espressif/binutils-gdb
1
0
Fork 0
mirror of https://github.com/espressif/binutils-gdb.git synced 2025-06-26 22:07:58 +08:00
Code Issues Packages Projects Releases Wiki Activity

Refactor Cortex-A8 erratum workaround in preparation

Browse Source 
2016-05-10  Thomas Preud'homme  <thomas.preudhomme@arm.com>

bfd/
	* elf32-arm.c (enum elf32_arm_stub_type): Delete
	arm_stub_a8_veneer_lwm enumerator.
	(arm_stub_a8_veneer_lwm): New unsigned constant to replace
	aforementioned enumerator.
	(struct elf32_arm_stub_hash_entry): Delete target_addend
	field and add source_value.
	(struct a8_erratum_fix): Delete addend field and add target_offset.
	(stub_hash_newfunc): Initialize source_value field amd remove
	initialization for target_addend.
	(arm_build_one_stub): Stop special casing Thumb relocations: promote
	the else to being always executed, moving the
	arm_stub_a8_veneer_b_cond specific code in it.  Remove
	stub_entry->target_addend from points_to computation.
	(cortex_a8_erratum_scan): Store in a8_erratum_fix structure the offset
	to target symbol from start of section rather than the offset from the
	stub address.
	(elf32_arm_size_stubs): Set stub_entry's source_value and target_value
	fields from struct a8_erratum_fix's offset and target_offset
	respectively.
	(make_branch_to_a8_stub): Rename target variable to loc.  Compute
	veneered_insn_loc and loc using stub_entry's source_value.
This commit is contained in:
Thomas Preud'homme
2016-05-10 15:28:10 +01:00
parent 15afaa63f3
commit 8d9d949014
2 changed files with 73 additions and 70 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
bfd/ChangeLog
View File

@ -1,3 +1,27 @@
2016-05-10 Thomas Preud'homme <thomas.preudhomme@arm.com>
* elf32-arm.c (enum elf32_arm_stub_type): Delete
arm_stub_a8_veneer_lwm enumerator.
(arm_stub_a8_veneer_lwm): New unsigned constant to replace
aforementioned enumerator.
(struct elf32_arm_stub_hash_entry): Delete target_addend
field and add source_value.
(struct a8_erratum_fix): Delete addend field and add target_offset.
(stub_hash_newfunc): Initialize source_value field amd remove
initialization for target_addend.
(arm_build_one_stub): Stop special casing Thumb relocations: promote
the else to being always executed, moving the
arm_stub_a8_veneer_b_cond specific code in it. Remove
stub_entry->target_addend from points_to computation.
(cortex_a8_erratum_scan): Store in a8_erratum_fix structure the offset
to target symbol from start of section rather than the offset from the
stub address.
(elf32_arm_size_stubs): Set stub_entry's source_value and target_value
fields from struct a8_erratum_fix's offset and target_offset
respectively.
(make_branch_to_a8_stub): Rename target variable to loc. Compute
veneered_insn_loc and loc using stub_entry's source_value.
2016-05-10 Thomas Preud'homme <thomas.preudhomme@arm.com> 2016-05-10 Thomas Preud'homme <thomas.preudhomme@arm.com>
(elf32_arm_merge_eabi_attributes): Add merging logic for (elf32_arm_merge_eabi_attributes): Add merging logic for

