Given a linker script fragment like this:
SECTIONS {
. = 0x1000;
.text : AT(0x100) { *(.text) }
.data : AT(0x200) { *(.data) }
.rodata : AT(0x300) { *(.rodata) }
}
and an input file containing sections, '.text', '.data.1', and
'.rodata', then we'd expect the linker to place '.text' and '.rodata' in
the obvious way, and the '.data.1' orphan section would be located after
the '.data' section (assuming similar section properties).
Further, I believe that the expectation would be that the LMA for the
orphan '.data.1' section would start from 0x200 (as there is no '.data'
content).
However, right now, the LMA for '.data.1' would be 0x101, following on
from the '.text' section, this is because the change in LMA for the
'.data' section is not noticed by the linker, if there's no content in
the '.data' section.
What can be even more confusing to a user (though the cause is obvious
once you understand what's going on) is that adding some content to
'.data' will cause the orphan '.data.1' to switch to an LMA based off of
0x200.
This commit changes the behaviour so that an empty section that is in
the default lma region, and sets its lma, will adjust the lma of the
default region, this change will then be reflected in following sections
within the default lma memory region.
There's a new test to cover this issue that passes on a range of
targets, however, some targets generate additional sections, or have
stricter memory region size requirements that make it harder to come
up with a generic pass pattern, that still tests the required
features. For now I've set the test to ignore these targets.
ld/ChangeLog:
* ldlang.c (lang_size_sections_1): Shortcut loop only after
tracking changes to the default regions LMA.
* testsuite/ld-elf/orphan-9.ld: Extend header comment.
* testsuite/ld-elf/orphan-10.d: New file.
* testsuite/ld-elf/orphan-10.s: New file.
* NEWS: Mention change in behaviour.
When picking an lma_region for an orphan section we currently create a
new lang_output_section_statement_type and then populate this with the
orphan section.
The problem is that the lang_output_section_statement_type has a prev
pointer that links back to the previous output section. For non-orphan
output sections, that are created in linker script order, the prev
pointer will point to the output section that appears previous in linker
script order, as you'd probably expect.
The problem is that orphan sections are placed after processing the
linker script, and so, in the case of an output section created for an
orphan input section, the prev pointer actually points to the last
output section created.
This causes some unexpected behaviour when the orphan section is not
placed after the last non-orphan section that was created.
For example, consider this linker script:
MEMORY {
TEXT : ORIGIN = 0x200, LENGTH = 0x10
RODATA : ORIGIN = 0x400, LENGTH = 0x10
}
SECTIONS {
.text : {*(.text) } AT>TEXT
.data : AT(0x300) { *(.data) }
.rodata : { *(.rodata) } AT>RODATA
}
If we are processing an orphan section '.data.1' and decide to place
this after '.data', then the output section created will have a prev
pointer that references the '.rodata' output section. The result of
this is that '.data.1' will actually be assigned to the RODATA lma
region, which is probably not the expected behaviour.
The reason why '.data.1' is placed into the lma region of the '.rodata'
section is that lma region propagation is done at the time we create the
output section, based on the previous output section pointer, which is
really just a last-output-section-created pointer at that point in time,
though the prev point is fixed up later to reflect the true order of the
output sections.
The solution I propose in this commit is to move the propagation of lma
regions into a separate pass of the linker, rather than performing this
as part of the enter/exit of output sections during linker script
parsing.
During this later phase we have all of the output sections to hand, and
the prev/next points have been fixed up by this point to reflect the
actual placement ordering.
There's a new test to cover this issue that passes on a range of
targets, however, some targets generate additional sections, or have
stricter memory region size requirements that make it harder to come
up with a generic pass pattern, that still tests the required
features. For now I've set the test to ignore these targets.
ld/ChangeLog:
* ldlang.c (lang_leave_output_section_statement): Move lma_region
logic to...
(lang_propagate_lma_regions): ...this new function.
(lang_process): Call new function.
* testsuite/ld-elf/orphan-9.d: New file.
* testsuite/ld-elf/orphan-9.ld: New file.
* testsuite/ld-elf/orphan-9.s: New file.
