Re: [RFC] Don't use linux/version.h anymore to indicate a per-driver version - Was: Re: [PATCH 03/37] Remove unneeded version.h includes from include/

[Mauro Carvalho Chehab <mchehab@infradead.org>]

Em 24-06-2011 18:10, Stefan Richter escreveu:
> On Jun 24 Devin Heitmueller wrote:
>> Really, this is all about applications being able to jam a hack into
>> their code that translates to "don't call this ioctl() with some
>> particular argument if it's driver W less than version X, because the
>> driver had a bug that is likely to panic the guy's PC".  Sure, it's a
>> crummy solution, but at this point it's the best that we have got.
> 
> The second best.  The best that we have got is that the user runs a fixed
> kernel.
> 
> Anyway; if this is the only purpose that this interface version¹ serves,
> then Mauro's subsystem-centralized solution has the benefit that it
> eliminates mistakes due to oversight by individual driver authors.
> Especially because the kind of implementation behavior changes that are
> tracked by this type of version datum are sometimes just discovered or
> documented in hindsight.  On the other hand, Mauro's solution is redundant
> to the uname(2) syscall.

Yes. That's why my initial proposal were to add some value to it by not associating
it with the kernel version, but with a number that will be incremented only if
the V4L2 API changes.

> 
> ¹) Yes, it is still an ABI version, nothing less.  With all its backwards
> and forwards compatibility ramifications.