Fri, Mar 21, 2025 at 09:37:09PM +0100, Benjamin Larsson kirjoitti: > On 10/02/2025 07:14, Greg KH wrote: > > On Sun, Feb 09, 2025 at 10:02:39PM +0100, Benjamin Larsson wrote: ... > I would argue that I follow the current flow of the code. In 8250.h we have: > > CONFIG_SERIAL_8250_PNP > CONFIG_SERIAL_8250_RSA These are historically parts of the main driver, RSA code theoretically can be removed. > CONFIG_SERIAL_8250_FINTEK I would love to see this being not part of main driver. > none of those enables COMPILE_TEST or the option to compile as a module. They all together may be compiled as a main driver module. Again, this is all historical and new code would need a very good justification why it can be held in a separate module. > Neither the Airoha code or the other code is not intended to be its own > separate module, Why not? > it is to be part of the 8250-driver. The 8250-driver can be > loaded as a module with or without the Airoha baud rate code. > > Implementing COMPILE_TEST when the 8250-driver does not support it seems > tricky. All the ways I could think of would result in messy code and logic. > I came to the conclusion that a smaller patch that reuses the current logic > was preferable. If that argument is not good enough then I need some > guidance how to implement something what would be accepted. -- With Best Regards, Andy Shevchenko