* NEWS: Mention change in behaviour.
lang_assignment_statement serves both assignments and asserts.
* ldlang.c (open_input_bfds): Check that lang_assignment_statement
is not an assert before referencing defsym.
Currently the EXCLUDE_FILE linker script construct can only be used
within the input section list, and applied only to the section pattern
immediately following the EXCLUDE_FILE. For example:
*.o (EXCLUDE_FILE (a.o) .text .rodata)
In this case all sections matching '.text' are included from all files
matching '*.o' but not from the file 'a.o'. All sections matching
'.rodata' are also included from all files matching '*.o' (incluing from
'a.o').
If the user wants to restrict the inclusion of section '.rodata' so that
this too is not taken from the file 'a.o' then the above example must be
extended like this:
*.o (EXCLUDE_FILE (a.o) .text EXCLUDE_FILE (a.o) .rodata)
However, due to the internal grammar of the linker script language the
snippet 'EXCLUDE_FILE (a.o) .text' is parsed by a pattern called
'wildcard_spec'. The same 'wildcard_spec' pattern is also used to parse
the input file name snippet '*.o' in the above examples. As a result of
this pattern reuse within the linker script grammar then the following
is also a valid linker script construct:
EXCLUDE_FILE (a.o) *.o (.text .rodata)
However, though the linker accepts this without complaint the
EXCLUDE_FILE part is silently ignored and has no effect.
This commit takes this last example and makes it a useful, valid,
construct. The last example now means to include sections '.text' and
'.rodata' from all files matching '*.o' except for the file 'a.o'.
If the list of input sections is long, and the user knows that the file
exclusion applies across the list then the second form might be a
clearer alternative to replicating the EXCLUDE_FILE construct.
I've added a set of tests for EXCLUDE_FILE to the linker, including
tests for the new functionality.
ld/ChangeLog:
* ldlang.h (struct lang_wild_statement_struct): Add
exclude_name_list field.
* ldlang.c (walk_wild_file_in_exclude_list): New function.
(walk_wild_consider_section): Use new
walk_wild_file_in_exclude_list function.
(walk_wild_file): Add call to walk_wild_file_in_exclude_list.
(print_wild_statement): Print new exclude_name_list field.
(lang_add_wild): Initialise new exclude_name_list field.
* testsuite/ld-scripts/exclude-file-1.d: New file.
* testsuite/ld-scripts/exclude-file-1.map: New file.
* testsuite/ld-scripts/exclude-file-1.t: New file.
* testsuite/ld-scripts/exclude-file-2.d: New file.
* testsuite/ld-scripts/exclude-file-2.map: New file.
* testsuite/ld-scripts/exclude-file-2.t: New file.
* testsuite/ld-scripts/exclude-file-3.d: New file.
* testsuite/ld-scripts/exclude-file-3.map: New file.
* testsuite/ld-scripts/exclude-file-3.t: New file.
* testsuite/ld-scripts/exclude-file-4.d: New file.
* testsuite/ld-scripts/exclude-file-4.map: New file.
* testsuite/ld-scripts/exclude-file-4.t: New file.
* testsuite/ld-scripts/exclude-file-a.s: New file.
* testsuite/ld-scripts/exclude-file-b.s: New file.
* testsuite/ld-scripts/exclude-file.exp: New file.
* ld.texinfo (Input Section Basics): Update description of
EXCLUDE_FILE to cover the new features.
* NEWS: Mention new EXCLUDE_FILE usage.
See https://sourceware.org/ml/binutils/2016-07/msg00091.html
This patch stop --gc-sections elf_gc_sweep_symbol localizing symbols
that ought to remain global.
The difficulty with always descending into output section statements
is that symbols defined by the script in such statements don't have
a bfd section when lang_do_assignments runs early in the link process.
There are two approaches to curing this problem. Either we can
create the bfd section early, or we can use a special section. This
patch takes the latter approach and uses bfd_und_section. (Creating
bfd sections early results in changed output section order, and thus
lots of testsuite failures. You can't create all output sections
early to ensure proper ordering as KEEP then stops empty sections
from being stripped.)
