[PATCH] D89049: [AIX][XCOFF] print out the traceback info

[David Tenty via Phabricator via llvm-commits]

daltenty added a comment.

The traceback decode format here is quite different from what binutils objdump would generate on AIX:

  $ objdump -D --private=traceback  xcoff-traceback-table.o
  
  xcoff-traceback-table.o:     file format aixcoff-rs6000
  00000000: .foo
   tags at 0000008c
   version: 0, lang: 0, global_link: 0, is_eprol: 0, has_tboff: 1, int_proc: 0
   has_ctl: 0, tocless: 0, fp_pres: 1, log_abort: 0, int_hndl: 0
   name_pres: 1, uses_alloca: 0, cl_dis_inv: 0, saves_cr: 0, saves_lr: 0
   stores_bc: 1, fixup: 0, fpr_saved: 0 , spare3: 0, gpr_saved: 1
   fixparms: 2    floatparms: 3    parm_on_stk: 1
   parminfo: 0x5a000000
   tb_offset: 0x00000088 (start=0x00000000)
   Name (len: 3): foo
   (end of tags at 000000a1)
  
  
  Disassembly of section .text:
  
  00000000 <.foo>
  ...

I'm not sure if having such a large discrepancy in decode format is a good idea for llvm-objdump.



================
Comment at: llvm/test/tools/llvm-objdump/XCOFF/disassemble-invalid-traceback-table.test:3
+
+# RUN: llvm-objdump -D --traceback-table --symbol-description %p/Inputs/xcoff-invalid-traceback-table.o | \
+# RUN:   FileCheck %s
----------------
Should we be expecting a non-zero exit code here?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D89049/new/

https://reviews.llvm.org/D89049