[Bug gas/3607] GAS errors on valid code complaining missing .cfi_endproc

[raj dot khem at gmail dot com]

------- Additional Comments From raj dot khem at gmail dot com  2006-12-01 
22:33 -------
Hi Nick

Yes its not reproducible with cvs binutils. I realised that after a while. Its 
one of the local patches I apply 
than this issue happens. This has been working. 

I debugged gas a bit. 

My patch calls make_relative_prefix() and if I apply my patch and disable 
make_relative_prefix() call it 
works ok too.

The problem seems to be that the check inside gas/dw2gencfi.c:769 
dot_cfi_startproc() assumes that  if 
(frchain_now->frch_cfi_data != NULL)

frchain_now->frch_cfi_data is NULL when initialised

when I look into the gas/subsegs.c:110

where the frchain_now is allocated then there is obstack_alloc call and this 
member is not set to NULL 
explicitly.

So unless obstack_alloc returns a 0 initialised memory area the above check in 
dot_cfi_startproc() is not 
correct 


-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=3607

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.