[Intel-gfx] [RFC] drm/i915: context support unit test-r2

[Ben Widawsky]

I've been sitting on this for over a week now. I figure I may as well
post this since I've been working on the implementation in libdrm and
the driver for the last week, although because of a bunch of
bureaucratic distractions this week, it amounted to only a day's worth
of work :P.

The changes basically implemented the slot mechanism that Daniel and
Kristian recommended. I'll admit that even this basic unit test has
changed a few times in the last few days, but it's pretty simple, and
shows the basic idea.

Most of the implementation details of the ctx related functions are
missing, but I think the names are descriptive enough. Let me know if
more details are needed, and I can add them.

I'd like to keep the implementation details such as, how I'll be pinning
the HW context memory, out of the API discussion - which is the
direction that I'd like this thread to move towards.

As a final note, the test copied the bad_blit test, which was the
easiest for me to reuse. I think blit is a really poor example for
context support, but it's also much easier for non-graphics people like
myself to understand.

Ben
-------------- next part --------------
A non-text attachment was scrubbed...
Name: gem_ctx_basic.c
Type: text/x-c
Size: 4458 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/intel-gfx/attachments/20110114/99e436f8/attachment.bin>