On Mon, Jun 30, 2025 at 11:44:03AM +0200, Marek Vasut wrote: > On 6/30/25 3:32 AM, Dmitry Torokhov wrote: > > Hi Marek, > > Hi, > > > On Tue, Jun 10, 2025 at 02:54:12AM +0200, Marek Vasut wrote: > > > Add the capability of polling the touch controller for events every > > > 16ms, which is useful on hardware that did integrate this touch > > > controller, but did not integrate the IRQ line, like the RasPi . > > > > > > Make use of the generic input poller code. Factor out the code > > > from goodix_ts_irq_handler() into generic goodix_ts_handler(), so > > > it can be used both by the IRQ handler and poller callback. > > > > > > Use of_client->irq to find out whether the interrupt line is present > > > or not, independent of whether this is OF or ACPI system. It is not > > > possible to register poller in case request_irq() fails, because the > > > request_irq() in this driver is deliberately called after the input > > > device was registered, and registering the generic poller at that point > > > is too late already. > > > > > > Signed-off-by: Marek Vasut <marek.vasut+renesas@xxxxxxxxxxx> > > > > There was another version of this patch that was submitted and reviewed > > by Hans, so I merged it. > > Is it the following patch, with malformed commit message and some odd > "LF-15225" subject tag ? > > https://patchwork.kernel.org/project/linux-input/patch/20250522020418.1963422-1-qijian.guo@xxxxxxx/ Yes, I cleaned the commit message. > > > BTW I think this version would end up trying to disable IRQ0... > How come ? Won't we hit disable_irq(client->irq) call in goodix_suspend()? In polling mode client->irq is 0. Thanks. -- Dmitry