The wrinkle with this approach is that some code that runs at
gc-sections time needs to be made aware of the odd defined symbols
using bfd_und_section.
bfd/
* elf64-x86-64.c (elf_x86_64_convert_load_reloc): Handle symbols
defined temporarily with bfd_und_section.
* elflink.c (_bfd_elf_gc_keep): Don't set SEC_KEEP for bfd_und_section.
* elfxx-mips.c (mips_elf_local_pic_function_p): Exclude defined
symbols with bfd_und_section.
ld/
* ldlang.c (lang_do_assignments_1): Descend into output section
statements that do not yet have bfd sections. Set symbol section
temporarily for symbols defined in such statements to the undefined
section. Don't error on data or reloc statements until final phase.
* ldexp.c (exp_fold_tree_1 <etree_assign>): Handle bfd_und_section
in expld.section.
* testsuite/ld-mmix/bpo-10.d: Adjust.
* testsuite/ld-mmix/bpo-11.d: Adjust.
It was like printf, which means you can't use bfd_set_error_handler to
hook in a function to do something and then call the original handler.
The patch also deletes some unused functions and makes pointers local.
bfd/
* bfd-in.h: Include stdarg.h.
* bfd.c (bfd_error_handler_type): Make like vprintf.
(_bfd_error_internal): Rename from _bfd_error_handler. Make static.
(error_handler_internal): New function, split out from..
(_bfd_default_error_handler): ..here. Rename to _bfd_error_handler.
(bfd_set_error_handler): Update.
(bfd_get_error_handler, bfd_get_assert_handler): Delete.
(_bfd_assert_handler): Make static.
* coffgen.c (null_error_handler): Update params.
* elf-bfd.h (struct elf_backend_data <link_order_error_handler>):
Don't use bfd_error_handler_type.
* elf64-mmix.c (mmix_dump_bpo_gregs): Likewise.
* elfxx-target.h (elf_backend_link_order_error_handler): Default
to _bfd_error_handler.
* libbfd-in.h (_bfd_default_error_handler): Don't declare.
(bfd_assert_handler_type): Likewise.
(_bfd_error_handler): Update.
* bfd-in2.h: Regenerate.
* libbfd.h: Regenerate.
ld/
* ldlang.c (ignore_bfd_errors): Update params.
Not much to see here, just renaming a function.
bfd/
* targets.c (bfd_seach_for_target): Rename to..
(bfd_iterate_over_targets): ..this. Rewrite doc.
* bfd-in2.h: Regenerate.
ld/
* ldlang.c (open_output): Replace bfd_search_for_target with
bfd_iterate_over_targets. Localize vars.
Many more places use abfd->my_archive rather than bfd_my_archive (abfd),
so let's make the code consistently use the first idiom.
bfd/
* bfd-in.h (bfd_my_archive): Delete.
* bfd-in2.h: Regenerate.
binutils/
* ar.c: Expand uses of bfd_my_archive.
* size.c: Likewise.
ld/
* ldlang.c: Expand uses of bfd_my_archive.
* ldmain.c: Likewise.
* ldmisc.c: Likewise.
* plugin.c: Likewise.
Move ELF relocation check after lang_gc_sections so that all the
reference counting code for plt and got relocs can be removed. This
only affects ELF targets which check relocations after opening all
input file.
* ldlang.c (lang_check_relocs): New function.
(lang_process): Call lang_check_relocs after lang_gc_sections.
* emultempl/elf32.em (gld${EMULATION_NAME}_before_parse): Don't
call _bfd_elf_link_check_relocs here.
NOCROSSREFS_TO is similar to the existing NOCROSSREFS command but only
checks one direction of cross referencing.
ld/ChangeLog
* ld.texinfo: Document NOCROSSREFS_TO script command.
* ldlang.h (struct lang_nocrossrefs): Add onlyfirst field.
(lang_add_nocrossref_to): New prototype.
* ldcref.c (check_local_sym_xref): Use onlyfirst to only look for
symbols defined in the first section.
(check_nocrossref): Likewise.
* ldgram.y (NOCROSSREFS_TO): New script command.
* ldlang.c (lang_add_nocrossref): Set onlyfirst to FALSE.
