Hashcat problems on Raspberry Pi 3 with VC4CL #81
Assignees
Labels
Comments
|
From the error location, it looks like the precompiled VC4CL-StdLib files were not found. Can you list the contents of:
|
|
In /usr/local/include/vc4cl-stdlib/ I have However I am missing the directories /usr/include/vc4cl-stdlib/, |
|
No you don't need to. But you should have these two files also in
Can you rerun the |
|
For vc4cl-stdlib I get For vc4c I get |
|
I also get errors when trying to compile the example. Is there a way to patch in the missing files? |
|
Yes, this command generates the precompiled headers required: Example (when running from VC4C project directory: This is also the command that should have been executed automatically. I will need to have a look as to why it was not run. |
|
When I run that command I get the error |
|
Does the relative path |
|
Yes, the file |
|
Does the user ( |
doe300
added a commit
to doe300/VC4C
that referenced
this issue
Aug 4, 2019
Also adds pecompilation of stdlib headers for make install, see doe300/VC4CL#81
|
Thank you, I reinstalled VC4C and now the files VC4CLStdLib.h.pch and VC4CLStdLib.bc are |
|
I managed to clear some errors by increasing the memory of the gpu, however I still get the error |
This error is due to some feature not yet implemented in VC4C. The other error I cannot really say without an error message, although it looks like it originates in the clang-frontend. Can you run the code in gdb, break on exceptions via |
|
Sorry, I'm quite new to this. Which code do I run in gdb and how do I find the errors? |
|
The command that you use to run hashcat. Assuming you run hashcat with the program Edit: Example with program the interactive shell will open... after some time, the exception will be thrown |
|
What happens to me is interesting. When running hashcat, I get my second error from before. When I run it in gdb, I get the following
This resets when I reboot however. |
doe300
added a commit
to doe300/VC4C
that referenced
this issue
Sep 14, 2019
Also adds helper function to insert while-loop into generated code. See doe300/VC4CL#81 Fixes: TestVC4C/emulator/test_async_copy
|
Thanks! Hashcat now runs, but during runtime I get this error |
|
I also needed to increase my swap file to 1 gb for hashcat to compile the opencl runtime binaries |
|
Hi there, is there any update on this? trying to run hashcat on my RPI but without success (using latest nightly VC4CL) and wondering whether is a user error or still lacking Hashcat support. Kind regards, |
|
The above error (
@ling1729, what compilation step exactly requires that much memory? |
|
When running hashcat, it gets to here It seems that it slowly uses more and more memory, and after letting it run for an hour or so it gets to the above error. When there isn't enough memory in the swap the cpu starts running at 100%, but even after leaving it running for a day or two it doesn't seem to be doing anything. Edit: This was for a rule based wordlist attack, it got stuck on 100% CPU for a brute force attack no matter what settings I changed |
|
Can you check which threads are running for your hashcat program when the memory usage increased that dramatically? Also, which of the threads exactly take up all the CPU? This can e.g. be done by installing |
|
It doesn't seem to use as much memory as before now, when running a wordlist it first uses around 200mb of ram running
using 1 core. After a bit, it starts running this whole time. After around 10 min, it outputs When running in brute force mode, it also runs clang for a bit, then the cpu gets stuck on 100% on all four cores from |
|
I looked into this the last few days and the same hashcat command ( Some analysis showed that hashcat itself seem to allocate >700MB on my machine of that at various positions, e.g.
I still will be trying to reduce the memory usage of different types and algorithms, but that might not be enough to be able to run that command on a Raspberry Pi at all, at least not without tuning the hashcat parameters and/or build options. |
|
Any idea what causes the cpu to be stuck on 100% while not progressing in brute force mode? |
|
Some:
|
|
@ling1729 could you run without errors? I got following errors: |
|
I can confirm that I am getting the same error on the latest commit |
|
I'm seeing a slightly different error with hashcat on a Raspberry Pi 2: |
|
Can anyone of you run the hashcat command with the |
|
|
|
So the problem is that hashcat read 64-bit kernel parameters, which are not yet supported (like most 64-bit integer related stuff). |
|
Ouch. I'm guessing that will take some time to implement, if you're able to do it at all? |
|
Support for 64-bit integer parameters by itself won't take that much time, I will try to have a look at that in the next days. But in general, also some things work with 64-bit integers (e.g. some bitwise operations, addition, memory access), they are actually not really supported so most likely fixing that particular error will only cause another compilation error for some other 64-bit operation... |
|
Well, let us know when 64 bit integers are working. I'll be happy to do further testing - I have a cluster of 5 RPi2s here I'm playing with, so I'll be happy to provide any bug reports I can. |
|
So the particular problem with not being able to pass in 64-bit integer directly to kernels should be fixed now. For the fix, all 3 repositories need to be updated. |
|
In particular this error from #81 (comment) seems to be fixed. But now hashcat got killed by oom_reaper. |
|
Yes, that is what I meant. For the out of memory, I also already debugged the reason for that in #81 (comment). So without modifying the hashcat code, I doubt it will actually work on a Raspberry Pi. |
Give your Pi 1G of swap. hashcat seems to need ~ 1.2-1.3G of memory just for the MD5 benchmark. |
|
Got much further this time, but still died during compilation (vc4cl-binary-1350418454.bin is empty, so I didn't attach it). vc4cl-ir-558203856.ll.zip |
|
@doe300, any followup here? |
|
I did not do anything specific in that directory and just tested with my latest development build, still getting the same errors. |
|
I threw some unsalted MD5 hashes at it, and verified that the same register errors show up - it's not just the benchmark (which makes sense, but I figured I'd verify it anyway). |
|
I'm also having the issue that Hashcat is not working on my Raspberry Pi 3B+. When I run hashcat with When running an example, it at first runs clang and then crashes with I installed VC4CL normally following the guide and compiled hashcat from source normally. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
When I try to run any Hashcat attack on a raspberry pi 3 with VC4CL, I get the error
clBuildProgram(): CL_COMPILE_PROGRAM_FAILURE, I have installed VC4CL according to here, https://github.com/aaguiar96/VC4CL-Install. Can anybody help me figure out what went wrong?The text was updated successfully, but these errors were encountered: