diff mbox series

[1/1] Force ELAN06FA touchpad I2C bus freq to 100KHz

Message ID 20250103051657.211966-2-rha051117@gmail.com
State New
Headers show
Series [1/1] Force ELAN06FA touchpad I2C bus freq to 100KHz | expand

Commit Message

Randolph Ha Jan. 3, 2025, 5:16 a.m. UTC
Some devices do not define valid bus frequencies for the ELAN06FA
touchpad in their ACPI table, and some controllers run them at
400KHz by default. The 06FA touchpad exhibits excessive smoothing
behaviors when run at 400KHz, so force the bus frequency to 100KHz.

Signed-off-by: Randolph Ha <rha051117@gmail.com>
---
 drivers/i2c/i2c-core-acpi.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

Comments

Mika Westerberg Jan. 5, 2025, 8:33 a.m. UTC | #1
On Fri, Jan 03, 2025 at 05:46:27PM -0600, R Ha wrote:
> Hello,
> 
> Thanks for reading my patch!
> 
> On Fri, Jan 3, 2025 at 3:33 AM Mika Westerberg
> <mika.westerberg@linux.intel.com> wrote:
> > What are those "some devices" and "some controllers"?
> 
> The "Some Devices" are the Lenovo V15 G4 IRU, which I use, and
> potentially the Lenovo V15 G4 AMN and Lenovo Ideapad Slim 3 15IAH8 as
> well (based on issue reports from other users [1]).
> The "Some Controllers" are the Designware I2C controller.
> 
> Sorry for not putting this in the commit message; I had tried to
> follow the comments for the quirk I copied in Commit 7574c0db2e68c
> ("i2c: acpi: Force bus speed to 400KHz if a Silead touchscreen is
> present"), which left them out.

In general it is good to follow the existing changelogs but in this case I
would prefer to add the details of the system in question (so we know what
systems the quirk is applied to).

> On Fri, Jan 3, 2025 at 3:33 AM Mika Westerberg
> <mika.westerberg@linux.intel.com> wrote:
> > Can you add the ACPI table snippet here too for reference?
> 
> I believe this is the correct snippet in my ACPI table (Again, V15 G4
> IRU). Tried to edit it down as much as I could, hopefully this tells
> everything. Please let me know how I should attach a longer snippet or
> the full ACPI table if needed.

Okay thanks for sharing. I don't see the "SPED" beeing assigned in the
below snipped though. I would expect this works in Windows? Have you
checked if it uses 100 kHz or 400kHz there?

