[PATCH v3] mfd: intel-lpss: Set d3hot_delay to 0

[PATCH v3] mfd: intel-lpss: Set d3hot_delay to 0

[Navon John Lukose]

Every runtime resume of an LPSS controller waits the default 10 ms
d3hot_delay in pci_power_up() for the D3hot->D0 transition. Nothing
misbehaves, but the wait adds latency. On Arrow Lake-H (Lenovo Yoga Pro 7
14IAH10) the touchpad's I2C controller autosuspends after 1 s idle, and
the i2c-hid interrupt handler for the first report afterwards takes
12.6-14.4 ms.

The PCI spec asks for those 10 ms, but Mika Westerberg pointed out that
the LPSS controllers are not really PCI-connected devices, so the delay
is likely not needed. intel-lpss-pci already clears d3cold_delay,
quirk_remove_d3hot_delay() drops d3hot_delay on several older Intel
chipsets, Cherry Trail among them, and commit 815b4448198f ("i3c:
mipi-i3c-hci-pci: Set d3hot_delay to 0 for Intel controllers") does the
same for the Intel I3C controllers.

Set d3hot_delay to 0 as well. The handler above then takes 1.1-1.9 ms,
with no I2C errors. Only Arrow Lake-H was tested.

Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Navon John Lukose <navonjohnlukose@gmail.com>
---

Notes:
    Changes in v3:
    - Describe the delay first, then why it is not needed (Ilpo).
    - Give Mika's full name (Ilpo).
    - Add Mika's Acked-by.
    
    Changes in v2:
    - Drop RFC and add the review discussion to the changelog (Andy).
    
    v2: https://lore.kernel.org/r/20261022040453.45797-1-navonjohnlukose@gmail.com
    RFC: https://lore.kernel.org/r/20261016114347.23633-1-navonjohnlukose@gmail.com

 drivers/mfd/intel-lpss-pci.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/mfd/intel-lpss-pci.c b/drivers/mfd/intel-lpss-pci.c
index 63a3fb58566..5672f033e5c 100644
--- a/drivers/mfd/intel-lpss-pci.c
+++ b/drivers/mfd/intel-lpss-pci.c
@@ -70,6 +70,7 @@ static int intel_lpss_pci_probe(struct pci_dev *pdev,
 		info->quirks = quirk_pci_info->driver_data;
 
 	pdev->d3cold_delay = 0;
+	pdev->d3hot_delay = 0;
 
 	/* Probably it is enough to set this for iDMA capable devices only */
 	pci_set_master(pdev);
-- 
2.56.0

Re: [PATCH v3] mfd: intel-lpss: Set d3hot_delay to 0

[Andy Shevchenko]

On Mon, Oct 26, 2026 at 01:06:13AM +0530, Navon John Lukose wrote:
> Every runtime resume of an LPSS controller waits the default 10 ms
> d3hot_delay in pci_power_up() for the D3hot->D0 transition. Nothing
> misbehaves, but the wait adds latency. On Arrow Lake-H (Lenovo Yoga Pro 7
> 14IAH10) the touchpad's I2C controller autosuspends after 1 s idle, and
> the i2c-hid interrupt handler for the first report afterwards takes
> 12.6-14.4 ms.
> 
> The PCI spec asks for those 10 ms, but Mika Westerberg pointed out that
> the LPSS controllers are not really PCI-connected devices, so the delay
> is likely not needed. intel-lpss-pci already clears d3cold_delay,
> quirk_remove_d3hot_delay() drops d3hot_delay on several older Intel
> chipsets, Cherry Trail among them, and commit 815b4448198f ("i3c:
> mipi-i3c-hci-pci: Set d3hot_delay to 0 for Intel controllers") does the
> same for the Intel I3C controllers.
> 
> Set d3hot_delay to 0 as well. The handler above then takes 1.1-1.9 ms,
> with no I2C errors. Only Arrow Lake-H was tested.

Okay, we may go with this and I believe if any issue happens somewhere we
will know that sooner than later.

Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v3] mfd: intel-lpss: Set d3hot_delay to 0

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 2319 bytes --]

On Mon, 26 Oct 2026, Navon John Lukose wrote:

> Every runtime resume of an LPSS controller waits the default 10 ms
> d3hot_delay in pci_power_up() for the D3hot->D0 transition. Nothing
> misbehaves, but the wait adds latency. On Arrow Lake-H (Lenovo Yoga Pro 7
> 14IAH10) the touchpad's I2C controller autosuspends after 1 s idle, and
> the i2c-hid interrupt handler for the first report afterwards takes
> 12.6-14.4 ms.
> 
> The PCI spec asks for those 10 ms, but Mika Westerberg pointed out that
> the LPSS controllers are not really PCI-connected devices, so the delay
> is likely not needed. intel-lpss-pci already clears d3cold_delay,
> quirk_remove_d3hot_delay() drops d3hot_delay on several older Intel
> chipsets, Cherry Trail among them, and commit 815b4448198f ("i3c:
> mipi-i3c-hci-pci: Set d3hot_delay to 0 for Intel controllers") does the
> same for the Intel I3C controllers.
> 
> Set d3hot_delay to 0 as well. The handler above then takes 1.1-1.9 ms,
> with no I2C errors. Only Arrow Lake-H was tested.

The new text reads much better, thanks,

Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>

--
 i.

> Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
> Signed-off-by: Navon John Lukose <navonjohnlukose@gmail.com>
> ---
> 
> Notes:
>     Changes in v3:
>     - Describe the delay first, then why it is not needed (Ilpo).
>     - Give Mika's full name (Ilpo).
>     - Add Mika's Acked-by.
>     
>     Changes in v2:
>     - Drop RFC and add the review discussion to the changelog (Andy).
>     
>     v2: https://lore.kernel.org/r/20261022040453.45797-1-navonjohnlukose@gmail.com
>     RFC: https://lore.kernel.org/r/20261016114347.23633-1-navonjohnlukose@gmail.com
> 
>  drivers/mfd/intel-lpss-pci.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/mfd/intel-lpss-pci.c b/drivers/mfd/intel-lpss-pci.c
> index 63a3fb58566..5672f033e5c 100644
> --- a/drivers/mfd/intel-lpss-pci.c
> +++ b/drivers/mfd/intel-lpss-pci.c
> @@ -70,6 +70,7 @@ static int intel_lpss_pci_probe(struct pci_dev *pdev,
>  		info->quirks = quirk_pci_info->driver_data;
>  
>  	pdev->d3cold_delay = 0;
> +	pdev->d3hot_delay = 0;
>  
>  	/* Probably it is enough to set this for iDMA capable devices only */
>  	pci_set_master(pdev);
>