This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
The real cause of the recent Irix 6.5 failure
- To: zackw at stanford dot edu, rth at redhat dot com, oldham at codesourcery dot com,David dot Billinghurst at riotinto dot com
- Subject: The real cause of the recent Irix 6.5 failure
- From: "H . J . Lu" <hjl at lucon dot org>
- Date: Sat, 11 Aug 2001 11:08:47 -0700
- Cc: gcc-bugs at gcc dot gnu dot org
Hi,
I don't believe Zack's patch is the real cause of the recent Irix 6.5
failure. I got the same problem on Irix 6.5:
stage1/xgcc -Bstage1/ -B/export/tools/mips-sgi-irix6.5/bin/ -DIN_GCC -g -O2 -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wtraditional -pedantic -Wno-long-long -DHAVE_CONFIG_H -DGENERATOR_FILE -o gengenrtl \ gengenrtl.o ` case "obstack.o" in ?*) echo obstack.o ;; esac ` ` case "" in ?*) echo ;; esac ` ` case "" in ?*) echo ;; esac ` ` case "" in ?*) echo ;; esac `
./gengenrtl -h > tmp-genrtl.h
# nm gengenrtl.o
gengenrtl.o ERROR: dwarf_elf_init: DW_DLE_DEBUG_LOC_NULL (40) .debug_loc section present but elf_getdata() failed or section is zero-length
When I generated gengenrtl.s and removed the empty .debug_loc section
by hand, the resulting gengenrtl worked fine.
It seems to me that it is an Irix 6.5 dynamic linker bug trigger by
the empty .debug_loc section. Now the questions are
1. Can we avoid creating the empty .debug_loc section?
2. How hard to find a fix for the Irix dynamic linker?
I think we should do #1.
H.J.