> Scope (_SB.PC00.I2C1)
> {
>     [...]
>     Device (TPD0)
>     {
>         [...]
>         CreateWordField (SBFB, \_SB.PC00.I2C1.TPD0._Y53._ADR, BADR)
> // _ADR: Address
>         CreateDWordField (SBFB, \_SB.PC00.I2C1.TPD0._Y53._SPE, SPED)
> // _SPE: Speed
>         CreateWordField (SBFG, 0x17, INT1)
>         CreateDWordField (SBFI, \_SB.PC00.I2C1.TPD0._Y54._INT, INT2)
> // _INT: Interrupts
>         Method (_INI, 0, NotSerialized)  // _INI: Initialize
>         {
>             If ((OSYS < 0x07DC))
>             {
>                 SRXO (0x09080011, One)
>             }
> 
>             INT1 = GNUM (0x09080011)
>             INT2 = INUM (0x09080011)
>             If ((TPTY == One))
>             {
>                 _HID = "ELAN06FA"
>                 _SUB = "ELAN0001"
>                 BADR = 0x15
>                 HID2 = One
>                 Return (Zero)
>             }
>             [...]
>         }
> 
>         Name (_HID, "XXXX0000")  // _HID: Hardware ID
>         Name (_CID, "PNP0C50" /* HID Protocol Device (I2C bus) */)  //
> _CID: Compatible ID
>         Name (_SUB, "XXXX0000")  // _SUB: Subsystem ID
>         Name (_S0W, 0x03)  // _S0W: S0 Device Wake State
>         Method (_DSM, 4, Serialized)  // _DSM: Device-Specific Method
>         {
>             If ((Arg0 == HIDG))
>             {
>                 Return (HIDD (Arg0, Arg1, Arg2, Arg3, HID2))
>             }
> 
>             If ((Arg0 == TP7G))
>             {
>                 Return (TP7D (Arg0, Arg1, Arg2, Arg3, SBFB, SBFG))
>             }
> 
>             Return (Buffer (One)
>             {
>                  0x00                                             // .
>             })
>         }
>         [...]
>         Method (_CRS, 0, NotSerialized)  // _CRS: Current Resource Settings
>         {
>             If ((OSYS < 0x07DC))
>             {
>                 Return (SBFI) /* \_SB_.PC00.I2C1.TPD0.SBFI */
>             }
> 
>             If ((TPDM == Zero))
>             {
>                 Return (ConcatenateResTemplate (I2CM (I2CX, BADR, SPED), SBFG))
>             }
> 
>             Return (ConcatenateResTemplate (I2CM (I2CX, BADR, SPED), SBFI))
>         }
>         [...]
>     }
> }
> 
> For comparison, the properties for a device that I think did set a
> proper speed was like this:
> If ((TPNP == 0xD64D))
> {
>     _HID = "GTCH7503"
>     HID2 = One
>     BADR = 0x10
>     SPED = 0x000F4240
>     Return (Zero)
> }
> 
> [1]: https://bbs.archlinux.org/viewtopic.php?id=297092
Randolph Ha Jan. 6, 2025, 9:08 a.m. UTC | #2
On Sun, Jan 5, 2025 at 2:34 AM Mika Westerberg
<mika.westerberg@linux.intel.com> wrote:
> I would expect this works in Windows? Have you
> checked if it uses 100 kHz or 400kHz there?

The touchpad does work in Windows. I couldn't find a way to check the
frequency there. I think that I need to do some more investigation
into how I can find the "proper" frequency, though if anyone has any
tips I would appreciate them greatly.

In Windows, my touchpad does not have a vendor-provided driver, but
uses the default I2C HID driver provided by Microsoft.
Actually, my touchpad is registered as two devices: An "I2C HID
Device" using the default "HIDI2C.inf" driver, and an "HID-compliant
touch pad", using the default "input.inf" driver. Both were provided
by Microsoft, so it appears there is no vendor-specific firmware here.

I initially assumed that the driver would default to 100KHz (as being
the lowest speed defined in
drivers/i2c/busses/i2c-designware-common.c), but after reading
Microsoft's documentation for I2C HID devices it appears that they
recommend 400KHz or more [1].
After reading this, I tried forcing the speed to 1MHz in the kernel,
which surprisingly also alleviated the laggy behavior. So it appears
that the speed can be either 100KHz or 1MHz, but I think that in
Windows it should be whatever the "default" defined by the Windows
driver is.

