dwarves.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* Segfault with LTO'ed Linux
@ 2020-11-26  8:01 Bill Wendling
  0 siblings, 0 replies; only message in thread
From: Bill Wendling @ 2020-11-26  8:01 UTC (permalink / raw)
  To: Arnaldo Carvalho de Melo, dwarves

I just ran the newest pahole release on the Linux kernel compiled with
clang plus LTO (link-time optimizations) and got a segmentation fault.
I placed the vmlinux file at the following URL:

https://drive.google.com/file/d/1KmrUFLBl42U1VZ3elp6zTlpK7rB7o4Jq/view?usp=sharing

The DWARF information generated by clang + LTO is definitely different
from a non-LTO build. We've seen issues with it in the past. This
however looks like a new issue.

[Note: I could only share the file with individuals with gmail
accounts, so I wasn't able to share it with the
"dwarves@vger.kernel.org" account.]

$ pahole -J -V tmp_vmlinux.btf
Found per-CPU symbol 'cpu_loops_per_jiffy' at address 0x10000
Found per-CPU symbol 'pmc_prev_left' at address 0x11330
Found per-CPU symbol 'perf_nmi_tstamp' at address 0x11530
Found per-CPU symbol 'bts_ctx' at address 0x12000
Found per-CPU symbol 'insn_buffer' at address 0x15000
Found per-CPU symbol 'pt_ctx' at address 0x15008
Found per-CPU symbol 'kvm_core_conflict' at address 0x150d8
Found per-CPU symbol 'svm_asid' at address 0x150e8
Found per-CPU symbol 'current_vcpu' at address 0x150f8
Found per-CPU symbol 'mmu_lock_acquired_time' at address 0x150f0
Found per-CPU symbol 'cpu_tsc_khz' at address 0x15100
Found per-CPU symbol 'current_vcpu' at address 0x15108
Found per-CPU symbol 'local_vcpu' at address 0x15110
Found per-CPU symbol 'last_nmi_rip' at address 0x15288
Found per-CPU symbol 'nmi_cr2' at address 0x15278
Found per-CPU symbol 'nmi_state' at address 0x15270
Found per-CPU symbol 'nmi_stats' at address 0x15294
Found per-CPU symbol 'swallow_nmi' at address 0x152a4
Found per-CPU symbol 'unhandled_nmi_count' at address 0x15290
Found per-CPU symbol 'update_debug_stack' at address 0x15280
Found per-CPU symbol 'cpu_devices' at address 0x15ab0
Found per-CPU symbol 'bp_per_reg' at address 0x15da0
Found per-CPU symbol 'cpu_debugreg' at address 0x15dc0
Found per-CPU symbol 'is_idle' at address 0x15e00
Found per-CPU symbol 'ssb_state' at address 0x15de8
Found per-CPU symbol 'in_kernel_fpu' at address 0x15e18
Found per-CPU symbol 'debug_stack_addr' at address 0x15e60
Found per-CPU symbol 'samples' at address 0x15ea8
Found per-CPU symbol 'last_mds_flush' at address 0x15f18
Found per-CPU symbol 'mds_core_conflict' at address 0x15ed0
Found per-CPU symbol 'mds_reasons' at address 0x15ee0
Found per-CPU symbol 'mds_sw_flush_buffer' at address 0x15ed8
Found per-CPU symbol 'mds_triggers' at address 0x15f00
Found per-CPU symbol 'mce_next_interval' at address 0x16050
Found per-CPU symbol 'mce_signaled' at address 0x15fa4
Found per-CPU symbol 'mce_timer' at address 0x16028
...
[1759] FUNC arch_setup_additional_pages type_id=1758
[1760] FUNC_PROTO (anon) return=381 args=(1750 bprm, 381 uses_interp)
[1761] FUNC compat_arch_setup_additional_pages type_id=1760
[1762] FUNC_PROTO (anon) return=381 args=(382 s)
[1763] FUNC vdso_setup type_id=1762
[1764] FUNC_PROTO (anon) return=381 args=(void)
[1765] FUNC init_vdso type_id=1764
Segmentation fault

-bw

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2020-11-26  8:01 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-11-26  8:01 Segfault with LTO'ed Linux Bill Wendling

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).