/usr/include/sys/mman.h:81:0: note: this is the location of the previous definition
#define SHARED 0x10
PR 27064
* deffilep.y (SHARED_K): Rename from SHARED. Update uses.
Match the condition used in `elf_vax_instantiate_got_entries' for the
creation of GOT entries in the processing of R_VAX_GOT32 relocations in
`elf_vax_check_relocs', removing incorrect warnings about a GOT addend
mismatch like:
./ld-new: tmpdir/got-local-ref-off-r.o: warning: GOT addend of 1 to `bar_hidden' does not match previous GOT addend of 0
./ld-new: tmpdir/got-local-ref-off-r.o: warning: GOT addend of 2 to `bar_hidden' does not match previous GOT addend of 0
and corresponding failures with the test cases newly added here:
FAIL: GOT test (executable hidden reference with offset)
FAIL: GOT test (executable visible reference with offset)
for symbols that are considered local for reasons other than having been
forced local with a version script, which is usually the ELF visibility.
Correct code is produced regardless, but the warning breaks `-Werror'
compilation and may upset people regardless.
Interestingly this shows with executable links only, because in shared
library links code from `elf_link_add_object_symbols' triggers:
/* If the symbol already has a dynamic index, but
visibility says it should not be visible, turn it into
a local symbol. */
switch (ELF_ST_VISIBILITY (h->other))
{
case STV_INTERNAL:
case STV_HIDDEN:
(*bed->elf_backend_hide_symbol) (info, h, TRUE);
dynsym = FALSE;
break;
}
that sets `h->forced_local' like with a version script.
Add suitable test cases including disassembly to verify correct code has
been produced where no warnings have been issued, and that warnings do
get issued where necessary. Do not verify (broken) code produced in the
latter case; we should probably make the warning an error, or preferably
actually start supporting GOT references with different addends as they
appear feasible with explicitly relocated GOT that we use.
bfd/
* elf32-vax.c (elf_vax_check_relocs) <R_VAX_GOT32>: Use
SYMBOL_REFERENCES_LOCAL rather than `h->forced_local' to check
whether the symbol referred is local or not.
ld/
* testsuite/ld-vax-elf/got-local-exe-off-hidden.dd: New test
dump.
* testsuite/ld-vax-elf/got-local-exe-off-visible.dd: New test
dump.
* testsuite/ld-vax-elf/got-local-lib-off-hidden.dd: New test
dump.
* testsuite/ld-vax-elf/got-local-lib-off-visible.ed: New test
dump.
* testsuite/ld-vax-elf/got-local-off-external.ed: New test dump.
* testsuite/ld-vax-elf/got-local-exe-off.xd: New test dump.
* testsuite/ld-vax-elf/got-local-lib-off.xd: New test dump.
* testsuite/ld-vax-elf/got-local.ld: New test linker script.
* testsuite/ld-vax-elf/got-local-aux-off.s: New test source.
* testsuite/ld-vax-elf/got-local-def-off.s: New test source.
* testsuite/ld-vax-elf/got-local-ref-off-external.s: New test
source.
* testsuite/ld-vax-elf/got-local-ref-off-hidden.s: New test
source.
* testsuite/ld-vax-elf/got-local-ref-off-visible.s: New test
source.
* testsuite/ld-vax-elf/vax-elf.exp: Run the new tests.
Since converting load to mov needs to rewrite the REX byte and we don't
know if there is a REX byte with GOTPCREL relocation, do it only for
GOTPCRELX relocations.
bfd/
PR ld/27016
* elf64-x86-64.c (elf_x86_64_convert_load_reloc): Convert load
to mov only for GOTPCRELX relocations.
ld/
PR ld/27016
* testsuite/ld-x86-64/x86-64.exp: Run pr27016a and pr27016b.
* testsuite/ld-x86-64/pr27016a.d: New file.
* testsuite/ld-x86-64/pr27016a.s: Likewise.
* testsuite/ld-x86-64/pr27016b.d: Likewise.
* testsuite/ld-x86-64/pr27016b.s: Likewise.
Prior to this patch
ld -shared --version-script=pr26979.ver pr26978a.o pr26978b.o
results in
ld: pr26978b.o: in function `foo_v1':
(.text+0x0): multiple definition of `foo@v1'
ld: pr26978b.o:(*IND*+0x0): multiple definition of `foo'
while
ld -shared --version-script=pr26979.ver pr26978b.o pr26978a.o
results in no error, but some odd dynamic symbols.
