[Intel-gfx] [PATCH 0/5] Add power feature debugfs disabling

[Jeff McGee]

On Thu, Feb 06, 2014 at 05:37:29PM +0100, Daniel Vetter wrote:
> On Thu, Feb 6, 2014 at 4:44 PM, Jeff McGee <jeff.mcgee at intel.com> wrote:
> > Our Android system validation tests are expecting these interfaces. That's
> > not igt, I know, but is supporting downstream test suites a priority? I can
> > get our val guys on the list to +1 the need for these patches. Likewise I
> > can request a developer from my team to review these patches. Or are you
> > looking specifically for someone outside our downstream product to 2nd the
> > need-for and quality of the patches?
> 
> I don't mind if something is only used in one product - if it's useful
> sooner or later other product teams will pick up on in. So upstreaming
> is the right thing. But if you add a debugfs for tests I also want to
> have the tests upstreamed for the following reasons:
> 
> - If a product team deems it useful to test something, our upstream QA
> should probably do the same.
> 
> - Without testcases actually using these on upstream there's a good
> chance that we'll break them. Which means forward-rolling to new
> forklifts will be more of a pain than strictly needed. We already have
> our issues with upstream collaboration, no need to make it more ugly.
> 
> - For interfaces used in tests/scripts I also want to do a bit of api
> review, which is best done by looking at the actual users. Ofc debugfs
> doesn't have strict abi guarantees imposed by the linux community like
> ioctls/sysfs which we're essentially never allowed to break. But
> change the interfaces still has its costs.
> 
> Cheers, Daniel

Thanks for the clarification. I think there is a willingness from our
Android teams, both dev and val, to invest in igt as we've just started to
do. I don't know the history well, but we obvious have a lot invested in
our own test suites, and so bridging those with igt will take time and
may occur on the back burner while the focus has to be on Android product.
-Jeff