I tried finding the I2C frequency using the steps defined in the
Microsoft guide to tracing I2C HID events using Logman [2], and this
was (what I believe to be) the relevant snippet for setting up the I2C
device, though it does not appear there is anything related to the
controller frequency:
[1]0004.0140::01/06/2025-02:34:21.359
[hidi2c][WDFDEVICE:0x000018796F20AFD8] Received WDFREQUEST
0x000018797575D918 (IOCTL_HID_READ_REPORT) on WDFQUEUE
0x000018796F20AC08
[1]0004.0140::01/06/2025-02:34:21.368
[hidi2c][WDFDEVICE:0x000018796F20AFD8] Error retrieving read
WDFREQUEST 0x0000000000000000 from ReportQueue on interrupt
status:0x8000001a(STATUS_NO_MORE_ENTRIES)
[3]0004.3184::01/06/2025-02:34:21.369
[hidi2c][WDFDEVICE:0x000018796F20AFD8] D0Exit to target state D4
[0]0004.3184::01/06/2025-02:34:21.369
[hidi2c][WDFDEVICE:0x000018796F20AFD8] Power command (opcode:0x801)
sent to device register 0x5
[0]0004.3184::01/06/2025-02:34:21.369
[hidi2c][WDFDEVICE:0x000018796F20AFD8] Successfully put device into
sleep power state
[0]0004.3184::01/06/2025-02:34:21.369
[hidi2c][WDFDEVICE:0x000018796F20AFD8] Deinitialized device HID state
[0]0004.3184::01/06/2025-02:34:21.369
[hidi2c][WDFDEVICE:0x000018796F20AFD8] WdfObjectDelete closed and
deleted SpbIoTarget
[1]0004.0150::01/06/2025-02:34:23.172 [hidi2c]Created WDF driver
object:0xFFFFE78696EE0E30
[3]0004.0150::01/06/2025-02:34:23.180
[hidi2c][WDFDEVICE:0x00001879716173C8] Created new device
[1]0004.2364::01/06/2025-02:34:23.183
[hidi2c][WDFDEVICE:0x00001879716173C8] I2C resource found at index 0
with connection id: 0x1
[1]0004.2364::01/06/2025-02:34:23.183
[hidi2c][WDFDEVICE:0x00001879716173C8] Interrupt resource found at
index 1
[1]0004.2364::01/06/2025-02:34:23.183
[hidi2c][WDFDEVICE:0x00001879716173C8] Created Spb IO Target and
resources
[1]0004.2364::01/06/2025-02:34:23.184
[hidi2c][WDFDEVICE:0x00001879716173C8] Found supported DSM function:1
[1]0004.2364::01/06/2025-02:34:23.184
[hidi2c][WDFDEVICE:0x00001879716173C8] HID Descriptor register address
0x1 retrieved from ACPI
[1]0004.2364::01/06/2025-02:34:23.184
[hidi2c][WDFDEVICE:0x00001879716173C8] D0Entry from from previous
state D4
[1]0004.2364::01/06/2025-02:34:23.185
[hidi2c][WDFDEVICE:0x00001879716173C8] HID descriptor retrieved from
register 0x1 : LEN 30 VER 0x100 VID 0x4f3 PID 0x32b9 VER 0x4 RDL 679
RDA 0x2 IRA 0x3 IRML 31 ORA 0x4 ORML 0 CRA 0x5 DRA 0x6
[1]0004.2364::01/06/2025-02:34:23.185
[hidi2c][WDFDEVICE:0x00001879716173C8] Initialized device HID state
[3]0004.2364::01/06/2025-02:34:23.185
[hidi2c][WDFDEVICE:0x00001879716173C8] Power command (opcode:0x800)
sent to device register 0x5
[3]0004.2364::01/06/2025-02:34:23.185
[hidi2c][WDFDEVICE:0x00001879716173C8] Successfully put device into on
power state
[3]0004.2364::01/06/2025-02:34:23.185
[hidi2c][WDFDEVICE:0x00001879716173C8] WdfIoQueueRetrieveNextRequest
failed to find idle notification request in IdleQueue WDFQUEUE
0x000018796F1CCFD8 status:0x8000001a(STATUS_NO_MORE_ENTRIES)
[1]0004.2364::01/06/2025-02:34:23.186
[hidi2c][WDFDEVICE:0x00001879716173C8] Failed querying registry value
for DoNotWaitForResetResponsestatus:0xc0000034(STATUS_OBJECT_NAME_NOT_FOUND)
[1]0004.2364::01/06/2025-02:34:23.186
[hidi2c][WDFDEVICE:0x00001879716173C8] Starting timer for reset
response - IO queue stopped
[1]0004.2364::01/06/2025-02:34:23.186
[hidi2c][WDFDEVICE:0x00001879716173C8] Reset command sent to device
register 0x5
[0]0004.0140::01/06/2025-02:34:23.188
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC4C48 (IOCTL_HID_GET_DEVICE_DESCRIPTOR) on WDFQUEUE
0x000018796258EA28
[0]0004.0140::01/06/2025-02:34:23.188
[hidi2c][WDFDEVICE:0x00001879716173C8] Returning device descriptor to
hidclass, report descriptor length 679
[0]0004.0140::01/06/2025-02:34:23.188
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC4C48 (IOCTL_HID_GET_DEVICE_DESCRIPTOR) with
STATUS_SUCCESS
[0]0004.0140::01/06/2025-02:34:23.188
[hidi2c][WDFDEVICE:0x00001879716173C8] Received reset response -
starting IO queue
[3]0004.0150::01/06/2025-02:34:23.189
