https://bugs.kde.org/show_bug.cgi?id=386923
--- Comment #6 from Giuseppe Bilotta <giuseppe.bilo...@gmail.com> --- Thanks for the extensive tracking, Milian, I'll find sometime later today to try the latest git master of heaptrack. FWIW, I do have the Intel platform available, but I can reproduce the deadlock by limiting the available platforms to just the Mesa one (using ocl-icd for libOpenCL and setting OCL_ICD_VENDORS so that only Mesa is found), so I don't thin Intel's TBB would be at fault in my case. Maybe Mesa does its own alloc meddling? I'm using mesa 17.2.4-1+b1 packaged in Debian sid. -- You are receiving this mail because: You are watching all bug changes.