... 0 NOTYPE GLOBAL DEFAULT 7 foo@v1
... 0 NOTYPE WEAK DEFAULT 7 foo@@v1
When linking an undecorated reference to foo against such a shared
library, ld complains about multiple definitions of foo@v1 while gold
creates a dynamic reference to foo@v1. That results in foo@v1 being
used at runtime.
While we could error in both cases, it is reasonable to say foo@v1 and
foo@@v1 are in fact the same symbol. (Same name, same version. The
only real difference is that foo@@v1 satisfies a reference to plain
foo, while foo@v1 does not.) Just as merging a weak undecorated sym
with a strong sym results in the strong sym prevailing, so should the
strong foo@v1 prevail. And since there is a definition that satisfies
plain foo, the foo@@v1 variety of dynamic symbol should be emitted at
the foo@v1 value. That makes the testcase that currently links
continue to produce a shared library, and that shared library can now
be used by both ld and gold with the same runtime behaviour as when
using gold with the odd dynamic symbol library.
bfd/
PR 26978
* elflink.c (_bfd_elf_add_default_symbol): Handle the case where
a new weak sym@@ver should be overridden by an existing sym@ver.
(elf_link_add_object_symbols): Don't _bfd_elf_add_default_symbol
for a new weak sym@ver when sym@@ver already exists.
* linker.c (link_action): Choose MIND for previous indirect,
current def, rather than MDEF.
(_bfd_generic_link_add_one_symbol <MIND>): Handle redefinition of
weak indirect symbol.
ld/
* testsuite/ld-elf/pr26978a.d,
* testsuite/ld-elf/pr26978a.s,
* testsuite/ld-elf/pr26978b.d,
* testsuite/ld-elf/pr26978b.s: New tests.
Add extended mnemonics used in the HLASM assembler. All of them are
just aliases for instructions we already support and help when
assembling code which was written for the HLASM assembler.
The HLASM mnemonics are documented here:
https://www.ibm.com/support/knowledgecenter/SSENW6_1.6.0/com.ibm.hlasm.v1r6.asm/asmr1023.pdf
See the 'Branching with extended mnemonic codes' chapter.
objdump will still print the existing mnemonics with the exception of
relative nop branches (i.e. conditional branches with an empty
condition code mask). Now we have jnop and jgnop which will be used
by objdump when possible.
The same change have been applied to the LLVM assembler:
https://reviews.llvm.org/D92185
opcodes/
* s390-opc.txt: Add extended mnemonics.
gas/
* testsuite/gas/s390/esa-g5.s: Test new extended mnemonics.
* testsuite/gas/s390/esa-g5.d: Likewise.
* testsuite/gas/s390/esa-z900.s: Likewise.
* testsuite/gas/s390/esa-z900.d: Likewise.
* testsuite/gas/s390/zarch-z900.s: Likewise.
* testsuite/gas/s390/zarch-z900.d: Likewise.
ld/
* testsuite/ld-s390/tlsbin_64.dd: The newly added jgnop mnemonic
replaces long relative branches with empty condition code mask.
A couple of lines in the vax-elf.exp test script exceed 80 characters;
wrap them.
ld/
* testsuite/ld-vax-elf/vax-elf.exp: Wrap excessively long lines
throughout.
Required for the expected "CU:" to be emitted for long
source-paths. See binutils/dwarf.c:
if (do_wide || strlen (directory) < 76)
printf (_("CU: %s/%s:\n"), directory, file_table[0].name);
else
printf ("%s:\n", file_table[0].name);
Linkonce sections and comdat groups can be mixed only if comdat groups
have only a single member with matching symbol table entries. Xfail
ld/26936 test:
1. If comdat groups always have more than one member.
2. If symbol table entries in linkonce and comdat group don't match.
3. If the assembly source file is renamed.
PR ld/26936
* testsuite/ld-elf/pr26936.d: Xfail targets which don't support
mixing linkonce and comdat sections.
Section ordering is important for _bfd_elf_map_sections_to_segments
and assign_file_positions_for_load_sections, which are only prepared
to handle sections in increasing LMA order. When zero size sections
are involved it is possible to have multiple sections at the same LMA.
In that case the zero size sections must sort before any non-zero size
sections regardless of their types.
bfd/
PR 26907
* elf.c (elf_sort_sections): Don't sort zero size !load sections
after load sections.
ld/
* testsuite/ld-elf/pr26907.ld,
* testsuite/ld-elf/pr26907.s,
* testsuite/ld-elf/pr26907.d: New test.
The ".persistent" section is for data that should be initialized during
load, but not during application reset.
The ".noinit" section is for data that should not be initialized during
load or application reset.
