Segmentation fault RpiZeroW

[dr-noob]

Segmentation fault RpiZeroW

[Dr-Noob]

Hi,

Thank you very much for your issue report. It looks like you found an important bug.

It's pretty hard to debug a segfault just looking at thousand lines of code, so I would need some additional information:

1. Build cpufetch with debug symbols (make clean; make debug).
2. Debug cpufetch(gdb cpufetch)
3. Run cpufetch (just r inside gdb)

And finally, paste here the output of the execution. It should tell you where did the segmentation fault occurred.

Thanks!

[felipefmartinez]

Starting program: /home/felipe/cpufetch/cpufetch

Program received signal SIGSEGV, Segmentation fault.
0xb6fbbb10 in memcpy () from /usr/lib/arm-linux-gnueabihf/libarmmem-v6l.so

[Dr-Noob]

Hi,

Sorry but that is not enough information to properly track the issue...Can you run cpufetch under valgrind? Steps:

1. Install valgrind (I hope it can be installed with sudo apt-get install valgrind).
2. Build cpufetch with debug symbols (make clean; make debug).
3. Run cpufetch with valgrind (valgrind ./cpufetch).

Paste the full output here. It should be clear enough to understand what's happening.

[felipefmartinez]

https://pastebin.com/KJgbikpf
I did it both with and without --leak-check=full. I'll post the other one in a second

[felipefmartinez]

https://pastebin.com/J703u5gi

[Dr-Noob]

It looks like this is going to be a hard bug to find.

I've just pushed a commit to bugfix2 branch. Change to it (git checkout bugfix2), and let's see how it goes.

[felipefmartinez]

https://pastebin.com/cedzPXin

Standard run (with make debug)

Valgrind on the way

[felipefmartinez]

https://pastebin.com/2FzSPbZC
Valgrind output of make debug of branch bugfix2

[Dr-Noob]

Try again with the latest commit in bugfix2 and paste the output, but just the output of cpufetch (no gdb or valgrind this time). Paste also the contents of the file /sys/devices/system/cpu/present. Thanks.

I expect to solve the "Microarchitecture: Unknown" part with this commit. For the issue of number of cores, I will wait to see the contents of the file I told you about, and I will decide what to do then.

[felipefmartinez]

Did a git pull, make clean, make

Here's the result

https://pastebin.com/dgw64Vvi

[Dr-Noob]

Great, the microarchitecture problem is solved, but you forgot to post the contents of the file I mentioned.

Would you mind to paste them here? Thanks

[felipefmartinez]

Shoot my bad. I just did a sudo cat /sys/devices/system/cpu/present

the result is a 0
that's it, jsut a fat 0 ?

[Dr-Noob]

Whooops!

Yes, it makes sense that it contains just a fat 0. The problem is that when I wrote the code to query the number of cores I didn't consider the case of a CPU with just one core, that's why in your case it was not working (Segmentation fault). In the previous commit I just made the code to always report that the CPU has 1 core (I did this for debugging purposes). I have pushed a commit in which the code reports the number of cores again, but considering the case of just one core.

So I need you to update again and run the latest commit. Everything should work well. If this is the case, I think we solved the issue! Let me know how it goes.

[felipefmartinez]

worked fine from the bugfix2 branch pulled just now.

[Dr-Noob]

Then I think we can consider the issue solved. Thank you very much for reporting this and help me by running all the tests, I have learned a lot!

I will merge the changes into master branch soon. Closing the issue.

[felipefmartinez]

No problem!

…

On Wed, Apr 7, 2021, 14:49 Dr-Noob ***@***.***> wrote: Then I think we can consider the issue solved. Thank you very much for reporting this and help me by running all the tests, I have learned a lot! I will merge the changes into master branch soon. Closing the issue. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#44 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFZ35CQASE3S6BP6FV64WPDTHSSKRANCNFSM42L23R4A> .