1 The powerpc assembler can generate .PPC.EMB.apuinfo sections in some
2 circumstances. These originate, I think, from the old e500 ABI; in
3 any case, they are only defined for 32-bit ABIs, and there is a
4 significant amount of code in bfd/elf32-ppc.c to handle them
5 appropriately, none of which is in elf64-ppc.c. However, the
6 assembler can generate them when building 64-bit code for processors
7 such as e5500 (which uses PPC_OPCODE_E500MC). In such circumstances,
8 they get generated as PROGBITS rather than NOTE sections, which
9 reportedly can cause problems in a Linux kernel build.
11 Since these sections are only defined in 32-bit ABIs, I think it's
12 most appropriate to avoid generating them for 64-bit output, which
15 Upstream-Status: Backport
19 From d77411f56a85225e3f7b674e7172c712f6cd61cb Mon Sep 17 00:00:00 2001
20 From: Joseph Myers <joseph@codesourcery.com>
21 Date: Fri, 25 Jan 2013 02:02:20 +0000
22 Subject: [PATCH] * config/tc-ppc.c (md_assemble): Do not generate
23 APUinfo sections for 64-bit output.
26 gas/ChangeLog | 5 +++++
27 gas/config/tc-ppc.c | 3 ++-
28 2 files changed, 7 insertions(+), 1 deletion(-)
30 diff --git a/gas/config/tc-ppc.c b/gas/config/tc-ppc.c
31 index fd15bea..0929e52 100644
32 --- a/gas/config/tc-ppc.c
33 +++ b/gas/config/tc-ppc.c
34 @@ -3249,7 +3249,8 @@ md_assemble (char *str)
37 /* Do we need/want an APUinfo section? */
38 - if ((ppc_cpu & (PPC_OPCODE_E500 | PPC_OPCODE_E500MC | PPC_OPCODE_VLE)) != 0)
39 + if ((ppc_cpu & (PPC_OPCODE_E500 | PPC_OPCODE_E500MC | PPC_OPCODE_VLE)) != 0
42 /* These are all version "1". */
43 if (opcode->flags & PPC_OPCODE_SPE)
