| Summary: | Crash when using attached template, if both OpenCL options are enabled | ||
|---|---|---|---|
| Product: | LibreOffice | Reporter: | Regina Henschel <rb.henschel> |
| Component: | Calc | Assignee: | Tomaz Vajngerl <quikee> |
| Status: | CLOSED NOTOURBUG | ||
| Severity: | major | CC: | michael.meeks, miguelangelrv, quikee, raal |
| Priority: | medium | Keywords: | bibisected, bisected, regression |
| Version: | 5.2.0.0.alpha0+ | ||
| Hardware: | x86 (IA32) | ||
| OS: | Windows (All) | ||
| Whiteboard: | |||
| Crash report or crash signature: | Regression By: | ||
| Bug Depends on: | |||
| Bug Blocks: | 97391 | ||
| Attachments: |
Crashing template
Backtrace as suggested |
||
|
Description
Regina Henschel
2016-05-28 12:09:58 UTC
I can't reproduce. Win10x64 Version: 5.2.0.0.alpha1+ Build ID: 7b704dfbdb23540ff6366fa60c73474bbda9dc26 CPU Threads: 4; OS Version: Windows 6.19; UI Render: default; TinderBox: Win-x86@39, Branch:master, Time: 2016-05-20_05:45:40 With OpenGL enable and disable. I'm on Windows 7, 32 bit. Goto Tools > Options > LibreOffice > OpenCL. Find the options (1) Allow use of Software Interpreter (even when OpenCL is not available) (2) Allow use of OpenCL. If both options are checked, then it crashes. If only one or no option is checked, it does not crash. The default is, that both options are checked. I can't reproduce in any case, with both, one or none. Maybe a not blacklisted OpenCL There are two log files on the cache folder with information about the devices. Opencl_devices.log Opencl_device.log For my PC the file content is:
opencl_devices.log
------------------
Device Index: 0
Selected: false
Device Name: AMD Phenom(tm) II X4 940 Processor
Device Vendor: AuthenticAMD
Device Version: OpenCL 1.2 AMD-APP (937.2)
Driver Version: 2.0 (sse2)
Device Type: cpu
Device Extensions: cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_gl_sharing cl_ext_device_fission cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_popcnt cl_khr_d3d10_sharing
Device OpenCL C Version: OpenCL C 1.2
Device Available: true
Device Compiler Available: true
Device Linker Available: false
Platform Name: AMD Accelerated Parallel Processing
Platform Vendor: Advanced Micro Devices, Inc.
Platform Version: OpenCL 1.2 AMD-APP (937.2)
Platform Profile: FULL_PROFILE
Platform Extensions: cl_khr_icd cl_amd_event_callback cl_amd_offline_devices cl_khr_d3d10_sharing
opengl_device.log
-----------------
DriverVersion: 8.970.100.7000
DriverDate: 11-16-2012
DeviceID: PCI\VEN_1002&DEV_9610&SUBSYS_75011462&REV_00
AdapterVendorID: 0x1002
AdapterDeviceID: 0x9610
AdapterSubsysID: 0x75011462
DeviceKey: System\CurrentControlSet\Control\Video\{331693BC-4FB7-4781-B597-9FB2E140569F}\0000
DeviceString: ATI Radeon HD 3200 Graphics
Maybe a backtrace using a TB39 build would be useful: https://wiki.documentfoundation.org/How_to_get_a_backtrace_with_WinDbg Created attachment 125357 [details]
Backtrace as suggested
Perhaps a try updating the graphics driver. http://support.amd.com/en-us/download/desktop/legacy?product=legacy2&os=Windows%20XP%20-%20Professional/Home&RenderOnServer=true not too much modern, but who knows. As there is a backtrace and this kink of bug are can't be reproduce without the same hardware, changed the status to new. This seems to have begun at the below commit. Adding Cc: to Tomaž ; Could you possibly take a look at this one? Thanks 97d366b4fa31a6aab550dbe001a893c4dd0e5bc3 is the first bad commit commit 97d366b4fa31a6aab550dbe001a893c4dd0e5bc3 Author: Norbert Thiebaud <nthiebaud@gmail.com> Date: Sat Apr 23 11:43:22 2016 -0700 source 334e2dc9c3da4519e31e0452a40d3a958c401876 author Tomaž Vajngerl <tomaz.vajngerl@collabora.co.uk> 2016-04-20 13:19:47 (GMT) committer Tomaž Vajngerl <quikee@gmail.com> 2016-04-22 11:27:05 (GMT) commit 334e2dc9c3da4519e31e0452a40d3a958c401876 (patch) tree bfe420e41e0e9c5e7767c4ce657bd5e29192d91f parent cb5438a3d16634eb267584122169758ab2b76930 (diff) opencl: better logging of devices and device selection Tomaz - I guess one to append to your queue. I have uninstalled the driver, without installing a new one. Now I have got a ATI Technologies Inc. driver version 8.632.1.2000 from 17.08.2009. That seems to be the default driver from Windows 7? With this driver it no longer crashes. So my previous driver seems to be a candidate for the blacklist. Does this still happen after [1]? Can anybody check? [1] https://cgit.freedesktop.org/libreoffice/core/commit/?id=40b0b9ab7703a165295b008f47df14d2ec076fb1 Do you have any OpenCL support at all with that driver you now are using? (If you remove the opencl_devices.log, does a new one appear? What does it contain?) When I remove all files in Cache, on restart a file "opengl_device.log" is generated. It contains:
DriverVersion: 8.632.1.2000
DriverDate: 8-17-2009
DeviceID: PCI\VEN_1002&DEV_9610&SUBSYS_75011462&REV_00
AdapterVendorID: 0x1002
AdapterDeviceID: 0x9610
AdapterSubsysID: 0x75011462
DeviceKey: System\CurrentControlSet\Control\Video\{331693BC-4FB7-4781-B597-9FB2E140569F}\0000
DeviceString: ATI Radeon HD 3200 Graphics
My current version of LO is Version: 5.3.0.0.alpha0+
Build ID: 3ab13873ebb6dc4738be2e2184ee4433a2447c1d
CPU Threads: 4; OS Version: Windows 6.1; UI Render: default;
TinderBox: Win-x86@39, Branch:master, Time: 2016-06-16_01:48:01
Locale: de-DE (de_DE)
Hi Regina, thanks for the feedback - we've added code now to test OpenCL by doing some real calculations on first-start and/or driver change - which should avoid this problem by crashing and disabling CL in the first instance. I guess we will collect stats from the crash-reporter on that over time as we add OCL information to it, but of course testing of that much appreciated vs. a recent build yourself. Otherwise - I guess closing this as NOTOURBUG is perhaps the best approach ? =) Thanks ! Hi Michael, I do not have installed the problematic driver any longer, so I cannot test it. Therefore I agree to close it. |