[PATCH] clk: eswin: Add CLK_IGNORE_UNUSED to NoC clock

[PATCH] clk: eswin: Add CLK_IGNORE_UNUSED to NoC clock

[dongxuyang]

From: Xuyang Dong <dongxuyang@eswincomputing.com>

The gate_noc_nsp_clk provides the essential clock source for NPU,
DSP, and PCIe subsystems. During kernel init, the clock framework
attempts to disable unused clocks when clk_ignore_unused kernel
parameter is not set.
However, gate_noc_nsp_clk is required to remain enabled for these
critical subsystems to function properly, causing PCIe boot failures
when auto-disabled.

Add CLK_IGNORE_UNUSED flag to gate_noc_nsp_clk to ensure it stays
enabled even when clk_ignore_unused is not specified in kernel
command line.

Fixes: cd44f127c1d4 ("clk: eswin: Add eic7700 clock driver")
Signed-off-by: Xuyang Dong <dongxuyang@eswincomputing.com>
---
 drivers/clk/eswin/clk-eic7700.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/clk/eswin/clk-eic7700.c b/drivers/clk/eswin/clk-eic7700.c
index be81d74192da..43a47fe16ab1 100644
--- a/drivers/clk/eswin/clk-eic7700.c
+++ b/drivers/clk/eswin/clk-eic7700.c
@@ -791,7 +791,8 @@ static struct eswin_clk_info eic7700_clks[] = {
 			EIC7700_CLK_MUX_CPU_ROOT_3MUX1_GFREE,
 			CLK_SET_RATE_PARENT, EIC7700_REG_OFFSET_U84, 27, 0),
 	ESWIN_GATE_TYPE(EIC7700_CLK_GATE_NOC_NSP_CLK, "gate_noc_nsp_clk",
-			EIC7700_CLK_DIV_NOC_NSP_DYNM, CLK_SET_RATE_PARENT,
+			EIC7700_CLK_DIV_NOC_NSP_DYNM,
+			CLK_SET_RATE_PARENT | CLK_IGNORE_UNUSED,
 			EIC7700_REG_OFFSET_NOC, 31, 0),
 	ESWIN_GATE_TYPE(EIC7700_CLK_GATE_BOOTSPI, "gate_clk_bootspi",
 			EIC7700_CLK_MUX_BOOTSPI_CLK_2MUX1_GFREE,
-- 
2.34.1

Re: [PATCH] clk: eswin: Add CLK_IGNORE_UNUSED to NoC clock

[Brian Masney]

On Fri, Jun 05, 2026 at 05:21:18PM +0800, dongxuyang@eswincomputing.com wrote:
> From: Xuyang Dong <dongxuyang@eswincomputing.com>
> 
> The gate_noc_nsp_clk provides the essential clock source for NPU,
> DSP, and PCIe subsystems. During kernel init, the clock framework
> attempts to disable unused clocks when clk_ignore_unused kernel
> parameter is not set.
> However, gate_noc_nsp_clk is required to remain enabled for these
> critical subsystems to function properly, causing PCIe boot failures
> when auto-disabled.
> 
> Add CLK_IGNORE_UNUSED flag to gate_noc_nsp_clk to ensure it stays
> enabled even when clk_ignore_unused is not specified in kernel
> command line.
> 
> Fixes: cd44f127c1d4 ("clk: eswin: Add eic7700 clock driver")
> Signed-off-by: Xuyang Dong <dongxuyang@eswincomputing.com>

Reviewed-by: Brian Masney <bmasney@redhat.com>

FYI, I posted a series to implement sync_state support for the clk
subsystem, which should fix these type of issues for the long term. If
you happen to have time, I would appreciate it if you have time to test
this series on your hardware, and drop a Tested-by.

https://lore.kernel.org/linux-clk/20260603-clk-sync-state-v1-0-457120eed200@redhat.com/

I have a new series ready that I'm planning to post tomorrow. I'm
waiting to hear back about the ongoing conversation on that thread.

Once this lands, there shouldn't be much need for the CLK_IGNORE_UNUSED
flag. I'm curious if you revert this patch in your tree if the need for the
flag goes away with the sync_state changes applied.

