| 9 Nov 2025 |
|  @mrargoz:matrix.org left the room. | 17:56:41 |
 Tristan Ross | I have a system with an ESWIN EIC770X CPU in it. Tried building hello and got an error about bootstrap-tools segfaulting. Seems like patchelf and all of the bootstrap tools don't want to execute. | 21:59:48 |
 Sergei Zimmerman (xokdvium) | What kernel is that running? Cross-compiled NixOS or vendor's system? Maybe some kernel-headers nonsense | 22:15:25 |
| Tristan Ross | 6.6.92-eic7x-2025.07 | 22:19:49 |
| Tristan Ross | https://github.com/DC-DeepComputing/DC-linux/ | 22:20:00 |
| Tristan Ross | It's running Ubuntu right now | 22:20:05 |
| Tristan Ross | I'm not doing cross, I'm doing native | 22:24:16 |
| Sergei Zimmerman (xokdvium) | Where is it segfaulting? somewhere in musl? | 22:24:23 |
| Tristan Ross | I can't get a backtrace | 22:24:33 |
| Tristan Ross | If I execute the ld-linux shared library, it does print. But grep, awk, and patchelf can't be executed. | 22:25:11 |
| Sergei Zimmerman (xokdvium) | So static executables built with nixpkgs musl are borked with that kernel? | 22:26:11 |
| Sergei Zimmerman (xokdvium) | As in they don't even make their way to main? | 22:26:21 |
| Tristan Ross | Yes | 22:26:41 |
| Tristan Ross | Even if I do start in GDB, I get a segfault on calling gcc | 22:26:59 |
| Tristan Ross | nproc works | 22:28:10 |
| Sergei Zimmerman (xokdvium) | Then maybe put a breakpoint on the elf entrypoint and step instruction-by-instruction? | 22:28:31 |
| Sergei Zimmerman (xokdvium) | I don't know if that will work even | 22:28:46 |
| Tristan Ross | Even if I set a breakpoint on _start, it still says segfault | 22:29:47 |
| Tristan Ross | If I do strace on awk, I get EPERM on execve | 22:30:51 |
| Tristan Ross | awk has 777 so it should be executable | 22:32:14 |
| Tristan Ross | AppArmor is enabled | 22:32:57 |
| Tristan Ross | It still segfaulted | 22:40:48 |
| Tristan Ross | Same EPERM with AppArmor off | 22:41:15 |
| Sergei Zimmerman (xokdvium) | bootstrap tarball looks very weird to me | 22:41:41 |
| Tristan Ross | Hmm, so it's a bootstrap tarball problem that probably should be fixed? | 22:42:19 |
| Sergei Zimmerman (xokdvium) | Yeah no idea. Do they happen to have a somewhat representative qemu-system setup to debug this without hardware? | 22:53:10 |
| Tristan Ross | I don't have a VM set up | 22:56:50 |
| 10 Nov 2025 |
 Alex | In reply to @rosscomputerguy:matrix.org
I have a system with an ESWIN EIC770X CPU in it. Tried building hello and got an error about bootstrap-tools segfaulting. Seems like patchelf and all of the bootstrap tools don't want to execute. I haven't seen it mentioned yet so worth a try: have you looked at the kernel log for hints? | 16:39:34 |
| Tristan Ross | In reply to @alex:tunstall.xyz
I haven't seen it mentioned yet so worth a try: have you looked at the kernel log for hints? I looked and the only thing I saw was EDID logging | 16:45:43 |
| Tristan Ross | LD vmlinux
BTFIDS vmlinux
FAILED elf_update(WRITE): no error
make[2]: *** [../scripts/Makefile.vmlinux:37: vmlinux] Error 255
make[2]: *** Deleting file 'vmlinux'
make[1]: *** [/build/source/Makefile:1165: vmlinux] Error 2
make: *** [../Makefile:234: __sub-make] Error 2
Wtf is this error? I'm trying to build the kernel for this system.
| 21:45:42 |
