Zhibin Li's GSoC status report - #12 of 12

Zhibin Li haoyouab at gmail.com
Mon Aug 19 16:56:23 UTC 2019

Hi all,

- Finished reworking of general DRM ioctl tests. In the tests I detect
config for libdrm or kernel headers like what I did in the decoders before
while the decoders now are using bundled headers in /strace/linux/drm/. I
believe this is the supposed way to handle the decoders and tests.
One problem that I found in detecting libdrm or kernel headers is that on
travis-ci env (namely ubuntu14.04) the form of checking members like:

AC_CHECK_MEMBERS([struct drm_mode_fb_cmd2.modifier],,, [#include <drm.h>])

is not working but it works on newer distribution like 16.04 or fedora.
I checked config.log and then found:

configure:16260: checking for struct drm_mode_fb_cmd2.modifier
configure:16260: gcc -std=gnu99 -c -g -O2  -I/usr/include/libdrm
conftest.c >&5
In file included from conftest.c:345:0:
/usr/include/libdrm/drm.h:132:2: error: unknown type name 'size_t'
  size_t name_len;   /**< Length of name buffer */
--- snippet ---

configure:16260: $? = 1

and this is exactly why HAVE_STRUCT_DRM_MODE_FB_CMD2_MODIFIER is #undef
although modifier field is present. The part I don't understand is that the
version of autoconf is 2.69 on all those OSes. Is it related to something
else? Anyway, I added AC_INCLUDES_DEFAULT to the include part and the
problem is resolved.

BTW, in order to simplify the tests, I just printed fix-number of elements
in those array printing functions. Would that be a problem? If so, I can
try to implement a tests-wide common print_array and I'm just not sure if
that will bring too much burden.

- Fixed a few issues found in the patch I sent last week while running the
tests. I will re-send it along with the tests.

- Created a blog page[1] for holding notes of GSoC 2019. It's a draft now
and will be updated according to you guys' feedbacks.

[1] https://haoyouab.github.io/2019/05/16/GSoC-2019.html

Priorities are outside of the scope of GSoC now but I will try to summarize
anyway considering DRM ioctl decoding still needs a lot of work.
- Continue to work on the patch I sent until they are good enough to be

Zhibin Li
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.strace.io/pipermail/strace-devel/attachments/20190820/65ccb687/attachment.html>

More information about the Strace-devel mailing list