Re: [REGRESSION][BISECTED] Lenovo IdeaPad touchpad does not work when idma64 is present in initramfs

[Mika Westerberg <mika.westerberg@linux.intel.com>]

On Tue, Jan 27, 2026 at 05:56:37AM -0500, correctmost wrote:
> > Okay can you set CONFIG_INTEL_IDMA64=n in .config? That should prevent it
> > from loading. Keep the rest as it was.
> 
> That change makes the IRQ message go away and the touchpad seems to work
> fine in my desktop environment (log attached).
> 
> These are the config changes that I tested with:
> - CONFIG_I2C_DESIGNWARE_CORE=m
> - CONFIG_I2C_DESIGNWARE_PLATFORM=m
> - CONFIG_INTEL_IDMA64=n

Okay thanks it looks good to me too.

I tried to go through the datasheets again but could not figure why the
LPSS I2C IP would keep the interrupt asserted. However, I think we can
hack^Hwork this around by simply not creating the IDMA64 device at all for
I2C. This should be fine to do because I2C is really not using DMA in the
first place so this is kind of just wasting "memory". So something like
below.

Please re-enable CONFIG_INTEL_IDMA64=m and the rest and try if this makes
it work without the irq splat. And if it does please try with the
"original" arch setup and see if that works too (keeping the patch applied
of course).

Andy, what's your thoughts on this?

diff --git a/drivers/mfd/intel-lpss.c b/drivers/mfd/intel-lpss.c
index 63d6694f7145..088db88b10d5 100644
--- a/drivers/mfd/intel-lpss.c
+++ b/drivers/mfd/intel-lpss.c
@@ -246,7 +246,8 @@ static int intel_lpss_assign_devs(struct intel_lpss *lpss)
 
 static bool intel_lpss_has_idma(const struct intel_lpss *lpss)
 {
-	return (lpss->caps & LPSS_PRIV_CAPS_NO_IDMA) == 0;
+	return (lpss->caps & LPSS_PRIV_CAPS_NO_IDMA) == 0 &&
+		lpss->type != LPSS_DEV_I2C;
 }
 
 static void intel_lpss_set_remap_addr(const struct intel_lpss *lpss)