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC4C48 (IOCTL_HID_GET_DEVICE_ATTRIBUTES) on WDFQUEUE
0x000018796258EA28
[3]0004.0150::01/06/2025-02:34:23.189
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC4C48 (IOCTL_HID_GET_DEVICE_ATTRIBUTES) with
STATUS_SUCCESS
[3]0004.0150::01/06/2025-02:34:23.189
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC4C48 (IOCTL_HID_GET_REPORT_DESCRIPTOR) on WDFQUEUE
0x000018796258EA28
[3]0004.0150::01/06/2025-02:34:23.209
[hidi2c][WDFDEVICE:0x00001879716173C8] Report descriptor of length 679
retrieved from register 0x2
[3]0004.0150::01/06/2025-02:34:23.209
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC4C48 (IOCTL_HID_GET_REPORT_DESCRIPTOR) with
STATUS_SUCCESS
[3]0004.0150::01/06/2025-02:34:23.209
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC4C48 (IOCTL_HID_READ_REPORT) on WDFQUEUE
0x000018796258EA28
[3]0004.0150::01/06/2025-02:34:23.209
[hidi2c][WDFDEVICE:0x00001879716173C8] Received first ping pong read -
enabling interrupt processing
[3]0004.0150::01/06/2025-02:34:23.209
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC48C8 (IOCTL_HID_READ_REPORT) on WDFQUEUE
0x000018796258EA28
[2]0790.0A3C::01/06/2025-02:34:23.225
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC5A48 (IOCTL_HID_GET_STRING) on WDFQUEUE
0x000018796258EA28
[2]0790.0A3C::01/06/2025-02:34:23.225
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC5A48 (IOCTL_HID_GET_STRING) with STATUS_SUCCESS
[2]0790.0A3C::01/06/2025-02:34:23.225
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC5A48 (IOCTL_HID_GET_FEATURE) on WDFQUEUE
0x000018796258EA28
[0]0790.0A3C::01/06/2025-02:34:23.233
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC5A48 (IOCTL_HID_GET_FEATURE) with STATUS_SUCCESS
[0]0790.0A3C::01/06/2025-02:34:23.234
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC5A48 (IOCTL_HID_GET_FEATURE) on WDFQUEUE
0x000018796258EA28
[2]0790.0A3C::01/06/2025-02:34:23.242
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC5A48 (IOCTL_HID_GET_FEATURE) with STATUS_SUCCESS
[2]0790.0A3C::01/06/2025-02:34:23.242
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC5A48 (IOCTL_HID_DEVICERESET_NOTIFICATION) on WDFQUEUE
0x000018796258EA28
[2]0790.0A3C::01/06/2025-02:34:23.242
[hidi2c][WDFDEVICE:0x00001879716173C8] Pended device reset
notification WDFREQUEST 0x0000187962CC5A48 to WDFQUEUE
0x0000187967F9EB08
[8]0004.020C::01/06/2025-02:34:23.243
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC3208 (IOCTL_HID_SET_FEATURE) on WDFQUEUE
0x000018796258EA28
[0]0004.0300::01/06/2025-02:34:23.243
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC1608 (IOCTL_HID_SET_FEATURE) on WDFQUEUE
0x000018796258EA28
[2]0004.020C::01/06/2025-02:34:23.244
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC3208 (IOCTL_HID_SET_FEATURE) with STATUS_SUCCESS
[2]0004.0300::01/06/2025-02:34:23.244
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC1608 (IOCTL_HID_SET_FEATURE) with STATUS_SUCCESS
[2]0790.0A3C::01/06/2025-02:34:23.247
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC1608 (IOCTL_HID_SET_FEATURE) on WDFQUEUE
0x000018796258EA28
[8]0004.0300::01/06/2025-02:34:23.247
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC3208 (IOCTL_HID_SET_FEATURE) on WDFQUEUE
0x000018796258EA28
[9]0004.020C::01/06/2025-02:34:23.247
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC0F08 (IOCTL_HID_SET_FEATURE) on WDFQUEUE
0x000018796258EA28
[2]0790.0A3C::01/06/2025-02:34:23.247
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC1608 (IOCTL_HID_SET_FEATURE) with STATUS_SUCCESS
[2]0004.0300::01/06/2025-02:34:23.248
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC3208 (IOCTL_HID_SET_FEATURE) with STATUS_SUCCESS
[2]0004.020C::01/06/2025-02:34:23.248
[hidi2c][WDFDEVICE:0x00001879716173C8] Completing WDFREQUEST
0x0000187962CC0F08 (IOCTL_HID_SET_FEATURE) with STATUS_SUCCESS
[3]0004.0140::01/06/2025-02:34:23.267
[hidi2c][WDFDEVICE:0x00001879716173C8] Received WDFREQUEST
0x0000187962CC0F08 (IOCTL_HID_READ_REPORT) on WDFQUEUE
0x000018796258EA28

