On 2012-08-15 11:48, Dan Carpenter wrote:
On Wed, Aug 15, 2012 at 11:16:20AM +0100, Ian Abbott wrote:
On 2012-08-15 08:14, Dan Carpenter wrote:
On Tue, Aug 14, 2012 at 04:31:28PM +0100, Ian Abbott wrote:
+#define IS_ISA_BOARD(board) (DO_ISA && (board)->bustype == isa_bustype)
+#define IS_PCI_BOARD(board) (DO_PCI && (board)->bustype == pci_bustype)
It would be better to make this an inline function.
#if IS_ENABLED(CONFIG_COMEDI_AMPLC_DIO200_ISA)
static inline bool is_isa_board(const struct dio200_board *board)
{
return board->bustype == isa_bustype;
}
#else
static inline bool is_isa_board(const struct dio200_board *board)
{
return 0;
}
#endif
I prefer the macro as it seems "easier" for the compiler to
recognize it as a constant in the case where DO_ISA expands to 0.
If we were just discussing this one function then, sure, we're not
going to complain about small style issues, but it's the whole file.
It's just riddled with #ifdefs. It can't be moved out of staging
until it is written according the same coding style as the rest of
the kernel.
The remaining #ifdefs are only used to fill in the list of supported
boards, the static "layout" data referenced by the supported boards (the
#ifdefs could be removed from the layout array with a moderate expansion
of module size), and to optionally make it a PCI driver. Currently, the
driver does not link any PCI functions if it is not configured to
support PCI boards, even if CONFIG_PCI is enabled (as long as the
compiler can optimize out the code that calls those functions).
I don't really want to hide the MODULE_DEVICE_TABLE(),
module_comedi_pci_driver() and module_comedi_driver() macro calls in
some other macros buried in some private header file, and I also want to
avoid compiler warnings about unused static functions and variables.
--
-=( Ian Abbott @ MEV Ltd. E-mail: <abbotti@xxxxxxxxx> )=-
-=( Tel: +44 (0)161 477 1898 FAX: +44 (0)161 718 3587 )=-
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