Targets utilizing the elf.sc linker script template can define
HAVE_{NOINIT,PERSISTENT}=yes to include the .noinit or .persistent
output sections in the generated linker script.
Targets with existing support for .noinit did not handle unique
.noinit.* and .gnu.linkonce.n.* sections the .noinit output section,
this patch also fixes that.
bfd/ChangeLog:
* elf.c (special_sections_g): Add .gnu.linkonce.n and .gnu.linkonce.p.
(special_sections_n): Add .noinit.
(special_sections_p): Add .persistent.
binutils/ChangeLog:
* testsuite/lib/binutils-common.exp (supports_noinit_section): New.
(supports_persistent_section): New.
gas/ChangeLog:
* testsuite/gas/elf/elf.exp: Run new tests.
* testsuite/gas/elf/section25.d: New test.
* testsuite/gas/elf/section25.s: New test.
* testsuite/gas/elf/section26.d: New test.
* testsuite/gas/elf/section26.s: New test.
ld/ChangeLog:
* emulparams/armelf.sh (OTHER_SECTIONS): Remove .noinit section
definition.
Define HAVE_{NOINIT,PERSISTENT}=yes.
* scripttempl/avr.sc (.noinit): Add .noinit.* and .gnu.linkonce.n.*
input section wildcard patterns.
* scripttempl/elf.sc: Define .noinit and .persistent sections when
HAVE_NOINIT or HAVE_PERSISTENT are defined to "yes".
* scripttempl/elf32msp430.sc (.noinit): Add .noinit.* and
.gnu.linkonce.n.*. input section wildcard patterns.
(.persistent): Add .persistent.* and
.gnu.linkonce.p.*. input section wildcard patterns.
* scripttempl/elfarcv2.sc (.noinit): Add .noinit.* and
.gnu.linkonce.n.*. input section wildcard patterns.
* scripttempl/pru.sc: Likewise.
* testsuite/ld-elf/noinit-sections-1.d: New test.
* testsuite/ld-elf/noinit-sections-2.d: New test.
* testsuite/ld-elf/noinit-sections-2.l: New test.
* testsuite/ld-elf/noinit-sections.s: New test.
* testsuite/ld-elf/persistent-sections-1.d: New test.
* testsuite/ld-elf/persistent-sections-2.d: New test.
* testsuite/ld-elf/persistent-sections-2.l: New test.
* testsuite/ld-elf/persistent-sections.s: New test.
Many targets fail this test due to -z noseparate-code not being
supported, or _start not being the proper entry symbol, or "as -g"
something other than "generate debug".
PR 26936
* testsuite/ld-elf/pr26936.d: Pass --gen-debug to gas rather than -g.
Only run when -shared -z options are supported.
* testsuite/ld-elf/pr26936b.s: Define more entry symbols.
When mixing linkonce and comdat sections, we need to keep searching to
get the real kept section.
bfd/
PR ld/26936
* elflink.c (_bfd_elf_check_kept_section): Get the real kept
section.
ld/
PR ld/26936
* testsuite/ld-elf/pr26936.d: New file.
* testsuite/ld-elf/pr26936a.s: Likewise.
* testsuite/ld-elf/pr26936b.s: Likewise.
* testsuite/ld-elf/pr26936c.s: Likewise.
Previously, ld merged duplicate output sections if such existed in
scripts, except for those with a constraint of SPECIAL. This makes
scripts with duplicate output section statements create duplicate
output sections in the linker output file.
* ldlang.c (lang_output_section_statement_lookup): Change "create"
parameter to a tristate, if 2 then always create a new output
section statement. Update all callers, with
lang_enter_output_section_statement using "2".
(map_input_to_output_sections): Don't ignore SPECIAL constraint
here.
* ldlang.h (lang_output_section_statement_type): Update prototype.
(lang_output_section_find): Update.
GOTPCRELX relocations can be transformed only when addend == -4. Add
tests for GOTPCRELX relocations with addend != -4.
PR gold/26939
* testsuite/ld-x86-64/pr26939-x32.d: New file.
* testsuite/ld-x86-64/pr26939.d: Likewise.
* testsuite/ld-x86-64/pr26939.s: Likewise.
* testsuite/ld-x86-64/x86-64.exp: Run PR gold/26939 tests.
This is to address the regressions addressed by Nic [1].
The regular expression pattern for the tls_ie-01 test was
too strict and raising false alarms. The new pattern only
looks for matches that should be there AND ignores the boiler
plates from the object dump.
[1] New failures for ARC targets in linker testsuite
https://sourceware.org/pipermail/binutils/2020-November/114177.html
ld/
* testsuite/ld-arc/tls_ie-01.d: Use a more general pattern.
