Wolfgang Grandegger wrote:
> Oliver Hartkopp wrote:
>> Oliver Hartkopp wrote:
>>> Hi Wolfgang,
>>>
>>> i created a patch for the Kconfig problem that has come up with the EMS USB
>>> config entry in 2.6.32-rcX.
>>>
>>> Can you check, if it's ok for you?
>>>
>> Just a small update using 'if' and 'endif' in sja1000/Kconfig to omit the
>> permanent "depends on CAN_SJA1000" stuff.
>
> I'm not a friend of using that type if if's in Kconfig.
But it's the usual way - i doublechecked that right now.
The only problem could be that the dependencies are missing, when the driver
is moved into another Kconfig environment. But this is not a real problem.
Another approach would be writing the config entries like this:
config CAN_OLD_DRIVERS
tristate "Prompt for old CAN drivers (e.g. no sysfs support)"
depends on CAN
default N
---help---
The old drivers do not support sysfs nor proper platform device
support. Some of the old drivers might only be configured by
module commandline options.
if CAN_OLD_DRIVERS
source "drivers/net/can/old/Kconfig"
endif
Would you prefer that?
> Apart from that
> the patch looks good. Just one minor nitpicking. I would write:
>
> +menuconfig CAN_SJA1000
> + depends on CAN_DEV && HAS_IOMEM
> + tristate "NXP/Philips SJA1000 CAN controller drivers"
Which is misleading.
We provide just *one* SJA1000 driver but several SJA1000 based devices.
I tend to change
"Philips/NXP SJA1000 CAN controller based devices"
to
"Philips/NXP SJA1000 based devices"
to reduce the bla bla :-)
>
> Thanks for overtaking.
I think fixing this up is important for 2.6.32 where the problem came in with
the USB driver. The later we fix this the bigger the mess ...
And AFAIK you need to cook a patch for the xstats for 2.6.32 anyway - and
you're already on heavy load.
Tnx,
Oliver
_______________________________________________
Socketcan-core mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/socketcan-core