(lang_add_nocrossref_to): New function.
* ldlex.l (NOCROSSREFS_TO): New token.
* NEWS: Mention NOCROSSREFS_TO.
* testsuite/ld-scripts/cross4.t: New file.
* testsuite/ld-scripts/cross5.t: Likewise.
* testsuite/ld-scripts/cross6.t: Likewise.
* testsuite/ld-scripts/cross7.t: Likewise.
* testsuite/ld-scripts/crossref.exp: Run 4 new NOCROSSREFS_TO
tests.
* ldlang.c (print_output_section_statement): Show minfo size
in target machine address units.
(print_reloc_statement): Likewise.
(print_padding_statement): Likewise.
(print_data_statement): Likewise. Ensure minimum print_dot
increment of one address unit.
and one extraneous occurrence.
* ldlang.c (TO_ADDR, TO_SIZE, opb_shift): Move earlier in file.
(lang_insert_orphan): Use TO_ADDR in __stop sym calculation.
(print_input_section): Don't use TO_ADDR when printing section
size.
(lang_size_sections_1): Use TO_ADDR in overlay lma calculation.
(lang_size_sections): Use TO_ADDR in relro end calculation.
PR 18452
* ldlang.c (maybe_overlays): New static var.
(lang_size_sections_1): Set it here.
(struct check_sec): New.
(sort_sections_by_lma): Adjust for array of structs.
(sort_sections_by_vma): New function.
(lang_check_section_addresses): Check both LMA and VMA for overlap.
* testsuite/ld-scripts/rgn-over7.d: Adjust.
Some places tested SEC_LOAD, others SEC_HAS_CONTENTS.
* ldlang.c (IS_TBSS): New macro, extracted from..
(IGNORE_SECTION): ..here.
(lang_size_sections_1): Use IS_TBSS and IGNORE_SECTION.
(lang_size_sections, lang_do_assignments_1): Use IS_TBSS.
PR ld/19803
* ldlang.c (lang_add_gc_name): New function. Adds the provided
symbol name to the list of gc symbols.
(lang_process): Call lang_add_gc_name with entry_symbol_default if
entry_symbol.name is NULL. Use lang_add_gc_name to add the init
and fini function names.
* pe-dll.c (process_def_file_and_drectve): Add exported names to
the gc symbol list.
* testsuite/ld-pe/pr19803.s: Do not export _testval symbol.
* testsuite/ld-pe/pr19803.d: Tweak expected output.
Another option might be to not bump "dot" for .tbss alignment in the
main section sizing loop, but that could leak some of the following
section into the TLS segment. Leakage shouldn't matter since it will
be to bytes past the end of .tdata, but for now this is a safer
option.
PR ld/19264
* ldlang.c (lang_size_sections): Don't ignore .tbss when
adjusting start of relro region.
The last patch missed handling the case where the ideal place to put
an orphan was after a non-existent output section statement, as can
happen when not using the builtin linker scripts. This patch uses the
updated flags for that case too, and extends the support to mmo and pe.
PR ld/19162
* emultempl/elf32.em (gld${EMULATION_NAME}_place_orphan): Pass
updated flags to lang_output_section_find_by_flags.
* emultempl/mmo.em (mmo_place_orphan): Merge flags for any
other input sections that might match a new output section to
decide placement.
* emultempl/pe.em (gld_${EMULATION_NAME}_place_orphan): Likewise.
* emultempl/pep.em (gld_${EMULATION_NAME}_place_orphan): Likewise.
* ldlang.c (lang_output_section_find_by_flags): Add sec_flags param.
* ldlang.h (lang_output_section_find_by_flags): Update prototype.
Makes these symbols defined before bfd_elf_size_dynamic_sections, to
avoid horrible hacks elsewhere. The exp_fold_tree undefweak change
is necessary to define undefweak symbols early too. The comment was
wrong. PROVIDE in fact defines undefweak symbols, via
bfd_elf_record_link_assignment.
PR ld/19175
* ldlang.c (lang_insert_orphan): Evaluate __start_* and __stop_*
symbol PROVIDE expressions.