I get the feedback recently that enable linker relaxations may fail to
build some program. Consider the following case,
.text
foo:
addi a0, a0, %pcrel_lo(.L2)
call foo
.L1: auipc a1, %pcrel_hi(data_g)
addi a1, a1, %pcrel_lo(.L1)
lui a2, %hi(data_g)
addi a2, a2, %lo(data_g)
lui a3, %tprel_hi(data_t)
add a3, a3, tp, %tprel_add(data_t)
addi a3, a3, %tprel_lo(data_t)
.L2: auipc a0, %pcrel_hi(data_g)
.data
.word 0x0
.global data_g
data_g: .word 0x1
.section .tbss
data_t: .word 0x0
The current ld reports `dangerous relocation error` when doing the
pcgp relaxation,
test.o: in function `foo':
(.text+0x0): dangerous relocation: %pcrel_lo missing matching %pcrel_hi
The .L2 auipc should not be removed since it is behind the corresponding
addi, so we record the information in the pcgp_relocs table to avoid
removing the auipc later. But current ld still remove it since we do not
update the pcgp_relocs table while doing other relaxations. I have two
solutions to fix the problem,
1. Update the pcgp_relocs table once we actually delete the code.
2. Add new relax pass to do the pcgp relaxations
At first I tried to do the first solution, and we need to update at
least three information - hi_sec_off of riscv_pcgp_lo_reloc, hi_sec_off
and hi_addr (symbol value) of riscv_pcgp_hi_reloc. Update the hi_sec_off
is simple, but it is more complicate to update the symbol value, since we
almost have to do parts the same works of _bfd_riscv_relax_call again in
the riscv_relax_delete_bytes to get the correct symbol value.
Compared with the first solution, the second one is more intuitive and
simple. We add a new relax pass to do the pcgp relaxations later, so
we will get all the information correctly in the _bfd_riscv_relax_call,
including the symbol value, without changing so much code. I do not see
any penalty by adding a new relax pass for now, so it should be fine
to delay the pcgp relaxations.
Besides, I have pass all riscv-gnu-toolchain regressions for this patch.
bfd/
* elfnn-riscv.c (_bfd_riscv_relax_section): Add a new relax pass
to do the pcgp relaxation later, after the lui and call relaxations,
but before the delete and alignment relaxations.
ld/
* emultempl/riscvelf.em (riscv_elf_before_allocation): Change
link_info.relax_pass from 3 to 4.
* testsuite/ld-riscv-elf/pcgp-relax.d: New testcase.
* testsuite/ld-riscv-elf/pcgp-relax.s: Likewise.
* testsuite/ld-riscv-elf/ld-riscv-elf.exp: Updated.
Some type kinds in CTF (functions, arrays, pointers, slices, and
cvr-quals) are intrinsically nameless: the ctt_name field in the CTF
is always zero, and the libctf API provides no way to set a name.
But the compiler can and does sometimes set names for some of these
kinds: in particular, the name it sets on CTF_K_FUNCTION types is the
means it uses to force the name of the function into the string table
so that it can point at it from the function info section.
So null out the name at hashing time so that the deduplicator can
correctly detect that e.g. function types identical but for name should
be considered truly identical, since they will not have a name when the
deduplicator re-emits them into the output.
ld/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* testsuite/ld-ctf/data-func-conflicted.d: Shrink the expected
size of the type section now that function types are being
deduplicated properly.
libctf/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* ctf-dedup.c (ctf_dedup_rhash_type): Null out the names of nameless
type kinds, just in case the input has named them.
The flags word is nonzero now (so all the tests have been adjusted to
not depend on its content): some of them have data objects and functions
in the data object and function info sections now, rather than in the
variable section or recorded nowhere. There is a new test for
parent/child relationships and index section emission.
ld/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* testsuite/ld-ctf/array.d: Adjust for nonzero flags word and
public symbols in the data section rather than variables: use
sysv hash style to keep test results the same on non-GNU targets.
* testsuite/ld-ctf/diag-cttname-null.d: Likewise.
* testsuite/ld-ctf/diag-cuname.d: Likewise.
* testsuite/ld-ctf/diag-parlabel.d: Likewise.
* testsuite/ld-ctf/slice.d: Likewise.
* testsuite/ld-ctf/function.d: Likewise, but in the function section.
* testsuite/ld-ctf/conflicting-cycle-1.B-1.d: Adjust for nonzero
flags word.
* testsuite/ld-ctf/conflicting-cycle-1.B-2.d: Likewise.
