/* SPARC-specific support for 64-bit ELF
- Copyright (C) 1993, 95, 96, 97, 98, 99, 2000
+ Copyright 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001
Free Software Foundation, Inc.
This file is part of BFD, the Binary File Descriptor library.
#include "sysdep.h"
#include "libbfd.h"
#include "elf-bfd.h"
+#include "opcode/sparc.h"
/* This is defined if one wants to build upward compatible binaries
with the original sparc64-elf toolchain. The support is kept in for
static void sparc64_elf_symbol_processing
PARAMS ((bfd *, asymbol *));
+static boolean sparc64_elf_copy_private_bfd_data
+ PARAMS ((bfd *, bfd *));
static boolean sparc64_elf_merge_private_bfd_data
PARAMS ((bfd *, bfd *));
+static boolean sparc64_elf_relax_section
+ PARAMS ((bfd *, asection *, struct bfd_link_info *, boolean *));
static boolean sparc64_elf_relocate_section
PARAMS ((bfd *, struct bfd_link_info *, bfd *, asection *, bfd_byte *,
Elf_Internal_Rela *, Elf_Internal_Sym *, asection **));
HOWTO(R_SPARC_GLOB_DAT, 0,0,00,false,0,complain_overflow_dont, bfd_elf_generic_reloc, "R_SPARC_GLOB_DAT",false,0,0x00000000,true),
HOWTO(R_SPARC_JMP_SLOT, 0,0,00,false,0,complain_overflow_dont, bfd_elf_generic_reloc, "R_SPARC_JMP_SLOT",false,0,0x00000000,true),
HOWTO(R_SPARC_RELATIVE, 0,0,00,false,0,complain_overflow_dont, bfd_elf_generic_reloc, "R_SPARC_RELATIVE",false,0,0x00000000,true),
- HOWTO(R_SPARC_UA32, 0,0,00,false,0,complain_overflow_dont, bfd_elf_generic_reloc, "R_SPARC_UA32", false,0,0x00000000,true),
+ HOWTO(R_SPARC_UA32, 0,2,32,false,0,complain_overflow_bitfield,bfd_elf_generic_reloc, "R_SPARC_UA32", false,0,0xffffffff,true),
#ifndef SPARC64_OLD_RELOCS
/* These aren't implemented yet. */
HOWTO(R_SPARC_PLT32, 0,0,00,false,0,complain_overflow_dont, sparc_elf_notsup_reloc, "R_SPARC_PLT32", false,0,0x00000000,true),
/* Due to the way how we handle R_SPARC_OLO10, each entry in a SHT_RELA
section can represent up to two relocs, we must tell the user to allocate
more space. */
-
+
static long
sparc64_elf_get_reloc_upper_bound (abfd, sec)
bfd *abfd ATTRIBUTE_UNUSED;
return _bfd_elf_get_dynamic_reloc_upper_bound (abfd) * 2;
}
-/* Read relocations for ASECT from REL_HDR. There are RELOC_COUNT of
+/* Read relocations for ASECT from REL_HDR. There are RELOC_COUNT of
them. We cannot use generic elf routines for this, because R_SPARC_OLO10
has secondary addend in ELF64_R_TYPE_DATA. We handle it as two relocations
for the same location, R_SPARC_LO10 and R_SPARC_13. */
entsize = rel_hdr->sh_entsize;
BFD_ASSERT (entsize == sizeof (Elf64_External_Rela));
-
+
count = rel_hdr->sh_size / entsize;
for (i = 0, relent = relents; i < count;
rel_hdr2 = NULL;
}
- asect->relocation = ((arelent *)
- bfd_alloc (abfd,
+ asect->relocation = ((arelent *)
+ bfd_alloc (abfd,
asect->reloc_count * 2 * sizeof (arelent)));
if (asect->relocation == NULL)
return false;
/* The sparc64_elf_slurp_one_reloc_table routine increments reloc_count. */
asect->reloc_count = 0;
-
+
if (!sparc64_elf_slurp_one_reloc_table (abfd, asect, rel_hdr, symbols,
dynamic))
return false;
-
- if (rel_hdr2
+
+ if (rel_hdr2
&& !sparc64_elf_slurp_one_reloc_table (abfd, asect, rel_hdr2, symbols,
dynamic))
return false;
if (rela_hdr->sh_type != SHT_RELA)
abort ();
- /* orelocation has the data, reloc_count has the count... */
+ /* orelocation has the data, reloc_count has the count... */
outbound_relocas = (Elf64_External_Rela *) rela_hdr->contents;
src_rela = outbound_relocas;
#define sparc64_elf_hash_table(p) \
((struct sparc64_elf_link_hash_table *) ((p)->hash))
-
+
/* Create a Sparc64 ELF linker hash table. */
static struct bfd_link_hash_table *
return &ret->root.root;
}
-
\f
/* Utility for performing the standard initial work of an instruction
relocation.
