if you just want it to work (no permant solution)
in asm/ioctl.h: lines 64, 65 and 66:
change the last argument of _IOC from "_IOC_TYPECHECK(size)" to
"sizeof(size)"
Changing kernel headers somehow strikes me as not the correct way. Does it even matter what the number values of these macros even are though as long as they dont step on any other's toes?
I'm gonna test things without sizeof in the macros.
and also the minor to MINOR changes (i used iminor though)
I thought iminor gave you an inode number. Not exactly the same as a device minor number is it? I know there is those two, i'd suspect it's for a reason.
I havent had any problems here.. using gcc3
good luck
Hallvar Helleseth
On Wed, 2003-10-15 at 06:40, Ed Sweetman wrote:
I'm using the cvs patch for 2.6.0-test7. I'm getting some kind of macro problems though. It keeps complaining about errors like this
drivers/char/fusion/fusiondev.c:368: error: syntax error before '[' token
drivers/char/fusion/fusiondev.c:372: error: syntax error before '[' token
drivers/char/fusion/fusiondev.c:387: error: syntax error before '[' token
drivers/char/fusion/fusiondev.c:400: error: syntax error before '[' token
drivers/char/fusion/fusiondev.c:411: error: syntax error before '[' token
the source is like this case FUSION_GET_ID:
the macro is in fusion.h
#define FUSION_GET_ID _IOR('F', 0x00, sizeof(int))
_IOR is defined in asm/ioctl.h
case's require the number to be a static int and even if they're all converted to if statements, I still cant get it to get rid of that [ error nor can i figure out where the [ comes from. _IOR returns a number, not a string last i checked and there is no [ even around the code.
Has anyone been able to compile fusion against 2.6.0-test7 and if so was there any hacking that was necessary
My config has devfs enabled, if that matters, and i doubt it would.
By the way, i had to change all the minor(dev) function calls to MINOR(dev).
--
Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe directfb-users" as subject.
