Roland McGrath mcgrathr@google.com
Thu Mar 13 17:32:00 GMT 2014

• Previous message (by thread): [RFC PATCH] Avoid emitting TEXTREL marker for R_ARM_NONE relocs.
• Next message (by thread): [RFC PATCH] Avoid emitting TEXTREL marker for R_ARM_NONE relocs.
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

---

Alan,

It was indeed .eh_frame that led to it.  Here's a trivial test case:

	$ head eh[123].s
	==> eh1.s <==
	.cfi_sections .eh_frame
	foo:
		.cfi_startproc
		.cfi_personality 0, pers
		.cfi_endproc

	==> eh2.s <==
	.cfi_sections .eh_frame
	bar:
		.cfi_startproc
		.cfi_personality 0, pers
		.cfi_endproc

	==> eh3.s <==
	.globl pers
	pers:
	nop
	$ ./ld/ld-new --eh-frame-hdr -shared -o eh.so eh[123].o
	$ readelf -dr eh.so

	Dynamic section at offset 0x260 contains 10 entries:
	  Tag        Type                         Name/Value
	 0x00000004 (HASH)                       0x100000d4
	 0x00000005 (STRTAB)                     0x100001b0
	 0x00000006 (SYMTAB)                     0x10000110
	 0x0000000a (STRSZ)                      64 (bytes)
	 0x0000000b (SYMENT)                     16 (bytes)
	 0x00000011 (REL)                        0x100001f0
	 0x00000012 (RELSZ)                      16 (bytes)
	 0x00000013 (RELENT)                     8 (bytes)
WRONG=>	 0x00000016 (TEXTREL)                    0x0
	 0x00000000 (NULL)                       0x0

	Relocation section '.rel.dyn' at offset 0x1f0 contains 2 entries:
	 Offset     Info    Type            Sym.Value  Sym. Name
UGLY=>	00000000  00000000 R_ARM_NONE
	1000022e  00000202 R_ARM_ABS32       00000000   pers
	$


It's no longer an urgent issue for me, but I think it deserves to be fixed.


Thanks,
Roland

---

• Previous message (by thread): [RFC PATCH] Avoid emitting TEXTREL marker for R_ARM_NONE relocs.
• Next message (by thread): [RFC PATCH] Avoid emitting TEXTREL marker for R_ARM_NONE relocs.
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

---

More information about the Binutils mailing list