#define ELF_DYNAMIC_INTERPRETER "/usr/lib/sparcv9/ld.so.1"
-
/* Fill in the .plt section. */
static void
{
const unsigned int nop = 0x01000000;
int i, j;
-
+
/* The first four entries are reserved, and are initially undefined.
We fill them with `illtrap 0' to force ld.so to do something. */
/* Now the tricky bit. Entries 32768 and higher are grouped in blocks of
160: 160 entries and 160 pointers. This is to separate code from data,
which is much friendlier on the cache. */
-
+
for (; i < nentries; i += 160)
{
int block = (i + 160 <= nentries ? 160 : nentries - i);
block = (index - LARGE_PLT_THRESHOLD) / 160;
ofs = (index - LARGE_PLT_THRESHOLD) % 160;
- return ((bfd_vma)(LARGE_PLT_THRESHOLD + block*160) * PLT_ENTRY_SIZE
+ return ((bfd_vma) (LARGE_PLT_THRESHOLD + block*160) * PLT_ENTRY_SIZE
+ ofs * 6*4);
}
+ last * 6*4
+ ofs * 8);
}
-
-
\f
/* Look through the relocs for a section during the first phase, and
allocate space in the global offset table or procedure linkage
srelgot = NULL;
sreloc = NULL;
- rel_end = relocs + sec->reloc_count;
+ rel_end = relocs + elf_section_data (sec)->rel_hdr.sh_size
+ / elf_section_data (sec)->rel_hdr.sh_entsize;
for (rel = relocs; rel < rel_end; rel++)
{
unsigned long r_symndx;
case R_SPARC_UA16:
/* When creating a shared object, we must copy these relocs
into the output file. We create a reloc section in
- dynobj and make room for the reloc.
+ dynobj and make room for the reloc.
But don't do this for debugging sections -- this shows up
with DWARF2 -- first because they are not loaded, and
break;
default:
- (*_bfd_error_handler)(_("%s: check_relocs: unhandled reloc type %d"),
+ (*_bfd_error_handler) (_("%s: check_relocs: unhandled reloc type %d"),
bfd_get_filename(abfd),
ELF64_R_TYPE_ID (rel->r_info));
return false;
{
int reg;
struct sparc64_elf_app_reg *p;
-
+
reg = (int)sym->st_value;
switch (reg & ~1)
{
if (**namep)
{
struct elf_link_hash_entry *h;
-
+
h = (struct elf_link_hash_entry *)
bfd_link_hash_lookup (info->hash, *namep, false, false, false);
s = bfd_get_section_by_name (dynobj, ".rela.plt");
BFD_ASSERT (s != NULL);
- /* The first plt entries are reserved, and the relocations must
- pair up exactly. */
- if (s->_raw_size == 0)
- s->_raw_size += (PLT_HEADER_SIZE/PLT_ENTRY_SIZE
- * sizeof (Elf64_External_Rela));
-
s->_raw_size += sizeof (Elf64_External_Rela);
/* The procedure linkage table size is bounded by the magnitude
{
if (! bfd_elf64_add_dynamic_entry (info, DT_TEXTREL, 0))
return false;
+ info->flags |= DF_TEXTREL;
}
/* Add dynamic STT_REGISTER symbols and corresponding DT_SPARC_REGISTER
if (app_regs [reg].name != NULL)
{
struct elf_link_local_dynamic_entry *entry, *e;
-
+
if (! bfd_elf64_add_dynamic_entry (info, DT_SPARC_REGISTER, 0))
return false;
return true;
}
\f
+#define SET_SEC_DO_RELAX(section) do { elf_section_data(section)->tdata = (void *)1; } while (0)