* testsuite/ld-ctf/conflicting-cycle-1.parent.d: Likewise.
* testsuite/ld-ctf/conflicting-cycle-2.A-1.d: Likewise.
* testsuite/ld-ctf/conflicting-cycle-2.A-2.d: Likewise.
* testsuite/ld-ctf/conflicting-cycle-2.parent.d: Likewise.
* testsuite/ld-ctf/conflicting-cycle-3.C-1.d: Likewise.
* testsuite/ld-ctf/conflicting-cycle-3.C-2.d: Likewise.
* testsuite/ld-ctf/conflicting-cycle-3.parent.d: Likewise.
* testsuite/ld-ctf/cross-tu-noncyclic.d: Likewise.
* testsuite/ld-ctf/cycle-1.d: Likewise.
* testsuite/ld-ctf/cycle-2.A.d: Likewise.
* testsuite/ld-ctf/cycle-2.B.d: Likewise.
* testsuite/ld-ctf/cycle-2.C.d: Likewise.
* testsuite/ld-ctf/diag-wrong-magic-number-mixed.d: Likewise.
* testsuite/ld-ctf/super-sub-cycles.d: Likewise.
* testsuite/ld-ctf/data-func-1.c: New test.
* testsuite/ld-ctf/data-func-2.c: Likewise.
* testsuite/ld-ctf/data-func-conflicted.d: Likewise.
This is embarrassing.
The whole point of CTF is that it remains intact even after a binary is
stripped, providing a compact mapping from symbols to types for
everything in the externally-visible interface of an ELF object: it has
connections to the symbol table for that purpose, and to the string
table to avoid duplicating symbol names. So it's a shame that the hooks
I implemented last year served to hook it up to the .symtab and .strtab,
which obviously disappear on strip, leaving any accompanying the CTF
dict containing references to strings (and, soon, symbols) which don't
exist any more because their containing strtab has been vaporized. The
original Solaris design used .dynsym and .dynstr (well, actually,
.ldynsym, which has more symbols) which do not disappear. So should we.
Thankfully the work we did before serves as guide rails, and adjusting
things to use the .dynstr and .dynsym was fast and easy. The only
annoyance is that the dynsym is assembled inside elflink.c in a fairly
piecemeal fashion, so that the easiest way to get the symbols out was to
hook in before every call to swap_symbol_out (we also leave in a hook in
front of symbol additions to the .symtab because it seems plausible that
we might want to hook them in future too: for now that hook is unused).
We adjust things so that rather than being offered a whole hash table of
symbols at once, libctf is now given symbols one at a time, with st_name
indexes already resolved and pointing at their final .dynstr offsets:
it's now up to libctf to resolve these to names as needed using the
strtab info we pass it separately.
Some bits might be contentious. The ctf_new_dynstr callback takes an
elf_internal_sym, and this remains an elf_internal_sym right down
through the generic emulation layers into ldelfgen. This is no worse
than the elf_sym_strtab we used to pass down, but in the future when we
gain non-ELF CTF symtab support we might want to lower the
elf_internal_sym to some other representation (perhaps a
ctf_link_symbol) in bfd or in ldlang_ctf_new_dynsym. We rename the
'apply_strsym' hooks to 'acquire_strings' instead, becuse they no longer
have anything to do with symbols.
There are some API changes to pieces of API which are technically public
but actually totally unused by anything and/or unused by anything but ld
so they can change freely: the ctf_link_symbol gains new fields to allow
symbol names to be given as strtab offsets as well as strings, and a
symidx so that the symbol index can be passed in. ctf_link_shuffle_syms
loses its callback parameter: the idea now is that linkers call the new
ctf_link_add_linker_symbol for every symbol in .dynsym, feed in all the
strtab entries with ctf_link_add_strtab, and then a call to
ctf_link_shuffle_syms will apply both and arrange to use them to reorder
the CTF symtab at CTF serialization time (which is coming in the next
commit).
Inside libctf we have a new preamble flag CTF_F_DYNSTR which is always
set in v3-format CTF dicts from this commit forwards: CTF dicts without
this flag are associated with .strtab like they used to be, so that old
dicts' external strings don't turn to garbage when loaded by new libctf.
Dicts with this flag are associated with .dynstr and .dynsym instead.
(The flag is not the next in sequence because this commit was written
quite late: the missing flags will be filled in by the next commit.)
Tests forthcoming in a later commit in this series.
bfd/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* elflink.c (elf_finalize_dynstr): Call examine_strtab after
dynstr finalization.
(elf_link_swap_symbols_out): Don't call it here. Call
ctf_new_symbol before swap_symbol_out.