119
bfd/elf32-arm.c
View File

@ -2632,11 +2632,12 @@ enum elf32_arm_stub_type
{ {
arm_stub_none, arm_stub_none,
DEF_STUBS DEF_STUBS
/* Note the first a8_veneer type. */
arm_stub_a8_veneer_lwm = arm_stub_a8_veneer_b_cond
}; };
#undef DEF_STUB #undef DEF_STUB
/* Note the first a8_veneer type. */
const unsigned arm_stub_a8_veneer_lwm = arm_stub_a8_veneer_b_cond;
typedef struct typedef struct
{ {
const insn_sequence* template_sequence; const insn_sequence* template_sequence;
@ -2666,8 +2667,12 @@ struct elf32_arm_stub_hash_entry
bfd_vma target_value; bfd_vma target_value;
asection *target_section; asection *target_section;
/* Offset to apply to relocation referencing target_value. */ /* Same as above but for the source of the branch to the stub. Used for
bfd_vma target_addend; Cortex-A8 erratum workaround to patch it to branch to the stub. As
such, source section does not need to be recorded since Cortex-A8 erratum
workaround stubs are only generated when both source and target are in the
same section. */
bfd_vma source_value;
/* The instruction which caused this stub to be generated (only valid for /* The instruction which caused this stub to be generated (only valid for
Cortex-A8 erratum workaround stubs at present). */ Cortex-A8 erratum workaround stubs at present). */
@ -2836,7 +2841,7 @@ struct a8_erratum_fix
bfd *input_bfd; bfd *input_bfd;
asection *section; asection *section;
bfd_vma offset; bfd_vma offset;
bfd_vma addend; bfd_vma target_offset;
unsigned long orig_insn; unsigned long orig_insn;
char *stub_name; char *stub_name;
enum elf32_arm_stub_type stub_type; enum elf32_arm_stub_type stub_type;
@ -3413,9 +3418,9 @@ stub_hash_newfunc (struct bfd_hash_entry *entry,
eh = (struct elf32_arm_stub_hash_entry *) entry; eh = (struct elf32_arm_stub_hash_entry *) entry;
eh->stub_sec = NULL; eh->stub_sec = NULL;
eh->stub_offset = 0; eh->stub_offset = 0;
eh->source_value = 0;
eh->target_value = 0; eh->target_value = 0;
eh->target_section = NULL; eh->target_section = NULL;
eh->target_addend = 0;
eh->orig_insn = 0; eh->orig_insn = 0;
eh->stub_type = arm_stub_none; eh->stub_type = arm_stub_none;
eh->stub_size = 0; eh->stub_size = 0;
@ -4450,65 +4455,36 @@ arm_build_one_stub (struct bfd_hash_entry *gen_entry,
BFD_ASSERT (nrelocs != 0 && nrelocs <= MAXRELOCS); BFD_ASSERT (nrelocs != 0 && nrelocs <= MAXRELOCS);
for (i = 0; i < nrelocs; i++) for (i = 0; i < nrelocs; i++)
if (template_sequence[stub_reloc_idx[i]].r_type == R_ARM_THM_JUMP24 {
|| template_sequence[stub_reloc_idx[i]].r_type == R_ARM_THM_JUMP19 Elf_Internal_Rela rel;
|| template_sequence[stub_reloc_idx[i]].r_type == R_ARM_THM_CALL bfd_boolean unresolved_reloc;
|| template_sequence[stub_reloc_idx[i]].r_type == R_ARM_THM_XPC22) char *error_message;
{ bfd_vma points_to =
Elf_Internal_Rela rel; sym_value + template_sequence[stub_reloc_idx[i]].reloc_addend;
bfd_boolean unresolved_reloc;
char *error_message;
enum arm_st_branch_type branch_type
= (template_sequence[stub_reloc_idx[i]].r_type != R_ARM_THM_XPC22
? ST_BRANCH_TO_THUMB : ST_BRANCH_TO_ARM);
bfd_vma points_to = sym_value + stub_entry->target_addend;
rel.r_offset = stub_entry->stub_offset + stub_reloc_offset[i]; rel.r_offset = stub_entry->stub_offset + stub_reloc_offset[i];
rel.r_info = ELF32_R_INFO (0, rel.r_info = ELF32_R_INFO (0,
template_sequence[stub_reloc_idx[i]].r_type); template_sequence[stub_reloc_idx[i]].r_type);
rel.r_addend = template_sequence[stub_reloc_idx[i]].reloc_addend; rel.r_addend = 0;
if (stub_entry->stub_type == arm_stub_a8_veneer_b_cond && i == 0) if (stub_entry->stub_type == arm_stub_a8_veneer_b_cond && i == 0)
/* The first relocation in the elf32_arm_stub_a8_veneer_b_cond[] /* The first relocation in the elf32_arm_stub_a8_veneer_b_cond[]
template should refer back to the instruction after the original template should refer back to the instruction after the original
branch. */ branch. We use target_section as Cortex-A8 erratum workaround stubs
points_to = sym_value; are only generated when both source and target are in the same
section. */
points_to = stub_entry->target_section->output_section->vma
+ stub_entry->target_section->output_offset
+ stub_entry->source_value;
/* There may be unintended consequences if this is not true. */ elf32_arm_final_link_relocate (elf32_arm_howto_from_type
BFD_ASSERT (stub_entry->h == NULL); (template_sequence[stub_reloc_idx[i]].r_type),
stub_bfd, info->output_bfd, stub_sec, stub_sec->contents, &rel,
/* Note: _bfd_final_link_relocate doesn't handle these relocations points_to, info, stub_entry->target_section, "", STT_FUNC,
properly. We should probably use this function unconditionally, stub_entry->branch_type,
rather than only for certain relocations listed in the enclosing (struct elf_link_hash_entry *) stub_entry->h, &unresolved_reloc,
conditional, for the sake of consistency. */ &error_message);
