Package: intel-opencl-icd
Version: 20.37.17906-1
Severity: critical
When this package is installed, any OpenCL program will abort with the
message
Abort was called at 42 line in file:
/build/intel-compute-runtime-WsWnhf/intel-compute-runtime-20.37.17906/shared/source/built_ins/built_ins.cpp
Application Error
This is reproducible with a simple `clinfo -l`, but even something like
LibreOffice with OpenCL support enabled will fail to start (making it
impossible to actually even the LibreOffice configuration to disable
OpenCL), which is the reason for the critical severity.
Running `clinfo -l` under gdb shows the following backtrace:
#0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
#1 0x77da6537 in __GI_abort () at abort.c:79
#2 0x76baab53 in NEO::abortExecution () at
./shared/source/helpers/abort.cpp:14
#3 0x76bc0b71 in NEO::abortUnrecoverable (line=line@entry=42,
file=file@entry=0x76f1b868
"/build/intel-compute-runtime-WsWnhf/intel-compute-runtime-20.37.17906/shared/source/built_ins/built_ins.cpp")
at ./shared/source/helpers/debug_helpers.cpp:24
#4 0x76ed07b4 in operator() (__closure=0x7fffdd60) at
./shared/source/built_ins/built_ins.cpp:42
#5 std::__invoke_impl&> (__f=...) at /usr/include/c++/10/bits/invoke.h:60
#6 std::__invoke&> (__fn=...) at /usr/include/c++/10/bits/invoke.h:95
#7 operator() (this=) at /usr/include/c++/10/mutex:717
#8 operator() (this=0x0) at /usr/include/c++/10/mutex:722
#9 _FUN () at /usr/include/c++/10/mutex:722
#10 0x7710b34f in __pthread_once_slow (once_control=0x55786400,
init_routine=0x77495fc0 <__once_proxy>) at pthread_once.c:116
#11 0x76ed0ae2 in __gthread_once (__func=,
__once=0x55786400) at
/usr/include/x86_64-linux-gnu/c++/10/bits/gthr-default.h:700
#12 std::call_once&> (__f=..., __once=...) at
/usr/include/c++/10/mutex:729
#13 NEO::BuiltIns::getSipKernel (this=0x55785ff0, type=,
device=...) at ./shared/source/built_ins/built_ins.cpp:64
#14 0x76be2516 in NEO::initSipKernel (type=, device=...)
at ./opencl/source/helpers/built_ins_helper.cpp:16
#15 0x76c29878 in NEO::Platform::initialize (this=0x557841b0,
devices=std::vector of length 1, capacity 1 = {...}) at
./opencl/source/platform/platform.cpp:145
#16 0x76bdc85d in clGetPlatformIDs (numEntries=,
platforms=, numPlatforms=) at
./opencl/source/api/api.cpp:100
#17 0x76bdce54 in clIcdGetPlatformIDsKHR (numEntries=0, platforms=0x0,
numPlatforms=0x7fffe04c) at ./opencl/source/api/api.cpp:136
#18 0x77f51ae3 in _find_and_check_platforms (num_icds=5) at
ocl_icd_loader.c:445
#19 __initClIcd () at ocl_icd_loader.c:652
#20 _initClIcd_real () at ocl_icd_loader.c:702
#21 0x77f524e3 in _initClIcd () at ocl_icd_loader.c:724
#22 clGetPlatformIDs (num_entries=0, platforms=0x0,
num_platforms=0x7fffe170) at ocl_icd_loader.c:846
#23 0x55564d42 in main (argc=2, argv=0x7fffe2d8) at
src/clinfo.c:3351
Note that no OpenCL ICD should _ever_ invoke abort: OpenCL has specific
ways to pass failure information up to the caller, those shold be used
instead.
(That aside, it's unclear why the failure even happens in the first
place.)
-- System Information:
Debian Release: bullseye/sid
APT prefers unstable-debug
APT policy: (500, 'unstable-debug'), (500, 'testing-debug'), (500,
'oldoldstable'), (500, 'unstable'), (500, 'testing'), (500, 'stable'), (1,
'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386
Kernel: Linux 5.8.0-2-amd64 (SMP w/12 CPU threads)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE,
TAINT_UNSIGNED_MODULE
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
Versions of packages intel-opencl-icd depends on:
ii libc6 2.31-4
ii libgcc-s1 10.2.0-17
ii libigc1 1.0.5353.1-2
ii libigdfcl1 1.0.5353.1-2
ii libigdgmm11 20.3.2+ds1-1
ii libstdc++6 10.2.0-17
ii ocl-icd-libopencl1 2.2.12-4
intel-opencl-icd recommends no packages.
intel-opencl-icd suggests no packages.
-- no debconf information