* ldexp.c (exp_fold_tree_1 <etree_provide>): Define undefweak
references.
Giving linker script symbols defined outside of output sections a
section-relative value early, leads to them being used in expressions
as if they were defined inside an output section. This can mean loss
of the section VMA, and wrong results.
ld/
PR ld/18963
* ldexp.h (struct ldexp_control): Add rel_from_abs.
(ldexp_finalize_syms): Declare.
* ldexp.c (new_rel_from_abs): Keep absolute for expressions
outside of output section statements. Set rel_from_abs.
(make_abs, exp_fold_tree, exp_fold_tree_no_dot): Clear rel_from_abs.
(struct definedness_hash_entry): Add final_sec, and comment.
(update_definedness): Set final_sec.
(set_sym_sections, ldexp_finalize_syms): New functions.
* ldlang.c (lang_process): Call ldexp_finalize_syms.
ld/testsuite
PR ld/18963
* ld-scripts/pr18963.d,
* ld-scripts/pr18963.t: New test.
* ld-scripts/expr.exp: Run it.
* ld-elf/provide-hidden-2.ld: Explicitly make "dot" absolute.
* ld-mips-elf/gp-hidden.sd: Don't care about _gp section.
* ld-mips-elf/no-shared-1-n32.d: Don't care about symbol shown at
start of .data section.
* ld-mips-elf/no-shared-1-n64.d: Likewise.
* ld-mips-elf/no-shared-1-o32.d: Likewise.
Replace the options --warn-orphan and --no-warn-orphan with a single
option --orphan-handling=MODE, where mode can be place, warn, error, and
discard.
Mode 'place' is the default, and is the current behaviour, placing the
orphan section into a suitable output section.
Mode 'warn' is the same as '--warn-orphan'. The orphan is also placed
using the same algorithm as for 'place'.
Mode 'error' is the same as '--warn-orphan' and '--fatal-warnings'.
Mode 'discard' assigns all output sections to the /DISCARD/ section.
ld/ChangeLog:
* ld.h (enum orphan_handling_enum): New.
(ld_config_type): Remove warn_orphan, add orphan_handling.
* ldemul.c (ldemul_place_orphan): Remove warning about orphan
sections.
* ldlang.c (ldlang_place_orphan): New function.
(lang_place_orphans): Call ldlang_place_orphan.
* ldlex.h (enum option_values): Remove OPTION_WARN_ORPHAN and
OPTION_NO_WARN_ORPHAN, add OPTION_ORPHAN_HANDLING.
* lexsup.c (ld_options): Remove 'warn-orphan' and
'no-warn-orphan', add 'orphan-handling'.
(parse_args): Remove handling for OPTION_WARN_ORPHAN and
OPTION_NO_WARN_ORPHAN, add handling for OPTION_ORPHAN_HANDLING.
* NEWS: Replace text about --warn-orphan with --orphan-handling.
* ld.texinfo (Options): Remove --warn-orphan entry and add
entry on --orphan-handling.
(Orphan Sections): Add reference to relevant command line options.
ld/testsuite/ChangeLog:
* ld-elf/elf.exp: Switch to rely on run_dump_test.
* ld-elf/orphan-5.l: Update expected output.
* ld-elf/orphan-5.d: New file.
* ld-elf/orphan-6.d: New file.
* ld-elf/orphan-6.l: New file.
* ld-elf/orphan-7.d: New file.
* ld-elf/orphan-7.map: New file.
* ld-elf/orphan-8.d: New file.
* ld-elf/orphan-8.map: New file.
Add a new command line option '--require-defined' to the linker. This
option operates identically to the '--undefined' option, except that if
the symbol is not defined in the final output file then the linker will
exit with an error.
When making use of --gc-section, or just when trying to pull in parts of
a library, it is not uncommon for a user to use the '--undefined'
command line option to specify a symbol that the user then expects to be
defined by one of the object files supplied to the link.
However, if for any reason the symbol is not satisfied by an object
provided to the link the user will be left with an undefined symbol in
the output file, instead of a defined symbol.