elf32_arm_final_link_relocate (elf32_arm_howto_from_type }
(template_sequence[stub_reloc_idx[i]].r_type),
stub_bfd, info->output_bfd, stub_sec, stub_sec->contents, &rel,
points_to, info, stub_entry->target_section, "", STT_FUNC,
branch_type, (struct elf_link_hash_entry *) stub_entry->h,
&unresolved_reloc, &error_message);
}
else
{
Elf_Internal_Rela rel;
bfd_boolean unresolved_reloc;
char *error_message;
bfd_vma points_to = sym_value + stub_entry->target_addend
+ template_sequence[stub_reloc_idx[i]].reloc_addend;
rel.r_offset = stub_entry->stub_offset + stub_reloc_offset[i];
rel.r_info = ELF32_R_INFO (0,
template_sequence[stub_reloc_idx[i]].r_type);
rel.r_addend = 0;
elf32_arm_final_link_relocate (elf32_arm_howto_from_type
(template_sequence[stub_reloc_idx[i]].r_type),
stub_bfd, info->output_bfd, stub_sec, stub_sec->contents, &rel,
points_to, info, stub_entry->target_section, "", STT_FUNC,
stub_entry->branch_type,
(struct elf_link_hash_entry *) stub_entry->h, &unresolved_reloc,
&error_message);
}
return TRUE; return TRUE;
#undef MAXRELOCS #undef MAXRELOCS
@ -5101,7 +5077,8 @@ cortex_a8_erratum_scan (bfd *input_bfd,
a8_fixes[num_a8_fixes].input_bfd = input_bfd; a8_fixes[num_a8_fixes].input_bfd = input_bfd;
a8_fixes[num_a8_fixes].section = section; a8_fixes[num_a8_fixes].section = section;
a8_fixes[num_a8_fixes].offset = i; a8_fixes[num_a8_fixes].offset = i;
a8_fixes[num_a8_fixes].addend = offset; a8_fixes[num_a8_fixes].target_offset =
target - base_vma;
a8_fixes[num_a8_fixes].orig_insn = insn; a8_fixes[num_a8_fixes].orig_insn = insn;
a8_fixes[num_a8_fixes].stub_name = stub_name; a8_fixes[num_a8_fixes].stub_name = stub_name;
a8_fixes[num_a8_fixes].stub_type = stub_type; a8_fixes[num_a8_fixes].stub_type = stub_type;
@ -5680,9 +5657,9 @@ elf32_arm_size_stubs (bfd *output_bfd,
stub_entry->stub_offset = 0; stub_entry->stub_offset = 0;
stub_entry->id_sec = link_sec; stub_entry->id_sec = link_sec;
stub_entry->stub_type = a8_fixes[i].stub_type; stub_entry->stub_type = a8_fixes[i].stub_type;
stub_entry->source_value = a8_fixes[i].offset;
stub_entry->target_section = a8_fixes[i].section; stub_entry->target_section = a8_fixes[i].section;
stub_entry->target_value = a8_fixes[i].offset; stub_entry->target_value = a8_fixes[i].target_offset;
stub_entry->target_addend = a8_fixes[i].addend;
stub_entry->orig_insn = a8_fixes[i].orig_insn; stub_entry->orig_insn = a8_fixes[i].orig_insn;
stub_entry->branch_type = a8_fixes[i].branch_type; stub_entry->branch_type = a8_fixes[i].branch_type;
@ -16277,7 +16254,7 @@ make_branch_to_a8_stub (struct bfd_hash_entry *gen_entry,
bfd_vma veneered_insn_loc, veneer_entry_loc; bfd_vma veneered_insn_loc, veneer_entry_loc;
bfd_signed_vma branch_offset; bfd_signed_vma branch_offset;
bfd *abfd; bfd *abfd;
unsigned int target; unsigned int loc;
stub_entry = (struct elf32_arm_stub_hash_entry *) gen_entry; stub_entry = (struct elf32_arm_stub_hash_entry *) gen_entry;
data = (struct a8_branch_to_stub_data *) in_arg; data = (struct a8_branch_to_stub_data *) in_arg;
@ -16288,9 +16265,11 @@ make_branch_to_a8_stub (struct bfd_hash_entry *gen_entry,
contents = data->contents; contents = data->contents;
/* We use target_section as Cortex-A8 erratum workaround stubs are only
generated when both source and target are in the same section. */
veneered_insn_loc = stub_entry->target_section->output_section->vma veneered_insn_loc = stub_entry->target_section->output_section->vma
+ stub_entry->target_section->output_offset + stub_entry->target_section->output_offset
+ stub_entry->target_value; + stub_entry->source_value;
veneer_entry_loc = stub_entry->stub_sec->output_section->vma veneer_entry_loc = stub_entry->stub_sec->output_section->vma
+ stub_entry->stub_sec->output_offset + stub_entry->stub_sec->output_offset
@ -16302,7 +16281,7 @@ make_branch_to_a8_stub (struct bfd_hash_entry *gen_entry,
branch_offset = veneer_entry_loc - veneered_insn_loc - 4; branch_offset = veneer_entry_loc - veneered_insn_loc - 4;
abfd = stub_entry->target_section->owner; abfd = stub_entry->target_section->owner;
target = stub_entry->target_value; loc = stub_entry->source_value;
/* We attempt to avoid this condition by setting stubs_always_after_branch /* We attempt to avoid this condition by setting stubs_always_after_branch
in elf32_arm_size_stubs if we've enabled the Cortex-A8 erratum workaround. in elf32_arm_size_stubs if we've enabled the Cortex-A8 erratum workaround.
@ -16363,8 +16342,8 @@ make_branch_to_a8_stub (struct bfd_hash_entry *gen_entry,
return FALSE; return FALSE;
} }
bfd_put_16 (abfd, (branch_insn >> 16) & 0xffff, &contents[target]); bfd_put_16 (abfd, (branch_insn >> 16) & 0xffff, &contents[loc]);
bfd_put_16 (abfd, branch_insn & 0xffff, &contents[target + 2]); bfd_put_16 (abfd, branch_insn & 0xffff, &contents[loc + 2]);
return TRUE; return TRUE;
} }