Let's get your change in the tree though. If we merged sync_state today,
it won't be eligible to be backported to the stable kernels, and the
CLK_IGNORE_UNUSED flag is an easier backport.

Brian

Re: Re: [PATCH] clk: eswin: Add CLK_IGNORE_UNUSED to NoC clock

[Xuyang Dong]

> 
> On Fri, Jun 05, 2026 at 05:21:18PM +0800, dongxuyang@eswincomputing.com wrote:
> > From: Xuyang Dong <dongxuyang@eswincomputing.com>
> > 
> > The gate_noc_nsp_clk provides the essential clock source for NPU,
> > DSP, and PCIe subsystems. During kernel init, the clock framework
> > attempts to disable unused clocks when clk_ignore_unused kernel
> > parameter is not set.
> > However, gate_noc_nsp_clk is required to remain enabled for these
> > critical subsystems to function properly, causing PCIe boot failures
> > when auto-disabled.
> > 
> > Add CLK_IGNORE_UNUSED flag to gate_noc_nsp_clk to ensure it stays
> > enabled even when clk_ignore_unused is not specified in kernel
> > command line.
> > 
> > Fixes: cd44f127c1d4 ("clk: eswin: Add eic7700 clock driver")
> > Signed-off-by: Xuyang Dong <dongxuyang@eswincomputing.com>
> 
> Reviewed-by: Brian Masney <bmasney@redhat.com>
> 
> FYI, I posted a series to implement sync_state support for the clk
> subsystem, which should fix these type of issues for the long term. If
> you happen to have time, I would appreciate it if you have time to test
> this series on your hardware, and drop a Tested-by.
> 
> https://lore.kernel.org/linux-clk/20260603-clk-sync-state-v1-0-457120eed200@redhat.com/
> 
> I have a new series ready that I'm planning to post tomorrow. I'm
> waiting to hear back about the ongoing conversation on that thread.
> 
> Once this lands, there shouldn't be much need for the CLK_IGNORE_UNUSED
> flag. I'm curious if you revert this patch in your tree if the need for the
> flag goes away with the sync_state changes applied.
> 
> Let's get your change in the tree though. If we merged sync_state today,
> it won't be eligible to be backported to the stable kernels, and the
> CLK_IGNORE_UNUSED flag is an easier backport.
> 

Hi Brian,

Thank you for your patch set above. However, we still need this fix patch 
at the moment, because gate_noc_nsp_clk still needs to be enabled when 
accessing PCIe.

Best regards,
Xuyang Dong

Re: Re: [PATCH] clk: eswin: Add CLK_IGNORE_UNUSED to NoC clock

[Brian Masney]

On Tue, Jun 23, 2026 at 2:58 AM Xuyang Dong
<dongxuyang@eswincomputing.com> wrote:
> > Reviewed-by: Brian Masney <bmasney@redhat.com>
> >
> > FYI, I posted a series to implement sync_state support for the clk
> > subsystem, which should fix these type of issues for the long term. If
> > you happen to have time, I would appreciate it if you have time to test
> > this series on your hardware, and drop a Tested-by.
> >
> > https://lore.kernel.org/linux-clk/20260603-clk-sync-state-v1-0-457120eed200@redhat.com/
> >
> > I have a new series ready that I'm planning to post tomorrow. I'm
> > waiting to hear back about the ongoing conversation on that thread.
> >
> > Once this lands, there shouldn't be much need for the CLK_IGNORE_UNUSED
> > flag. I'm curious if you revert this patch in your tree if the need for the
> > flag goes away with the sync_state changes applied.
> >
> > Let's get your change in the tree though. If we merged sync_state today,
> > it won't be eligible to be backported to the stable kernels, and the
> > CLK_IGNORE_UNUSED flag is an easier backport.
> >
>
> Hi Brian,
>
> Thank you for your patch set above. However, we still need this fix patch
> at the moment, because gate_noc_nsp_clk still needs to be enabled when
> accessing PCIe.

Yes, I agree this is also needed.

Brian