On Mon, 19 Sep 2022, Andy Shevchenko wrote: > On Fri, Sep 16, 2022 at 06:44:49PM +0300, Andy Shevchenko wrote: > > On Fri, Sep 16, 2022 at 06:36:16PM +0300, Andy Shevchenko wrote: > > > On Fri, Sep 16, 2022 at 04:38:04PM +0300, Ilpo Järvinen wrote: > > > > Side note: > > > > $ git grep -n -w setup_irq -- drivers/tty/ > > drivers/tty/serial/8250/8250_core.c:382: .setup_irq = univ8250_setup_irq, > > drivers/tty/serial/8250/8250_port.c:2341: retval = up->ops->setup_irq(up); > > > > which rises a question of whether we need the setup_irq member in the > > respective structure. > > And to be confident, the files that include linux/serial_8250.h and have > setup_irq word in them (recursive `git grep`) in entire source tree: > > drivers/tty/serial/8250/8250_core.c > drivers/tty/serial/8250/8250_port.c > include/linux/serial_8250.h Thanks. I know about it already. The whole struct uart_8250_ops seems useless indirection layer (and IIRC the results of my history dig, it was added w/o any justification). It's on my kill list already :-) but I'll wait a bit with the cleanups to avoid -linus vs -next conflicts. -- i.
