On 2013-06-27 17:50, Ian Abbott wrote:
On 2013-06-27 17:08, Ian Abbott wrote:
The "ni_labpc" module acts as a low-level comedi driver for various
National Instruments Lab-PC ISA cards if `CONFIG_NI_LABPC_ISA` is
enabled, and also as a common module depended on by the "ni_labpc_cs"
and "ni_labpc_pci" modules. For ISA cards, it optionally supports the
use of an ISA DMA channel if `CONFIG_ISA_DMA_API` is enabled, using
conditionally compiled code. This conditionally compiled code also
relies on `virt_to_bus()`, which only exists if `CONFIG_VIRT_TO_BUS` is
enabled. Therefore, support for ISA DMA should only be compiled in if
both `CONFIG_ISA_DMA_API` and `CONFIG_VIRT_TO_BUS` are defined.
The `CONFIG_NI_LABPC_ISA` option currently depends on `VIRT_TO_BUS` at
configure time and checks `CONFIG_ISA_DMA_API` at compile time. It
seems better to check both options at compile time.
Conditionally define a new macro `NI_LABPC_ISA_DMA` iff both
`CONFIG_ISA_DMA_API` and `CONFIG_VIRT_TO_BUS` are defined, and change
the conditional compilation tests to use the new macro instead of
`CONFIG_ISA_DMA_API`.
Also, in the "Kconfig", remove the `VIRT_TO_BUS` dependency from the
`COMEDI_NI_LABPC_ISA` option since the dependency is now checked at
compile time.
On second thoughts, perhaps I should create a new module and header file
and a new conditionally selected config option to handle the ISA DMA
support in this module, as that would remove a bunch of #ifdefs (or at
least hide them in the header file).
Might as well dump this patch now. I've already split the ISA DMA
support out into a separate module locally. I just need to split my
patch up for submission.
--
-=( 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/driverdev-devel