[1]: https://learn.microsoft.com/en-us/windows-hardware/design/component-guidelines/touchpad-device-bus-connectivity#acpi-table-entries
[2]: https://learn.microsoft.com/en-us/windows-hardware/drivers/hid/event-tracing#using-logmanexe
diff mbox series

Patch

diff --git a/drivers/i2c/i2c-core-acpi.c b/drivers/i2c/i2c-core-acpi.c
index 14ae0cfc325e..b10f52e12fe8 100644
--- a/drivers/i2c/i2c-core-acpi.c
+++ b/drivers/i2c/i2c-core-acpi.c
@@ -355,6 +355,18 @@  static const struct acpi_device_id i2c_acpi_force_400khz_device_ids[] = {
 	{}
 };
 
+static const struct acpi_device_id i2c_acpi_force_100khz_device_ids[] = {
+	/*
+	 * When a 400KHz freq is used on this model of ELAN touchpad instead
+	 * of 100Khz, excessive smoothing (similar to when there is noise in
+	 * the signal) is intermittently applied. As some devices' ACPI
+	 * tables do not specify the 100KHz frequency requirement, it is
+	 * necessary to force the speed to 100KHz.
+	 */
+	{ "ELAN06FA", 0 },
+	{}
+};
+
 static acpi_status i2c_acpi_lookup_speed(acpi_handle handle, u32 level,
 					   void *data, void **return_value)
 {
@@ -373,6 +385,9 @@  static acpi_status i2c_acpi_lookup_speed(acpi_handle handle, u32 level,
 	if (acpi_match_device_ids(adev, i2c_acpi_force_400khz_device_ids) == 0)
 		lookup->force_speed = I2C_MAX_FAST_MODE_FREQ;
 
+	if (acpi_match_device_ids(adev, i2c_acpi_force_100khz_device_ids) == 0)
+		lookup->force_speed = I2C_MAX_STANDARD_MODE_FREQ;
+
 	return AE_OK;
 }