add generic stub for unsupported ops on ioman(X)

[israpps]

• improves readability (IMO)
• unified return value for unsupported ops (-ENOTSUP) making it clear from EE side that the operation that just failed is simply stubbed on the IOP

[fjtrujy]

instead of using macros you can also use:

static inline int not_supported_op (void) {return -134;}

[israpps]

You're right, I'll apply the change later.

Although, I would like to read a review of @uyjulian on this regard, specially with the DVR related IRX modules

[uyjulian]

Suggest to use ENOSYS instead of making up a new error code 134

[israpps]

Suggest to use ENOSYS instead of making up a new error code 134

134 corresponds to -ENOTSUP without including errno everywhere.

[fjtrujy]

Suggest to use ENOSYS instead of making up a new error code 134

134 corresponds to -ENOTSUP without including errno everywhere.

But I think this is where the issue is coming from, to include errno in every file where it is needed it is a good practice. In this way you don't care about the "actual value" however you care about the "meaning" ENOSYS.

[israpps]

Suggest to use ENOSYS instead of making up a new error code 134

134 corresponds to -ENOTSUP without including errno everywhere.

But I think this is where the issue is coming from, to include errno in every file where it is needed it is a good practice. In this way you don't care about the "actual value" however you care about the "meaning" ENOSYS.

I thought including errno also includes the array of descriptions that comes with it. Or that is stripped away by the compiler? If that's the case I'll correct it

[fjtrujy]

Suggest to use ENOSYS instead of making up a new error code 134

134 corresponds to -ENOTSUP without including errno everywhere.

But I think this is where the issue is coming from, to include errno in every file where it is needed it is a good practice. In this way you don't care about the "actual value" however you care about the "meaning" ENOSYS.

I thought including errno also includes the array of descriptions that comes with it. Or that is stripped away by the compiler? If that's the case I'll correct it

It shouldn’t increase binary size, it must be exactly the same, if not something else is wrong

[fjtrujy]

Why have you changed the value here?

[israpps]

-ENOTSUP. but as s64

[fjtrujy]

I’m ok merging this change @rickgaiser @uyjulian .
However this improvement is quite difficult to check if other devs will maintain.
Just to put an example, uyjulian has created 2 extra PRs where he adds more IOP modules, he is not using that function there…

[israpps]

Just to put an example, uyjulian has created 2 extra PRs where he adds more IOP modules, he is not using that function there…

Well of course he's not using something that has not been merged yet.😬

If this gets merged let's just propagate the change to any open PR.

[uyjulian]

Including errno.h won't increase binary size as it is just preprocessor macro definitions. Basically the equivalent of copy and paste.

[israpps]

Including errno.h won't increase binary size as it is just preprocessor macro definitions. Basically the equivalent of copy and paste.

excellent. I thought it did. at least my local copy of errno.h (did not check the one on ps2sdk) had a char** buffer for the descriptions of errno values

[fjtrujy]

Resolve first conflicts and then I will review it

[israpps]

done, should b ok now

[fjtrujy]

It should work nice! And I don't expect any memory increase at all

[israpps]

It should work nice! And I don't expect any memory increase at all

I hope so lol