In some cases the above behaviour is what the user wants, in other cases
though we can do better. The '--require-defined' option tries to fill
this gap. The symbol passed to the '--require-defined' option is
treated exactly as if the symbol was passed to '--undefined', however,
before the linker exits a check is made that all symbols passed to
'--require-defined' are actually defined, if any are not then the link
will fail with an error.
ld/ChangeLog:
* ld.texinfo (Options): Document --require-defined option.
* ldlang.c (struct require_defined_symbol): New structure.
(require_defined_symbol_list): New variable.
(ldlang_add_require_defined): New function.
(ldlang_check_require_defined_symbols): New function.
(lang_process): Check required symbols are defined.
* ldlang.h (ldlang_add_require_defined): Declare.
* ldlex.h (enum option_values): Add OPTION_REQUIRE_DEFINED_SYMBOL.
* lexsup.c (ld_options): Add '--require-defined' entry.
(parse_args): Handle '--require-defined' entry.
* NEWS: Mention new '--require-defined' option.
ld/testsuite/ChangeLog:
* ld-undefined/require-defined-1.d: New file.
* ld-undefined/require-defined-2.d: New file.
* ld-undefined/require-defined-3.d: New file.
* ld-undefined/require-defined-4.d: New file.
* ld-undefined/require-defined-5.d: New file.
* ld-undefined/require-defined.exp: New file.
* ld-undefined/require-defined.s: New file.
The linker tries to put the end of the last section in the relro
segment exactly on a page boundary, because the relro segment itself
must end on a page boundary. If for any reason this can't be done,
padding is inserted. Since the end of the relro segment is typically
between .got and .got.plt, padding effectively increases the size of
the GOT. This isn't nice for targets and code models with limited GOT
addressing.
The problem with the current code is that it doesn't cope very well
with aligned sections in the relro segment. When making .got aligned
to a 256 byte boundary for PowerPC64, I found that often the initial
alignment attempt failed and the fallback attempt to be less than
adequate. This is a particular problem for PowerPC64 since the
distance between .got and .plt affects the size of plt call stubs,
leading to "stubs don't match calculated size" errors.
So this rewrite takes a direct approach to calculating a new relro
base. Starting from the last section in the segment, we calculate
where it must start to position its end on the boundary, or as near as
possible considering alignment requirements. The new start then
becomes the goal for the previous section to end, and so on for all
sections. This of course ignores the possibility that user scripts
will place . = ALIGN(xxx); in the relro segment, or provide section
address expressions. In those cases we might fail, but the old code
probably did too, and a fallback is provided.
ld/
* ldexp.h (struct ldexp_control): Delete dataseg.min_base. Add
data_seg.relro_offset.
* ldexp.c (fold_binary <DATA_SEGMENT_ALIGN>): Don't set min_base.
(fold_binary <DATA_SEGMENT_RELRO_END>): Do set relro_offset.
* ldlang.c (lang_size_sections): Rewrite code adjusting relro
segment base to line up last section on page boundary.
ld/testsuite/
* ld-x86-64/pr18176.d: Update.
gold marks _init and _fini via symbol, ld marks them via section
(default scripts set .init and .fini section KEEP). This makes it
possible for people to write their own _init and not bother to put the
function into the right section.
PR ld/18223
* ldlang.c (lang_process): Add _init and _fini to gc_sym_list.
Adjusting the start of the relro segment in order to make it end
exactly on a page boundary runs into difficulties when sections in the
relro segment are aligned; Adjusting the start by (next_page - end)
sometimes results in more than that adjustment occurring at the end,
overrunning the page boundary. So when that occurs we try a new lower
start position by masking the adjusted start with the maximum section
alignment. However, we didn't consider that this masked start address
may in fact be before the initial relro base, which is silly since
that can only increase padding at the relro end.
I've also moved some calculations closer to where they are used, and
comments closer to the relevant statements.
* ldlang.c (lang_size_sections): When alignment of sections
results in relro base adjustment being too large, don't go lower
than the initial value.
* ldexp.c (fold_binary <DATA_SEGMENT_RELRO_END>): Comment.
* scripttempl/elf.sc (DATA_SEGMENT_ALIGN): Omit SEGMENT_SIZE
alignment when SEGMENT_SIZE is the same as MAXPAGESIZE.
