Bug 73583

Summary:	[PNV/ILK]igt/gem_double_irq_loop NSPT
Product:	DRI	Reporter:	lu hua <huax.lu>
Component:	DRM/Intel	Assignee:	Daniel Vetter <daniel>
Status:	CLOSED FIXED	QA Contact:	Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity:	major
Priority:	high	CC:	intel-gfx-bugs
Version:	unspecified
Hardware:	All
OS:	Linux (All)
Whiteboard:
i915 platform:		i915 features:

Description lu hua 2014-01-14 02:44:27 UTC

System Environment:
--------------------------
Platform:  Pineview/Ironlake/Sandybridge
Kernel:    (drm-intel-fixes)09f2344d0896eb458a639b922224c7a202c11599

Bug detailed description:
---------------------------
It's NSPT on Pineview/Ironlake/Sandybridge with -fixes, -queued and -nightly kernel.

Bisect shows: cb9e8376fac2c96c2e6bffa511b3087d916286d0 is the first bad commit
commit cb9e8376fac2c96c2e6bffa511b3087d916286d0
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Fri Jan 10 09:21:15 2014 +0100

    tests: fixup igt_simple_main fallout

    Oops, I should look at compilier warnings a bit better.

    Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

output:
IGT-Version: 1.5-gcb9e837 (x86_64) (Linux: 3.13.0-rc3_drm-intel-fixes_a885b3_20131220+ x86_64)
Test requirement not met in function __real_main94, file gem_double_irq_loop.c:104:
Last errno: 2, No such file or directory
Test requirement: (!((((devid) == 0x0102 || (devid) == 0x0112 || (devid) == 0x0122 || (devid) == 0x0106 || (devid) == 0x0116 || (devid) == 0x0126 || (devid) == 0x010A) || (((devid) == 0x0152 || (devid) == 0x0162 || (devid) == 0x0156 || (devid) == 0x0166 || (devid) == 0x015a || (devid) == 0x016a) || (((devid) == 0x0402 || (devid) == 0x0406 || (devid) == 0x040A || (devid) == 0x040B || (devid) == 0x040E || (devid) == 0x0C02 || (devid) == 0x0C06 || (devid) == 0x0C0A || (devid) == 0x0C0B || (devid) == 0x0C0E || (devid) == 0x0A02 || (devid) == 0x0A06 || (devid) == 0x0A0A || (devid) == 0x0A0B || (devid) == 0x0A0E || (devid) == 0x0D02 || (devid) == 0x0D06 || (devid) == 0x0D0A || (devid) == 0x0D0B || (devid) == 0x0D0E) || ((devid) == 0x0412 || (devid) == 0x0416 || (devid) == 0x041A || (devid) == 0x041B || (devid) == 0x041E || (devid) == 0x0C12 || (devid) == 0x0C16 || (devid) == 0x0C1A || (devid) == 0x0C1B || (devid) == 0x0C1E || (devid) == 0x0A12 || (devid) == 0x0A16 || (devid) == 0x0A1A || (devid) == 0x0A1B || (devid) == 0x0A1E || (devid) == 0x0D12 || (devid) == 0x0D16 || (devid) == 0x0D1A || (devid) == 0x0D1B || (devid) == 0x0D1E) || ((devid) == 0x0422 || (devid) == 0x0426 || (devid) == 0x042A || (devid) == 0x042B || (devid) == 0x042E || (devid) == 0x0C22 || (devid) == 0x0C26 || (devid) == 0x0C2A || (devid) == 0x0C2B || (devid) == 0x0C2E || (devid) == 0x0A22 || (devid) == 0x0A26 || (devid) == 0x0A2A || (devid) == 0x0A2B || (devid) == 0x0A2E || (devid) == 0x0D22 || (devid) == 0x0D26 || (devid) == 0x0D2A || (devid) == 0x0D2B || (devid) == 0x0D2E)) || ((devid) == 0x0f30 || (devid) == 0x0f31 || (devid) == 0x0f32 || (devid) == 0x0f33)) || ((((devid) & 0xff00) != 0x1600) ? 0 : ((((devid) & 0x00f0) >> 4) > 3) ? 0 : (((devid) & 0x000f) == 0x2) ? 1 : (((devid) & 0x000f) == 0x6) ? 1 : (((devid) & 0x000f) == 0xb) ? 1 : (((devid) & 0x000f) == 0xa) ? 1 : (((devid) & 0x000f) == 0xd) ? 1 : (((devid) & 0x000f) == 0xe) ? 1 : 0))))
not (yet) implemented for pre-snb

Reproduce steps:
-------------------------
1. ./gem_double_irq_loop

Comment 1 Daniel Vetter 2014-01-14 10:12:12 UTC

I've just run this on my snb, and it seems to work.

The test itself is supposed to get skipped on older platforms like ilk/pnv, which is the bug I've fixed in the patch. Can you please double check that on gen6 and later it works and earlier it skips?

Comment 2 lu hua 2014-01-16 01:55:15 UTC

(In reply to comment #1)
> I've just run this on my snb, and it seems to work.
> 
> The test itself is supposed to get skipped on older platforms like ilk/pnv,
> which is the bug I've fixed in the patch. Can you please double check that
> on gen6 and later it works and earlier it skips?

Retest on Sandybridge, It works well.

About PNV/ILK, Test on earlier commit(7775fca2df81), the return code is 0, result shows pass. I guess the test was incorrect, commit(cb9e8376fa) fixed this issue, Is right? So skip on latest commit, It is expected.  

output on commit(7775fca2df81):
IGT-Version: 1.5-g7775fca (x86_64) (Linux: 3.13.0-rc7_drm-intel-nightly_07db3e_20140115+ x86_64)
not (yet) implemented for pre-snb

Comment 3 Daniel Vetter 2014-01-17 06:35:53 UTC

(In reply to comment #2)
> (In reply to comment #1)
> > I've just run this on my snb, and it seems to work.
> > 
> > The test itself is supposed to get skipped on older platforms like ilk/pnv,
> > which is the bug I've fixed in the patch. Can you please double check that
> > on gen6 and later it works and earlier it skips?
> 
> Retest on Sandybridge, It works well.
> 
> About PNV/ILK, Test on earlier commit(7775fca2df81), the return code is 0,
> result shows pass. I guess the test was incorrect, commit(cb9e8376fa) fixed
> this issue, Is right? So skip on latest commit, It is expected.  
> 
> output on commit(7775fca2df81):
> IGT-Version: 1.5-g7775fca (x86_64) (Linux:
> 3.13.0-rc7_drm-intel-nightly_07db3e_20140115+ x86_64)
> not (yet) implemented for pre-snb

Yes exactly, my refactoring in dda85fb108469e3d3c8ce19751f928c70e9fba4c broke the testcase (by switching from SKIP to SUCCESS) and this commit fixed it again.

Since you've confirmed that snb still works all is working well, so we can close this.

Comment 4 lu hua 2014-01-17 08:08:34 UTC

Verified.Fixed.

Comment 5 Elizabeth 2017-10-06 14:40:36 UTC

Closing old verified.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.