+#define SEC_DO_RELAX(section) (elf_section_data(section)->tdata == (void *)1)
+
+static boolean
+sparc64_elf_relax_section (abfd, section, link_info, again)
+ bfd *abfd ATTRIBUTE_UNUSED;
+ asection *section ATTRIBUTE_UNUSED;
+ struct bfd_link_info *link_info ATTRIBUTE_UNUSED;
+ boolean *again;
+{
+ *again = false;
+ SET_SEC_DO_RELAX (section);
+ return true;
+}
+\f
/* Relocate a SPARC64 ELF section. */
static boolean
sgot = splt = sreloc = NULL;
rel = relocs;
- relend = relocs + input_section->reloc_count;
+ relend = relocs + elf_section_data (input_section)->rel_hdr.sh_size
+ / elf_section_data (input_section)->rel_hdr.sh_entsize;
for (; rel < relend; rel++)
{
int r_type;
(!info->shared || info->no_undefined
|| ELF_ST_VISIBILITY (h->other)))))
return false;
- relocation = 0;
+
+ /* To avoid generating warning messages about truncated
+ relocations, set the relocation's address to be the same as
+ the start of this section. */
+
+ if (input_section->output_section != NULL)
+ relocation = input_section->output_section->vma;
+ else
+ relocation = 0;
}
}
off &= ~1;
else
{
- bfd_put_64 (output_bfd, relocation, sgot->contents + off);
local_got_offsets[r_symndx] |= 1;
if (info->shared)
asection *srelgot;
Elf_Internal_Rela outrel;
+ /* The Solaris 2.7 64-bit linker adds the contents
+ of the location to the value of the reloc.
+ Note this is different behaviour to the
+ 32-bit linker, which both adds the contents
+ and ignores the addend. So clear the location. */
+ bfd_put_64 (output_bfd, 0, sgot->contents + off);
+
/* We need to generate a R_SPARC_RELATIVE reloc
for the dynamic linker. */
srelgot = bfd_get_section_by_name(dynobj, ".rela.got");
+ srelgot->reloc_count));
++srelgot->reloc_count;
}
+ else
+ bfd_put_64 (output_bfd, relocation, sgot->contents + off);
}
relocation = sgot->output_offset + off - got_base;
}
relocation = (splt->output_section->vma
+ splt->output_offset
+ sparc64_elf_plt_entry_offset (h->plt.offset));
+ if (r_type == R_SPARC_WPLT30)
+ goto do_wplt30;
goto do_default;
case R_SPARC_OLO10:
}
break;
+ case R_SPARC_WDISP30:
+ do_wplt30:
+ if (SEC_DO_RELAX (input_section)
+ && rel->r_offset + 4 < input_section->_raw_size)
+ {
+#define G0 0
+#define O7 15
+#define XCC (2 << 20)
+#define COND(x) (((x)&0xf)<<25)
+#define CONDA COND(0x8)
+#define INSN_BPA (F2(0,1) | CONDA | BPRED | XCC)
+#define INSN_BA (F2(0,2) | CONDA)
+#define INSN_OR F3(2, 0x2, 0)
+#define INSN_NOP F2(0,4)
+
+ bfd_vma x, y;
+
+ /* If the instruction is a call with either:
+ restore
+ arithmetic instruction with rd == %o7
+ where rs1 != %o7 and rs2 if it is register != %o7
+ then we can optimize if the call destination is near
+ by changing the call into a branch always. */
+ x = bfd_get_32 (input_bfd, contents + rel->r_offset);
+ y = bfd_get_32 (input_bfd, contents + rel->r_offset + 4);
+ if ((x & OP(~0)) == OP(1) && (y & OP(~0)) == OP(2))
