Solar Designer
Solar Designer
Are you sure? Here's someone else reporting this same issue on Arch Linux and then confirming it's gone in our latest code: https://github.com/openwall/john/issues/3572#issuecomment-2017858799 Can you please include copy-paste from your...
Thank you! Your screenshot does show that the kernel paths have changed, so is for a newer version, yet the same error occurs for another OpenCL kernel (the very first...
Also, can you please try this (when in the `run` directory of your manual build from this repo): ``` sed -ri 's/^( *)inline /\1INLINE /' `grep -rwl inline opencl` ./john...
> Also, can you please try this (when in the `run` directory of your manual build from this repo): For these changes to fully work, also need to add the...
I've implemented the potential fix above in #5780.
Thank you very much for running these tests @fruitbang! So it looks like the proposed fix is working to get past the kernel build issues. As to runtime failures, unfortunately...
> default branch or [#5780](https://github.com/openwall/john/pull/5780) ? At this time, we're more interested in testing of #5780. Thanks!
> Benchmarking: LM-opencl [DES BS OpenCL/mask accel]... 0: OpenCL CL_INVALID_WORK_GROUP_SIZE (-54) error in opencl_lm_b_plug.c:1306 - Failed enqueue kernel lm_bs_*. > LWS=128 GWS=131072 x17940 DONE > Raw: 27395M c/s real, 1128G...
Thank you very much! Only `3 out of 100 tests have FAILED` plus 1 format crashing (or causing the next to crash) is an unusually good result for an AMD...
As to which specific formats fail, only `pgpdisk-opencl` overlaps with what we see in #5709.