(elf_link_output_extsym): Call ctf_new_dynsym before
swap_symbol_out.
(bfd_elf_final_link): Likewise.
* elf.c (swap_out_syms): Pass in bfd_link_info. Call
ctf_new_symbol before swap_symbol_out.
(_bfd_elf_compute_section_file_positions): Adjust.
binutils/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* readelf.c (dump_section_as_ctf): Use .dynsym and .dynstr, not
.symtab and .strtab.
include/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* bfdlink.h (struct elf_sym_strtab): Replace with...
(struct elf_internal_sym): ... this.
(struct bfd_link_callbacks) <examine_strtab>: Take only a
symstrtab argument.
<ctf_new_symbol>: New.
<ctf_new_dynsym>: Likewise.
* ctf-api.h (struct ctf_link_sym) <st_symidx>: New.
<st_nameidx>: Likewise.
<st_nameidx_set>: Likewise.
(ctf_link_iter_symbol_f): Removed.
(ctf_link_shuffle_syms): Remove most parameters, just takes a
ctf_dict_t now.
(ctf_link_add_linker_symbol): New, split from
ctf_link_shuffle_syms.
* ctf.h (CTF_F_DYNSTR): New.
(CTF_F_MAX): Adjust.
ld/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* ldelfgen.c (struct ctf_strsym_iter_cb_arg): Rename to...
(struct ctf_strtab_iter_cb_arg): ... this, changing fields:
<syms>: Remove.
<symcount>: Remove.
<symstrtab>: Rename to...
<strtab>: ... this.
(ldelf_ctf_strtab_iter_cb): Adjust.
(ldelf_ctf_symbols_iter_cb): Remove.
(ldelf_new_dynsym_for_ctf): New, tell libctf about a single
symbol.
(ldelf_examine_strtab_for_ctf): Rename to...
(ldelf_acquire_strings_for_ctf): ... this, only doing the strtab
portion and not symbols.
* ldelfgen.h: Adjust declarations accordingly.
* ldemul.c (ldemul_examine_strtab_for_ctf): Rename to...
(ldemul_acquire_strings_for_ctf): ... this.
(ldemul_new_dynsym_for_ctf): New.
* ldemul.h: Adjust declarations accordingly.
* ldlang.c (ldlang_ctf_apply_strsym): Rename to...
(ldlang_ctf_acquire_strings): ... this.
(ldlang_ctf_new_dynsym): New.
(lang_write_ctf): Call ldemul_new_dynsym_for_ctf with NULL to do
the actual symbol shuffle.
* ldlang.h (struct elf_strtab_hash): Adjust accordingly.
* ldmain.c (bfd_link_callbacks): Wire up new/renamed callbacks.
libctf/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* ctf-link.c (ctf_link_shuffle_syms): Adjust.
(ctf_link_add_linker_symbol): New, unimplemented stub.
* libctf.ver: Add it.
* ctf-create.c (ctf_serialize): Set CTF_F_DYNSTR on newly-serialized
dicts.
* ctf-open-bfd.c (ctf_bfdopen_ctfsect): Check for the flag: open the
symtab/strtab if not present, dynsym/dynstr otherwise.
* ctf-archive.c (ctf_arc_bufpreamble): New, get the preamble from
some arbitrary member of a CTF archive.
* ctf-impl.h (ctf_arc_bufpreamble): Declare it.
The naming of the ctf_file_t type in libctf is a historical curiosity.
Back in the Solaris days, CTF dictionaries were originally generated as
a separate file and then (sometimes) merged into objects: hence the
datatype was named ctf_file_t, and known as a "CTF file". Nowadays, raw
CTF is essentially never written to a file on its own, and the datatype
changed name to a "CTF dictionary" years ago. So the term "CTF file"
refers to something that is never a file! This is at best confusing.
The type has also historically been known as a 'CTF container", which is
even more confusing now that we have CTF archives which are *also* a
sort of container (they contain CTF dictionaries), but which are never
referred to as containers in the source code.
So fix this by completing the renaming, renaming ctf_file_t to
ctf_dict_t throughout, and renaming those few functions that refer to
CTF files by name (keeping compatibility aliases) to refer to dicts
instead. Old users who still refer to ctf_file_t will see (harmless)
pointer-compatibility warnings at compile time, but the ABI is unchanged
(since C doesn't mangle names, and ctf_file_t was always an opaque type)
and things will still compile fine as long as -Werror is not specified.
All references to CTF containers and CTF files in the source code are
fixed to refer to CTF dicts instead.