+ {
+ if (((y & OP3(~0)) == OP3(0x3d) /* restore */
+ || ((y & OP3(0x28)) == 0 /* arithmetic */
+ && (y & RD(~0)) == RD(O7)))
+ && (y & RS1(~0)) != RS1(O7)
+ && ((y & F3I(~0))
+ || (y & RS2(~0)) != RS2(O7)))
+ {
+ bfd_vma reloc;
+
+ reloc = relocation + rel->r_addend - rel->r_offset;
+ reloc -= (input_section->output_section->vma
+ + input_section->output_offset);
+ if (reloc & 3)
+ goto do_default;
+
+ /* Ensure the branch fits into simm22. */
+ if ((reloc & ~(bfd_vma)0x7fffff)
+ && ((reloc | 0x7fffff) != MINUS_ONE))
+ goto do_default;
+ reloc >>= 2;
+
+ /* Check whether it fits into simm19. */
+ if ((reloc & 0x3c0000) == 0
+ || (reloc & 0x3c0000) == 0x3c0000)
+ x = INSN_BPA | (reloc & 0x7ffff); /* ba,pt %xcc */
+ else
+ x = INSN_BA | (reloc & 0x3fffff); /* ba */
+ bfd_put_32 (input_bfd, x, contents + rel->r_offset);
+ r = bfd_reloc_ok;
+ if (rel->r_offset >= 4
+ && (y & (0xffffffff ^ RS1(~0)))
+ == (INSN_OR | RD(O7) | RS2(G0)))
+ {
+ bfd_vma z;
+ unsigned int reg;
+
+ z = bfd_get_32 (input_bfd,
+ contents + rel->r_offset - 4);
+ if ((z & (0xffffffff ^ RD(~0)))
+ != (INSN_OR | RS1(O7) | RS2(G0)))
+ break;
+
+ /* The sequence was
+ or %o7, %g0, %rN
+ call foo
+ or %rN, %g0, %o7
+
+ If call foo was replaced with ba, replace
+ or %rN, %g0, %o7 with nop. */
+
+ reg = (y & RS1(~0)) >> 14;
+ if (reg != ((z & RD(~0)) >> 25)
+ || reg == G0 || reg == O7)
+ break;
+
+ bfd_put_32 (input_bfd, INSN_NOP,
+ contents + rel->r_offset + 4);
+ }
+ break;
+ }
+ }
+ }
+ /* FALLTHROUGH */
+
default:
do_default:
r = _bfd_final_link_relocate (howto, input_bfd, input_section,
{
/* Assume this is a call protected by other code that
detect the symbol is undefined. If this is the case,
- we can safely ignore the overflow. If not, the
+ we can safely ignore the overflow. If not, the
program is hosed anyway, and a little warning isn't
going to help. */
break;
}
-
+
name = h->root.root.string;
}
else
asection *srela;
Elf_Internal_Rela rela;
- /* This symbol has an entry in the PLT. Set it up. */
+ /* This symbol has an entry in the PLT. Set it up. */
BFD_ASSERT (h->dynindx != -1);
rela.r_offset += (splt->output_section->vma + splt->output_offset);
rela.r_info = ELF64_R_INFO (h->dynindx, R_SPARC_JMP_SLOT);
+ /* Adjust for the first 4 reserved elements in the .plt section
+ when setting the offset in the .rela.plt section.
+ Sun forgot to read their own ABI and copied elf32-sparc behaviour,
+ thus .plt[4] has corresponding .rela.plt[0] and so on. */
+
bfd_elf64_swap_reloca_out (output_bfd, &rela,
((Elf64_External_Rela *) srela->contents
- + h->plt.offset));
+ + (h->plt.offset - 4)));
if ((h->elf_link_hash_flags & ELF_LINK_HASH_DEF_REGULAR) == 0)
{
/* Mark the symbol as undefined, rather than as defined in
the .plt section. Leave the value alone. */
sym->st_shndx = SHN_UNDEF;
+ /* If the symbol is weak, we do need to clear the value.
