Juan Quintela <quint...@redhat.com> writes: > Ricardo Ribalda Delgado <ricardo.riba...@gmail.com> wrote: >> Hello Juan >> >> Thanks for your comments. About the indentation error... Do you have >> some kind of auto indent script(like the kernel code has). It is >> making me crazy trying to collaborate with a lot of projects an all of >> them with different styles. > > Don't even start this discussion yet again :) > Short answer: we don't have. > Long answer: any indenter that would pass qemu code style will be turing > complete (at least) and possibly will also pass the Turing test.
Among the things I'm paid for is tolerating ugly code. >>>> +#include <stdlib.h> >>>> +#include "../qemu-common.h" >>>> +#include "../qemu-char.h" >>>> +#include "../console.h" >>> >>> You can remove the "../" from those, Makefile sets correct include paths >>> for this to work. >> >> >> Ok. I used the mssmouse.c as reference. I can change that. (I guess >> that I should also replace "" with <>) Use #include "name.h" for includes that come with the source or are generated by the build. > Everything uses "". I just looked when reviewing this patch at msmouse, > and that one could also take some cleanup. This happens a lot in qemu, > you search for another driver for inspiration, and Murphy gets just the > one with the wrong examples. > >>> Why does the lenght of the FIFO changes here? I think this change in >>> independent of the rest of the patch (no knowledge about the 16550A to >>> know if it should be 16 or 32). >> >> I have to send a 2x10 bytes package, and it does not fit the the 16 >> bytes buffer.... Any other suggestion about how to do it? > > Nope, I am not a 16550A guru at all. No sure if your change will break > anything else or no, that is why I asked. The 16550 FIFOs are 16 bytes. The chips with larger FIFOs have different model numbers, e.g. 16650 with 32 byte FIFOs. I can't see how the FIFO length limits your packet size. Could you explain?