LTO output objects have an STT_FILE symbol using the name of the file,
a temporary file. This results in executables that can't be exactly
reproduced, so the file name needs to be dropped. We don't want to
lose all file symbols when linking a mix of lto and non-lto objects as
a file symbol can be used to figure which source file generated a
given local symbol. So lto output objects need to be marked.
I chose to mark lto output objects with a new bfd flag. This flag is
also used to fix a bug in the link-once handling; An object being
loaded after "loading_lto_outputs" is set might be one extracted from
an archive to satisfy new references from lto objects, not an lto
object itself.
The new flag is copied from archive to elements, and the same done
for no_export. This fixes a bug in that --exclude-libs doesn't work
with thin archives. I'm not completely happy with this part of the
patch and may revist this to avoid the hack in
_bfd_look_for_bfd_in_cache.
PR ld/17973
include/
* bfdlink.h (struct bfd_link_info): Delete loading_lto_outputs.
bfd/
* bfd.c (struct bfd): Add lto_output.
* linker.c (_bfd_handle_already_linked): Explicitly test for
objects added by the lto plugin.
* opncls.c (_bfd_new_bfd_contained_in): Copy lto_output and
no_export flags from archive.
* archive.c (open_nested_file): New function, setting lto_output
and no_export, extracted from..
(find_nested_archive): ..here. Flip params. Rename from
_bfd_find_nested_archive.
(_bfd_get_elt_at_filepos): Correct var typo. Use open_nested_file.
(_bfd_look_for_bfd_in_cache): Copy no_export.
* elflink.c (elf_link_add_object_symbols): Remove now unnecessary
my_archive->no_export test.
(elf_link_input_bfd): Drop existing lto_output STT_FILE syms.
Don't use the file name when adding lto_output STT_FILE sym.
* bfd-in2.h: Regenerate.
ld/
* ldlang.h (struct lang_input_statement_flags): Add lto_output.
* ldlang.c (lang_process): Don't set loading_lto_outputs.
* ldfile.c (ldfile_try_open_bfd): Transfer entry flags.lto_output
to bfd.
* plugin.c (add_input_file, add_input_library): Set flags.lto_output.
The changes to reorder sections for better relro protection on powerpc64,
3e2b0f31, 23283c1b, and 5ad18f16, run into a problem with xlc.
xlc -qdatalocal puts global variables into .toc, which means that .toc
must be writable. The simplest way to accomplish this is to edit the
linker script to remove .toc sections from .got on detecting xlc object
files.
bfd/
* elf64-ppc.h (struct ppc64_elf_params): Add "object_in_toc".
* elf64-ppc.c (ppc64_elf_add_symbol_hook): Assume that global symbols
in .toc indicate xlc compiled code that might require a rw .toc.
ld/
* emulparams/elf64ppc.sh (INITIAL_READWRITE_SECTIONS): Define.
* emultempl/ppc64elf.em (params): Init new field.
(ppc_after_open): New function.
(LDEMUL_AFTER_OPEN): Define.
* ldlang.c (lang_final): Whitespace fix.
ld/testsuite/
* ld-powerpc/tocvar.d, * ld-powerpc/tocvar.s: New test.
* ld-powerpc/tocnovar.d, * ld-powerpc/tocnovar.s: New test.
* ld-powerpc/powerpc.exp: Run tocvar and tocnovar.
This patch fixes PR 4643 by allowing symbols in the LENGTH and ORIGIN
fields of MEMORY regions. Previously, only constants and constant
expressions are allowed.
For the AVR target, this helps define memory constraints more
accurately (per device), without having to create a ton of device
specific linker scripts.
ld/
PR 4643
* ldexp.c (fold_name): Fold LENGTH only after
lang_first_phase_enum.
* ldgram.y (memory_spec): Don't evaluate ORIGIN and LENGTH
rightaway.
* ldlang.h (struct memory_region_struct): Add origin_exp and
length_exp fields.
* ldlang.c (lang_do_memory_regions): New.
(lang_memory_region_lookup): Initialize origin_exp and
length_exp fields.
