Re: [PATCH RFC] HID: cp2112: add IRQ chip handling

On Mon, Nov 21, 2016 at 12:07 PM, Benjamin Tissoires
[off-list ref] wrote:

The GPIO part doesn't provide interrupts when GPIO are toggled.
So use a polling mechanism if someone requests a GPIO as an IRQ.

Signed-off-by: Benjamin Tissoires <redacted>
---

I am currently running this patch in my local tree.
It allows to attach a SMBus device completely on the CP2112, with a simple
definition at the end of probe:

        {
                struct i2c_client *client;
                struct i2c_hid_platform_data pdata = {
                        .hid_descriptor_address = 0x20,
                };
                struct i2c_board_info synaptics_info = {
                        I2C_BOARD_INFO("hid", 0x2c),
                        .platform_data = &pdata,
                };
                int irq = cp2112_allocate_irq(dev, 2);

                if (irq <= 0) {
                        dev_err(dev->gc.parent, "Failed to translate GPIO to IRQ\n");
                        goto err_sysfs_remove;
                }

                synaptics_info.irq = irq;
                irq_set_irq_type(irq, IRQ_TYPE_LEVEL_LOW);

                hid_device_io_start(hdev);

                /* give time for the device to initialize */
                msleep(500);

                client = i2c_new_device(&dev->adap, &synaptics_info);
                if (!client)
                        hid_err(hdev, "failed allocating Synaptics device\n");

        }

I wonder if we want to consider this upstream, given that the driver is only
used for development.

Ciao Benjamin,

for me it makes sense to push it upstream.
I also coded something similar 2+ years ago, but then my cp2112 got
broken before I could clean-up and submit the patches.
As far as I remember, Ellen was using cp2112 in a system in production.
For sure this driver is not widely used, but definitively it's not for
development use only.

Best Regards
Antonio