Re: [RFC c/r 2/4] [RFC] syscalls, x86: Add __NR_kcmp syscall v7

[ebiederm@xmission.com (Eric W. Biederman)]

Cyrill Gorcunov <gorcunov@openvz.org> writes:

> On Fri, Jan 27, 2012 at 12:33:07PM -0800, Eric W. Biederman wrote:
>> KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> writes:
>> >> +			ret = kcmp_ptr((long)filp1, (long)filp2, KCMP_FILE);
>> >> +		else
>> >> +			ret = -ENOENT;
>> >
>> > If my remember is correct, Andrew pointed out EINVAL is better than ENOENT.
>> 
>> Ah yes.  And really what it should be is
>> 		if (!filp1 || !filp2)
>>  			return -EBADF;
>> 
>> At least EBADF is what you return if it is your process that doesn't
>> have the filedescriptor.
>> 
>
> Eric, I've sent out version with
>
> 		if (filp1 && filp2)
> 			...
> 		else
> 			ret = -EBADF;
>
> maybe you're lookin into previous version?

Yeah.  Comments and patch passing in the night.  It looks like you have
it right in your latest patch.

>> >> +			       KCMP_SYSVSEM);
>> >> +#else
>> >> +		ret = -EINVAL;
>> >
>> > ENOTSUP is better, I think. because of, EINVAL implicitly mean _caller_ is wrong.
>> > but in this case, it is not bad. only the kernel doesn't have enough
>> > feature.
>> 
>> Careful a type compiled out should in principle match a type whose
>> support has not been implemented. That is the default case should match
>> what happens when you don't compile in sysvipc support.
>
> I don't get it :) Will -EINVAL be enough or not?

At the present time the only way we can get -EINVAL by not supporting a
type so there is no pressing need for something different.

However in the general case EINVAL is a pretty generic failure mode and
having something more precise that you can use to figure out what you
did wrong when calling a system call tends to help a great deal.

So I am favor of using a better error code if EOPNOTSUP or ENOTTY if we
can convince ourselves it is the proper error code.

Mostly it is bike shedding but it is a detail that getting it right will
help users of this interface in the long run.

Eric