(lang_process): Call lang_do_memory_regions.
ld/testsuite/
* ld-scripts/memory.t: Define new symbol tred.
* ld-scripts/memory_sym.t: New.
* ld-scripts/script.exp: Perform MEMORY with symbols test, and
conditionally check values of linker symbols.
When creating a linker mapfile (using -Map=MAPFILE), we previously would
always try to evaluate the expression from a PROVIDE statement.
However, this is not always safe, consider:
PROVIDE (foo = 0x10);
PROVIDE (bar = foo);
In this example, if neither 'foo' or 'bar' is needed, then while
generating the linker mapfile evaluating the expression for 'foo' is
harmless (just the value 0x10). However, evaluating the expression for
'bar' requires the symbol 'foo', which is undefined. This used to cause
a fatal error.
This patch changes the behaviour, so that when the destination of the
PROVIDE is not defined (that is the PROVIDE is not going to provide
anything) the expression is not evaluated, and instead a special string
is displayed to indicate that the linker is discarding the PROVIDE
statement.
This change not only fixes the spurious undefined symbol error, but also
means that a user can now tell if a PROVIDE statement has provided
anything by inspecting the linker mapfile, something that could not be
done before.
ld/ChangeLog:
* ldlang.c (print_assignment): Only evaluate the expression for a
PROVIDE'd assignment when the destination is being defined.
Display a special message for PROVIDE'd symbols that are not being
provided.
ld/testsuite/ChangeLog:
* ld-scripts/provide-4.d: New file.
* ld-scripts/provide-4-map.d: New file.
* ld-scripts/provide-4.t: New file.
* ld-scripts/provide-5.d: New file.
* ld-scripts/provide-5.s: New file.
* ld-scripts/provide-5-map.d: New file.
* ld-scripts/provide-5.t: New file.
* ld-scripts/provide.exp: Run the provide-4.d and provide-5.d
tests.
Running lang_common before garbage collection means slightly less work
in garbage collection code, since common symbols should no longer
appear there. It does have the side effect of keeping linker script
symbols (at least those defined outside of sections) global too,
hence some testsuite churn.
bfd/
PR 17165
* elf-bfd.h (ELF_COMMON_DEF): Note that this might be true for
linker script assignments too.
* elflink.c (elf_gc_sweep_symbol): Don't drop ELF_COMMON_DEF syms.
(bfd_elf_gc_mark_dynamic_ref_symbol): Similarly.
ld/
PR 17165
* ldlang.c (lang_process): Run lang_common before lang_gc_sections.
ld/testsuite/
* ld-gc/pr14265.d,
* ld-cris/tls-gc-68.d,
* ld-cris/tls-gc-69.d,
* ld-cris/tls-gc-70.d,
* ld-cris/tls-gc-71.d,
* ld-cris/tls-gc-75.d,
* ld-cris/tls-gc-76.d,
* ld-cris/tls-gc-79.d,
* ld-mmix/bpo-10.d,
* ld-mmix/bpo-11.d: Update.
This moves support code for DEFINED to ldexp.c where it is used,
losing the lang_ prefix on identifiers. Two new functions are needed
to initialize and clean up to hash table, but other than that there
are no functional changes here.
* ldexp.c (struct definedness_hash_entry, definedness_table)
(definedness_newfunc, symbol_defined, update_definedness): Move
and rename from..
* ldlang.h (struct lang_definedness_hash_entry): ..here,..
* ldlang.c (lang_definedness_table, lang_definedness_newfunc)
(lang_symbol_defined, lang_update_definedness): ..and here.
* ldexp.c (ldexp_init, ldexp_finish): New functions, extracted from..
* ldlang.c (lang_init, lang_finish): ..here.
* ldexp.h (ldexp_init, ldexp_finish): Declare.
* ldlang.h (lang_symbol_defined, lang_update_definedness): Delete.
* ldmain.c (main): Call ldexp_init and ldexp_finish.
Andrew Burgess
Alan Modra
Nick Clifton
Igor Kudrin
Ambrogino Modigliani
H.J. Lu
Matthew Fortune
Dan Gisselquist
Senthil Kumar Selvaraj
Tristan Gingold