+ Otherwise, the PLT entry would provide a definition for
+ the symbol even if the symbol wasn't defined anywhere,
+ and so the symbol would never be NULL. */
+ if ((h->elf_link_hash_flags & ELF_LINK_HASH_REF_REGULAR_NONWEAK)
+ == 0)
+ sym->st_value = 0;
}
}
return true;
}
\f
-/* Functions for dealing with the e_flags field. */
+/* Functions for dealing with the e_flags field. */
+
+/* Copy backend specific data from one object module to another */
+static boolean
+sparc64_elf_copy_private_bfd_data (ibfd, obfd)
+ bfd *ibfd, *obfd;
+{
+ if ( bfd_get_flavour (ibfd) != bfd_target_elf_flavour
+ || bfd_get_flavour (obfd) != bfd_target_elf_flavour)
+ return true;
+
+ BFD_ASSERT (!elf_flags_init (obfd)
+ || (elf_elfheader (obfd)->e_flags
+ == elf_elfheader (ibfd)->e_flags));
+
+ elf_elfheader (obfd)->e_flags = elf_elfheader (ibfd)->e_flags;
+ elf_flags_init (obfd) = true;
+ return true;
+}
/* Merge backend specific data from an object file to the output
object file when linking. */
elf_flags_init (obfd) = true;
elf_elfheader (obfd)->e_flags = new_flags;
}
-
+
else if (new_flags == old_flags) /* Compatible flags are ok */
;
-
+
else /* Incompatible flags */
{
error = false;
-
+
+#define EF_SPARC_ISA_EXTENSIONS \
+ (EF_SPARC_SUN_US1 | EF_SPARC_SUN_US3 | EF_SPARC_HAL_R1)
+
if ((ibfd->flags & DYNAMIC) != 0)
{
/* We don't want dynamic objects memory ordering and
architecture to have any role. That's what dynamic linker
should do. */
- new_flags &= ~(EF_SPARCV9_MM | EF_SPARC_SUN_US1 | EF_SPARC_HAL_R1);
+ new_flags &= ~(EF_SPARCV9_MM | EF_SPARC_ISA_EXTENSIONS);
new_flags |= (old_flags
- & (EF_SPARCV9_MM
- | EF_SPARC_SUN_US1
- | EF_SPARC_HAL_R1));
+ & (EF_SPARCV9_MM | EF_SPARC_ISA_EXTENSIONS));
}
else
{
/* Choose the highest architecture requirements. */
- old_flags |= (new_flags & (EF_SPARC_SUN_US1 | EF_SPARC_HAL_R1));
- new_flags |= (old_flags & (EF_SPARC_SUN_US1 | EF_SPARC_HAL_R1));
- if ((old_flags & (EF_SPARC_SUN_US1 | EF_SPARC_HAL_R1))
- == (EF_SPARC_SUN_US1 | EF_SPARC_HAL_R1))
+ old_flags |= (new_flags & EF_SPARC_ISA_EXTENSIONS);
+ new_flags |= (old_flags & EF_SPARC_ISA_EXTENSIONS);
+ if ((old_flags & (EF_SPARC_SUN_US1 | EF_SPARC_SUN_US3))
+ && (old_flags & EF_SPARC_HAL_R1))
{
error = true;
(*_bfd_error_handler)
{
FILE *file = (FILE *) filep;
int reg, type;
-
+
if (ELF_ST_TYPE (((elf_symbol_type *) symbol)->internal_elf_sym.st_info)
!= STT_REGISTER)
return NULL;
bfd *abfd;
{
unsigned long mach = bfd_mach_sparc_v9;
-
- if (elf_elfheader (abfd)->e_flags & EF_SPARC_SUN_US1)
+
+ if (elf_elfheader (abfd)->e_flags & EF_SPARC_SUN_US3)
+ mach = bfd_mach_sparc_v9b;
+ else if (elf_elfheader (abfd)->e_flags & EF_SPARC_SUN_US1)
mach = bfd_mach_sparc_v9a;
return bfd_default_set_arch_mach (abfd, bfd_arch_sparc, mach);
}
/* internal relocations per external relocations.
For link purposes we use just 1 internal per
1 external, for assembly and slurp symbol table
- we use 2. */
+ we use 2. */
1,
64, /* arch_size */
8, /* file_align */
#define bfd_elf64_bfd_link_hash_table_create \
sparc64_elf_bfd_link_hash_table_create
-
+
#define elf_info_to_howto \
sparc64_elf_info_to_howto
#define bfd_elf64_get_reloc_upper_bound \
sparc64_elf_canonicalize_dynamic_reloc
#define bfd_elf64_bfd_reloc_type_lookup \
sparc64_elf_reloc_type_lookup
+#define bfd_elf64_bfd_relax_section \
+ sparc64_elf_relax_section
#define elf_backend_create_dynamic_sections \
_bfd_elf_create_dynamic_sections
sparc64_elf_print_symbol_all
#define elf_backend_output_arch_syms \
sparc64_elf_output_arch_syms
-
+#define bfd_elf64_bfd_copy_private_bfd_data \
+ sparc64_elf_copy_private_bfd_data
#define bfd_elf64_bfd_merge_private_bfd_data \
sparc64_elf_merge_private_bfd_data
projects / deliverable / binutils-gdb.git / blobdiff