So. The deadlock seems to be a leak on lock_non_0_ref. Every other caller is simply stacked behind RWLOCK on thq.lock_write, which is held by the thread in send -> msg_alloc.

Apr 17 2015, 6:29 AM · enlightenment-git

Apr 13 2015

c assigned T2287: eina_thread_queue deadlocks on FreeBSD to raster.

Apr 13 2015, 6:04 PM · enlightenment-git

c added a comment to T2287: eina_thread_queue deadlocks on FreeBSD.

Thread 3 (Thread 809c06c00 (LWP 101243)):
#0  0x000000080216bc69 in pthread_spin_trylock () from /lib/libthr.so.3
#1  0x00000008018dbd6d in eina_spinlock_take (spinlock=<optimized out>) at ../src/lib/eina/eina_inline_lock_posix.x:683
#2  eina_thread_queue_send (thq=<optimized out>, size=<optimized out>, allocref=0x7fffdfffdf58) at lib/eina/eina_thread_queue.c:385
#3  0x00000000004084e4 in th31_do (data=<optimized out>, th=<optimized out>)
    at tests/ecore/ecore_test_ecore_thread_eina_thread_queue.c:268
#4  0x0000000801d23a0b in _ecore_direct_worker (work=0x809ce4020) at lib/ecore/ecore_thread.c:424
#5  0x00000008018cdee2 in _eina_internal_call (context=0x809c3cb80) at lib/eina/eina_thread.c:109
#6  0x000000080216c6f4 in ?? () from /lib/libthr.so.3
#7  0x0000000000000000 in ?? ()
Backtrace stopped: Cannot access memory at address 0x7fffdfffe000

Apr 13 2015, 2:50 PM · enlightenment-git

c created T2287: eina_thread_queue deadlocks on FreeBSD.

Apr 13 2015, 2:21 PM · enlightenment-git

c added a comment to D2335: tempget: Add more temperature sources for FreeBSD.

In D2335#12, @q66 wrote:

dude, please set your git config right... because of the lack of your correct git config it put me as the author of the commit. I forgot to check and now it's pushed. Fix it for your next revision

Apr 13 2015, 3:25 AM · enlightenment-git

Apr 12 2015

c added a comment to D2314: Makefile_Evas.am: Invoke bash for gen_shaders*.sh.

Thanks!

Apr 12 2015, 4:22 PM · efl

c retitled D2335: tempget: Add more temperature sources for FreeBSD from to tempget: Add more temperature sources for FreeBSD.

Apr 12 2015, 9:28 AM · enlightenment-git

Apr 9 2015

c retitled D2326: epulse: e_mod_main:_gc_orient: Don't use uninitialized w, h from to epulse: e_mod_main:_gc_orient: Don't use uninitialized w, h.

Apr 9 2015, 7:36 AM · Restricted Project

c edited this Maniphest Task.

Apr 9 2015, 2:39 AM · enlightenment-git

Apr 8 2015

c updated the diff for D2314: Makefile_Evas.am: Invoke bash for gen_shaders*.sh.

Rebase against master, add new 3d SHADERS as well.

Apr 8 2015, 6:55 PM · efl

c planned changes to D2314: Makefile_Evas.am: Invoke bash for gen_shaders*.sh.

This is incomplete. I will update for master (adds shaders_3d) and invoke with bash specifically (uses non-sh features and shebang line specifies bash).

Apr 8 2015, 6:44 PM · efl

c retitled D2315: elm_spinner.c: Use correct fabs() for floating point spin_speed from to elm_spinner.c: Use correct fabs() for floating point spin_speed.

Apr 8 2015, 4:30 PM · Restricted Project

c retitled D2314: Makefile_Evas.am: Invoke bash for gen_shaders*.sh from to Makefile_Evas.am: Invoke sh for gen_shaders (fixes src-on-NFS builds on FreeBSD).

Apr 8 2015, 4:04 PM · efl

c retitled D2313: ecore_audio/pulse: Include libgen.h on FreeBSD to define basename(3) from to ecore_audio/pulse: Include libgen.h on FreeBSD to define basename(3).

Apr 8 2015, 3:52 PM · efl

c added a comment to D2302: ecore_main.c: Refactor purely duplicate ecore_main_fd_handler_file_add().

In D2302#8, @jpeg wrote:

Instead, we dislike the lack of authorship. Your name gets lost. And this actually adds burden to us reviewers as we need to manually amend the commit to add your name. Otherwise git log will show patches that were authored by someone else under our name (not a huge deal, but it actually matters to some of us). Please configure git in your EFL/Elm/E repos:

Apr 8 2015, 5:48 AM · efl

Apr 7 2015

c updated the diff for D2302: ecore_main.c: Refactor purely duplicate ecore_main_fd_handler_file_add().

Missed the header definition — did not expect a function with underscore-prefix to be exposed outside of the CU. Fixed.

Apr 7 2015, 7:39 PM · efl

c added a reviewer for D2302: ecore_main.c: Refactor purely duplicate ecore_main_fd_handler_file_add(): efl.

Apr 7 2015, 6:14 PM · efl

c retitled D2302: ecore_main.c: Refactor purely duplicate ecore_main_fd_handler_file_add() from to ecore_main.c: Refactor purely duplicate ecore_main_fd_handler_file_add().

Apr 7 2015, 6:12 PM · efl

c abandoned D2301: T2274: Fix trivial use-after-free in mixer/pa.c.

In D2301#10, @zmike wrote:

I appreciate your submission, and I would be grateful it if you could change your local configuration to add in the authorship info prior to updating this patch with it.

Apr 7 2015, 3:51 PM · enlightenment-git

c added a comment to D2301: T2274: Fix trivial use-after-free in mixer/pa.c.

Are you kidding me? It's completely trivial and you can credit yourself if you like. This patch is released into the public domain or MIT-licensed if you live somewhere that the public domain doesn't exist.

Apr 7 2015, 2:37 PM · enlightenment-git

c edited this Maniphest Task.

Apr 7 2015, 2:21 PM · enlightenment-git

c retitled D2301: T2274: Fix trivial use-after-free in mixer/pa.c from to T2274: Fix trivial use-after-free in mixer/pa.c.

Apr 7 2015, 2:20 PM · enlightenment-git

c created T2274: Trivial use-after-free in e17 modules/mixer/pa.c ca. line 744.

Apr 7 2015, 2:01 PM · enlightenment-git