On Tue, 1 Jan 2019, I wrote:
There are no [nvram] ioctls common to all architectures. So your example
becomes,
static long nvram_misc_ioctl(struct file *file, unsigned int cmd,
unsigned long arg)
{
if (ops->ioctl)
return ops->ioctl(file, cmd, arg);
return -ENOTTY;
}
And then my objection is the same: m68k and x86 now have to duplicate
their common ops->ioctl() implementation.
Perhaps code duplication is inevitable. Either you punt the ioctl
implementation from the char misc driver (and duplicate that ioctl
implementation under arch/) or else you duplicate the char misc driver.
Maybe this dilemma explains the situation we have now* which is duplicated
drivers (drivers/char/nvram.c and drivers/char/generic_nvram.c). But this
explanation doesn't seem to offer any solution. Re-using either of the
existing drivers seems to be impossible.
Different interpretations of the NVRAM Kconfig symbol accross the tree are
not helping. And having separate Kconfig symbols (NVRAM and GENERIC_NVRAM)
for the two drivers doesn't help either. But maybe the NVRAM symbol can be
dropped from arch/powerpc and all of the powerpc drivers...
* Actually, we presently have duplicated misc device drivers AND
duplicated ioctl implementations too, for good measure. See
arch/powerpc/kernel/nvram_64.c and drivers/char/generic_nvram.c.
--
