Alan Cox writes: > On Thu, 24 May 2007 09:45:37 -0400 > David Woodhouse <[EMAIL PROTECTED]> wrote: > > > On Thu, 2007-05-24 at 14:41 +0100, Alan Cox wrote: > > > Most people copied the x86 behaviour which makes it easy to transplant. > > > Some are just smoking something (see ioctls.h for sh-64 and weep), others > > > have slightly odd behaviour for historical compatibility reasons (Sparc) > > > > Likewise, I assume the lack of IBSHIFT on PowerPC is because of AIX? > > I assume nobody ever got around to it. CIBAUD is in the PowerPC System V > API supplement for example and is supported by other Power OS's.
The guys that did the original ppc linux port (mainly Gary Thomas and Cort Dougan, I believe) tended to follow the alpha and x86 linux ports rather than the System V ABI. There aren't any SysV-based OSes for PowerPC still extant so it doesn't really matter. And yes, nobody has got around to implementing CIBAUD/IBSHIFT. > > Why bother introducing new IBSHIFT stuff when it can be declared > > obsolete already -- if you want different input and output baud rates, > > just set BOTHER and have different values c_ispeed and c_ospeed. > > BOTHER is for the output speed. BOTHER << IBSHIFT is for the input speed > if it different to the default (and you need B0 << IBSHIFT to know if an > input speed is being specifically set in the first place). > > CIBAUD and IBSHIFT come from standards documents so it appears to make > sense at some level to support those standards. As everyone else in the > Power world supports it I don't see why Linux should be different. Sure, we can add them. I suggest IBSHIFT = 16 and CIBAUD = 0xff0000 (or 077600000 if we are going to continue the peculiar convention of using octal for the C* constants). Paul. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/