Further (smaller) renamings of annoyingly-named functions to come, as
part of the process of souping up queries across whole archives at once
(needed for the function info and data object sections).
binutils/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* objdump.c (dump_ctf_errs): Rename ctf_file_t to ctf_dict_t.
(dump_ctf_archive_member): Likewise.
(dump_ctf): Likewise. Use ctf_dict_close, not ctf_file_close.
* readelf.c (dump_ctf_errs): Rename ctf_file_t to ctf_dict_t.
(dump_ctf_archive_member): Likewise.
(dump_section_as_ctf): Likewise. Use ctf_dict_close, not
ctf_file_close.
gdb/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* ctfread.c: Change uses of ctf_file_t to ctf_dict_t.
(ctf_fp_info::~ctf_fp_info): Call ctf_dict_close, not ctf_file_close.
include/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* ctf-api.h (ctf_file_t): Rename to...
(ctf_dict_t): ... this. Keep ctf_file_t around for compatibility.
(struct ctf_file): Likewise rename to...
(struct ctf_dict): ... this.
(ctf_file_close): Rename to...
(ctf_dict_close): ... this, keeping compatibility function.
(ctf_parent_file): Rename to...
(ctf_parent_dict): ... this, keeping compatibility function.
All callers adjusted.
* ctf.h: Rename references to ctf_file_t to ctf_dict_t.
(struct ctf_archive) <ctfa_nfiles>: Rename to...
<ctfa_ndicts>: ... this.
ld/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* ldlang.c (ctf_output): This is a ctf_dict_t now.
(lang_ctf_errs_warnings): Rename ctf_file_t to ctf_dict_t.
(ldlang_open_ctf): Adjust comment.
(lang_merge_ctf): Use ctf_dict_close, not ctf_file_close.
* ldelfgen.h (ldelf_examine_strtab_for_ctf): Rename ctf_file_t to
ctf_dict_t. Change opaque declaration accordingly.
* ldelfgen.c (ldelf_examine_strtab_for_ctf): Adjust.
* ldemul.h (examine_strtab_for_ctf): Likewise.
(ldemul_examine_strtab_for_ctf): Likewise.
* ldeuml.c (ldemul_examine_strtab_for_ctf): Likewise.
libctf/ChangeLog
2020-11-20 Nick Alcock <nick.alcock@oracle.com>
* ctf-impl.h: Rename ctf_file_t to ctf_dict_t: all declarations
adjusted.
(ctf_fileops): Rename to...
(ctf_dictops): ... this.
(ctf_dedup_t) <cd_id_to_file_t>: Rename to...
<cd_id_to_dict_t>: ... this.
(ctf_file_t): Fix outdated comment.
<ctf_fileops>: Rename to...
<ctf_dictops>: ... this.
(struct ctf_archive_internal) <ctfi_file>: Rename to...
<ctfi_dict>: ... this.
* ctf-archive.c: Rename ctf_file_t to ctf_dict_t.
Rename ctf_archive.ctfa_nfiles to ctfa_ndicts.
Rename ctf_file_close to ctf_dict_close. All users adjusted.
* ctf-create.c: Likewise. Refer to CTF dicts, not CTF containers.
(ctf_bundle_t) <ctb_file>: Rename to...
<ctb_dict): ... this.
* ctf-decl.c: Rename ctf_file_t to ctf_dict_t.
* ctf-dedup.c: Likewise. Rename ctf_file_close to
ctf_dict_close. Refer to CTF dicts, not CTF containers.
* ctf-dump.c: Likewise.
* ctf-error.c: Likewise.
* ctf-hash.c: Likewise.
* ctf-inlines.h: Likewise.
* ctf-labels.c: Likewise.
* ctf-link.c: Likewise.
* ctf-lookup.c: Likewise.
* ctf-open-bfd.c: Likewise.
* ctf-string.c: Likewise.
* ctf-subr.c: Likewise.
* ctf-types.c: Likewise.
* ctf-util.c: Likewise.
* ctf-open.c: Likewise.
(ctf_file_close): Rename to...
(ctf_dict_close): ...this.
(ctf_file_close): New trivial wrapper around ctf_dict_close, for
compatibility.
(ctf_parent_file): Rename to...
(ctf_parent_dict): ... this.
(ctf_parent_file): New trivial wrapper around ctf_parent_dict, for
compatibility.
* libctf.ver: Add ctf_dict_close and ctf_parent_dict.
binutils/ChangeLog:
* testsuite/binutils-all/readelf-maskos-1a.d: Fix test for unrecognized
bit set in SHF_MASKOS range.
