driver+0xe4/0x208
[    3.193267]  driver_register+0x68/0x124
[    3.196825]  __platform_driver_register+0x24/0x30
[    3.200559]  wcss_sec_driver_init+0x1c/0x28
[    3.205419]  do_one_initcall+0x50/0x210
[    3.209413]  kernel_init_freeable+0x23c/0x298
[    3.213232]  kernel_init+0x20/0x120
[    3.217746]  ret_from_fork+0x10/0x20
[    3.221046] ---[ end trace 0000000000000000 ]---
[    3.225373] remoteproc remoteproc0: releasing q6wcss
---
 drivers/clk/qcom/gcc-ipq5018.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/clk/qcom/gcc-ipq5018.c b/drivers/clk/qcom/gcc-ipq5018.c
index dcda2be8c1a51950248050882620d63d75eb1ca5..64792cda06202157441222a0bdbf6dc883343054 100644
--- a/drivers/clk/qcom/gcc-ipq5018.c
+++ b/drivers/clk/qcom/gcc-ipq5018.c
@@ -1340,6 +1340,7 @@ static struct clk_branch gcc_sleep_clk_src = {
 			.name = "gcc_sleep_clk_src",
 			.parent_data = gcc_sleep_clk_data,
 			.num_parents = ARRAY_SIZE(gcc_sleep_clk_data),
+			.flags = CLK_IS_CRITICAL,
 			.ops = &clk_branch2_ops,
 		},
 	},

---
base-commit: 7d31f578f3230f3b7b33b0930b08f9afd8429817
change-id: 20251128-ipq5018-sleep-clk-fix-1ff61966fcf8

Best regards,
-- 
George Moussalem <george.moussalem@outlook.com>

On 11/28/25 12:03 PM, George Moussalem via B4 Relay wrote:
> From: George Moussalem <george.moussalem@outlook.com>
> 
> The sleep clock never be disabled. To avoid the kernel trying to disable
> it and keep it always on, flag it as critical.
> 
> Fixes: e3fdbef1bab8 ("clk: qcom: Add Global Clock controller (GCC) driver for IPQ5018")
> Signed-off-by: George Moussalem <george.moussalem@outlook.com>
> ---

Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

Konrad