Hi Jason, I have currently seen both. I agree on the principles as it is simplifying the code a little bit. I including this clean up in this patch Add devicetree structure for a future v2 submission Best Regards Christophe -----Original Message----- From: Jason Gunthorpe [mailto:jgunthorpe@xxxxxxxxxxxxxxxxxxxx] Sent: mercredi 8 octobre 2014 18:29 To: Christophe RICARD Cc: peterhuewe@xxxxxx; ashley@xxxxxxxxxxxxx; tpmdd@xxxxxxxxxxxx; devicetree@xxxxxxxxxxxxxxx; tpmdd-devel@xxxxxxxxxxxxxxxxxxxxx; Christophe Henri RICARD; Jean-Luc BLANC Subject: Re: [tpmdd-devel] [PATCH 07/16] tpm/tpm_i2c_stm_st33: Add devicetree structure On Wed, Oct 08, 2014 at 07:47:58AM +0200, Christophe RICARD wrote: > >>+ if (interrupts) { > >>+ r = devm_gpio_request_one(&client->dev, pdata->io_serirq, > >>+ GPIOF_IN, "TPM IO_SERIRQ"); > >Similarly, I wonder if pdata->io_serirq is just duplication of > >client->irq and that should be set by the creator instead? > pdata->io_serirq stores the gpio number which will be converted into > irq number. pdata->io_serirq is only use by static platform > configuration not devicetree configuration Right, but the driver never uses it as a GPIO, so accepting a GPIO is actually less flexible - a platform may connect the TPM to a dedicated IRQ pin, for instance. The creator should just specify the irq in client->irq, however that is typically done.. Jason -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html