X86 LTS

[nickdesaulniers]

branched off of #40 . Let's see what's green for x86.

[nathanchance]

4.19 needs at least ClangBuiltLinux/linux@81b4568 applied to it.

I'm looking into the 4.14 fail now.

[nickdesaulniers]

I'm looking into the 4.14 fail now.

FATAL: kernel too old

[   29.860060] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00007f00
[   29.860060] 
[   29.861196] CPU: 0 PID: 1 Comm: init Not tainted 4.14.80 #1
[   29.861710] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.1-1 04/01/2014
[   29.862527] Call Trace:
[   29.863291]  dump_stack+0xa6/0x100
[   29.863692]  panic+0xf4/0x2b0
[   29.863989]  ? do_exit+0x88f/0x910
[   29.864303]  do_exit+0x905/0x910
[   29.864688]  do_group_exit+0xc2/0xd0
[   29.865101]  SYSC_exit_group+0xb/0x10
[   29.865364]  SyS_exit_group+0x5/0x10
[   29.865575]  do_syscall_64+0x6a/0x80
[   29.865819]  entry_SYSCALL_64_after_hwframe+0x42/0xb7
[   29.866424] RIP: 0033:0x7f3ab2ff16a6
[   29.866687] RSP: 002b:00007fff08a33ee8 EFLAGS: 00000206 ORIG_RAX: 00000000000000e7
[   29.867305] RAX: ffffffffffffffda RBX: 00007f3ab2ff7088 RCX: 00007f3ab2ff16a6
[   29.867703] RDX: 000000000000007f RSI: 000000000000003c RDI: 000000000000007f
[   29.868160] RBP: 00007fff08a340d0 R08: 00000000000000e7 R09: 00007fff08a33e18
[   29.868558] R10: ffff8000f7528864 R11: 0000000000000206 R12: 00007fff08ad72dc
[   29.869074] R13: 0000000000000010 R14: 0000000000000000 R15: 00007f3ab3000190
[   29.870599] Kernel Offset: 0x1f000000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff)
[   29.871855] ---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x00007f00

A passing x86_64 kernel has:

[   19.359141] Freeing unused kernel image memory: 2000K
[   19.366007] Freeing unused kernel image memory: 1920K
[   19.367007] rodata_test: all tests were successful
[   19.368651] Run /sbin/init as init process
[   19.795726] EXT4-fs (sda): re-mounted. Opts: (null)
[   19.892707] mkdir (1178) used greatest stack depth: 13072 bytes left

A failing 4.14 kernel has:

[   29.730921] Freeing unused kernel memory: 2000K
[   29.735362] Freeing unused kernel memory: 136K
[   29.736465] rodata_test: all tests were successful
FATAL: kernel too old
[   29.860060] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00007f00
<above panic>

4.4 and 4.9 have the same failures as arm.

[nickdesaulniers]

FATAL: kernel too old

I wonder if glibc in our buildroot is failing to parse our kernel version string correctly?

Or it looks like our buildroot might need to have the correct headers: https://forums.xilinx.com/t5/Embedded-Linux/Linux-build-for-MicroZED-FATAL-kernel-too-old/td-p/812467

[nathanchance]

I'll check the configuration, thanks for the link!

[nathanchance]

Yeah I think that's probably what it is because the default header value is 4.17. I've changed it to 4.4.

[nathanchance]

Thanks for the tip @nickdesaulniers!

Linux version 4.14.80 (nathan@flashbox) (clang version 8.0.0-r346189 (https://github.com/nathanchance/clang 79a31f952e7f1c8f8977f896bffc4f080a764d0a) (https://git.llvm.org/git/llvm a8f890f4777f53bd9243bfdd9ed2b5d171cc8c03)) #1 SMP Mon Nov 12 19:20:17 MST 2018

#56

[nickdesaulniers]

build 269 shows x86 4.14 passing.

build 271 shows x86 4.19.

[nathanchance]

nathanchance@7277201

https://travis-ci.com/nathanchance/continuous-integration/jobs/158116418

[dileks]

On Tue, Nov 13, 2018 at 11:55 AM Nathan Chancellor ***@***.***> wrote: ***@***.*** <nathanchance@7277201>

I would keep the echo-line and do an "exit 0" instead in ... Subject: [PATCH 1/2] DO-NOT-UPSTREAM: x86: Avoid warnings/errors due to lack of asm goto ... @echo Compiler lacks asm-goto support. + @EXIT 0 +KBUILD_CFLAGS += -D__BPF_TRACING__ ... - sed@ -

…

https://travis-ci.com/nathanchance/continuous-integration/jobs/158116418 — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#44 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAsJyCH0WmA0_0T8kEpPBI0AnM1hLL4Fks5uuqU7gaJpZM4YYzTp> .

[nathanchance]

I would keep the echo-line and do an "exit 0"

I'm curious, why? It doesn't avoid potential merge conflicts and we don't need to see that message every build since the whole reason the patch exists is because Clang doesn't support asm-goto.

[dileks]

On Tue, Nov 13, 2018 at 6:40 PM Nathan Chancellor ***@***.***> wrote: I would keep the echo-line and do an "exit 0" I'm curious, why? It doesn't avoid potential merge conflicts and we don't need to see that message every build since the whole reason the patch exists is because Clang doesn't support asm-goto. I hope to see asm-goto support for clang-8, then I do not care :-).

- sed@ -

…

[nickdesaulniers]

build 271 shows x86 4.19.

The job exceeded the maximum log length, and has been terminated.

Heh, looks like I'll have to send ClangBuiltLinux/linux@81b4568 back to stable 4.19 (possibly further)

[nickdesaulniers]

ready for review.

[nathanchance]

LGTM. Needs to be rebased on master. Also, we could enable 4.14 right now, it doesn't need any special patches applied to it: https://travis-ci.com/nathanchance/continuous-integration/jobs/158594813

[nickdesaulniers]

only cron jobs added. bypassing presubmit.

[nickdesaulniers]

Also, we could enable 4.14 right now

done