On 2017-04-24 19:29, Andy Shevchenko wrote:
> There are Exar quirks in 8250_port which belong to 8250_exar module.
> Move them to the correct module and do not contaminate generic code with it.
>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
> ---
> Just compile tested. I have no hardware to test this approach. It would
> be nice if you guys (Jan, Sudip) can take it and continue on real
> hardware. Why am I interested in this? It would make my life slightly
> easier for runtime PM support developing for 8250.
>
> TODO:
> - move get_divisor() / set_divisor() quirks as well
> - perhaps move startup() / shutdown() either
> - patch can be split a bit since there is at least one useless check
> in exar_handle_irq() for port type (this might be separate clean up)
>
> I would be pleased if this just works after being applied.
>
> drivers/tty/serial/8250/8250_exar.c | 57 ++++++++++++++++++++++++++++-
> drivers/tty/serial/8250/8250_port.c | 73 -------------------------------------
> 2 files changed, 56 insertions(+), 74 deletions(-)
>
...
> @@ -684,18 +681,6 @@ EXPORT_SYMBOL_GPL(serial8250_rpm_put_tx);
> static void serial8250_set_sleep(struct uart_8250_port *p, int sleep)
> {
> unsigned char lcr = 0, efr = 0;
> - /*
> - * Exar UARTs have a SLEEP register that enables or disables
> - * each UART to enter sleep mode separately. On the XR17V35x the
> - * register is accessible to each UART at the UART_EXAR_SLEEP
> - * offset but the UART channel may only write to the corresponding
> - * bit.
> - */
> - if ((p->port.type == PORT_XR17V35X) ||
> - (p->port.type == PORT_XR17D15X)) {
> - serial_out(p, UART_EXAR_SLEEP, sleep ? 0xff : 0);
> - return;
> - }
>
> if (p->capabilities & UART_CAP_SLEEP) {
> if (p->capabilities & UART_CAP_EFR) {
Which patch(es) does this depend on? The hunk above is different in
Linus' tree and also tty.git.
Jan
Attachment:
signature.asc
Description: OpenPGP digital signature
