- Update config files.
- Refresh patches.suse/stack-unwind.
Reenable dwarf stack unwinder for C. Assembler is not dwarf-friendly
annotated (removed in upstream and neither
patches.arch/x86_64-unwind-annotations is enabled (yet)), so dwarf
unwinder does not work there (yet).
This should be fixed by Josh Poimbeouf's stacktool later. But until
then, dwarf-unwind at least C and use frame pointers for asm.
I.e. no more question marks for most cases:
[<ffffffff81008415>] try_stack_unwind+0x175/0x190
[<ffffffff81007293>] dump_trace+0x93/0x3a0
[<ffffffff8100847b>] show_trace_log_lvl+0x4b/0x60
[<ffffffff810076ac>] show_stack_log_lvl+0x10c/0x180
[<ffffffff81008515>] show_stack+0x25/0x50
[<ffffffff81364619>] dump_stack+0x4b/0x72
[<ffffffffa01a302e>] init1+0xe/0x20 [aaa]
[<ffffffff810002f3>] do_one_initcall+0xb3/0x200
[<ffffffff8116e4da>] do_init_module+0x5f/0x1f6
[<ffffffff810ee87e>] load_module+0x15fe/0x1b60
[<ffffffff810eefe8>] SYSC_finit_module+0x98/0xc0
[<ffffffff810ef02e>] SyS_finit_module+0xe/0x10
[<ffffffff816b07b6>] entry_SYSCALL_64_fastpath+0x16/0x75
DWARF2 unwinder stuck at entry_SYSCALL_64_fastpath+0x16/0x75
Leftover inexact backtrace:
<nothing>
Instead of the funny:
[<ffffffff81364619>] ? dump_stack+0x4b/0x72
[<ffffffffa022a02e>] ? init1+0xe/0x20 [aaa]
[<ffffffff810002f3>] ? do_one_initcall+0xb3/0x200
[<ffffffff8116e4a2>] ? do_init_module+0x27/0x1f6
[<ffffffff8116e4da>] ? do_init_module+0x5f/0x1f6
[<ffffffff810ee87e>] ? load_module+0x15fe/0x1b60
[<ffffffff810eaa40>] ? __symbol_put+0x50/0x50
[<ffffffff811e9501>] ? kernel_read+0x41/0x60
[<ffffffff810eefe8>] ? SYSC_finit_module+0x98/0xc0
[<ffffffff810ef02e>] ? SyS_finit_module+0xe/0x10
[<ffffffff816b07b6>] ? entry_SYSCALL_64_fastpath+0x16/0x75
where __symbol_put and do_init_module+0x27 should not be there at all.