* testsuite/binutils-all/readelf-maskos-1b.d: Likewise.
* testsuite/binutils-all/readelf-maskos-unknown.s: New test.
ld/ChangeLog:
* testsuite/ld-elf/retain3.s: Move symbolic reference into writeable
.data section from read-only .text section.
* testsuite/ld-elf/retain5.d: Don't pass --print-gc-sections for test
that doesn't require it.
* testsuite/ld-elf/retain6a.d: Adjust test.
* testsuite/ld-elf/retain6main.s: Move symbolic reference into writeable
.data section from read-only .text section.
The SHF_GNU_RETAIN section flag is an extension to the GNU ELF OSABI.
It is defined as follows:
=========================================================
Section Attribute Flags
+-------------------------------------+
| Name | Value |
+-------------------------------------+
| SHF_GNU_RETAIN | 0x200000 (1 << 21) |
+-------------------------------------+
SHF_GNU_RETAIN
The link editor should not garbage collect the section.
=========================================================
The .section directive accepts the "R" flag, which indicates
SHF_GNU_RETAIN should be applied to the section.
There is not a direct mapping of SHF_GNU_RETAIN to the BFD
section flag SEC_KEEP. Keeping these flags distinct allows
SHF_GNU_RETAIN sections to be explicitly removed by placing them in
/DISCARD/.
bfd/ChangeLog:
* elf-bfd.h (enum elf_gnu_osabi): Add elf_gnu_osabi_retain.
(struct elf_obj_tdata): Increase has_gnu_osabi to 4 bits.
* elf.c (_bfd_elf_make_section_from_shdr): Set elf_gnu_osabi_retain
for SHF_GNU_RETAIN.
(_bfd_elf_final_write_processing): Report if SHF_GNU_RETAIN is
not supported by the OSABI.
Adjust error messages.
* elflink.c (elf_link_input_bfd): Copy enabled has_gnu_osabi bits from
input BFD to output BFD.
(bfd_elf_gc_sections): gc_mark the section if SHF_GNU_RETAIN is set.
binutils/ChangeLog:
* NEWS: Announce SHF_GNU_RETAIN support.
* readelf.c (get_elf_section_flags): Handle SHF_GNU_RETAIN.
Recognize SHF_GNU_RETAIN and SHF_GNU_MBIND only for supported OSABIs.
* testsuite/binutils-all/readelf.exp: Run new tests.
Don't run run_dump_test when there isn't an assembler available.
* testsuite/lib/binutils-common.exp (supports_gnu_osabi): Adjust
comment.
* testsuite/binutils-all/readelf-maskos-1a.d: New test.
* testsuite/binutils-all/readelf-maskos-1b.d: New test.
* testsuite/binutils-all/readelf-maskos.s: New test.
* testsuite/binutils-all/retain1.s: New test.
* testsuite/binutils-all/retain1a.d: New test.
* testsuite/binutils-all/retain1b.d: New test.
gas/ChangeLog:
* NEWS: Announce SHF_GNU_RETAIN support.
* config/obj-elf.c (obj_elf_change_section): Merge SHF_GNU_RETAIN bit
between section declarations.
(obj_elf_parse_section_letters): Handle 'R' flag.
Handle numeric flag values within the SHF_MASKOS range.
(obj_elf_section): Validate SHF_GNU_RETAIN usage.
* doc/as.texi: Document 'R' flag to .section directive.
* testsuite/gas/elf/elf.exp: Run new tests.
* testsuite/gas/elf/section10.d: Unset SHF_GNU_RETAIN bit.
* testsuite/gas/elf/section10.s: Likewise.
* testsuite/gas/elf/section22.d: New test.
* testsuite/gas/elf/section22.s: New test.
* testsuite/gas/elf/section23.s: New test.
* testsuite/gas/elf/section23a.d: New test.
* testsuite/gas/elf/section23b.d: New test.
* testsuite/gas/elf/section23b.err: New test.
* testsuite/gas/elf/section24.l: New test.
* testsuite/gas/elf/section24.s: New test.
* testsuite/gas/elf/section24a.d: New test.
* testsuite/gas/elf/section24b.d: New test.
include/ChangeLog:
* elf/common.h (SHF_GNU_RETAIN): Define.
ld/ChangeLog:
* NEWS: Announce support for SHF_GNU_RETAIN.
* ld.texi (garbage collection): Document SHF_GNU_RETAIN.
(Output Section Discarding): Likewise.
* testsuite/ld-elf/elf.exp: Run new tests.
* testsuite/ld-elf/retain1.s: New test.
* testsuite/ld-elf/retain1a.d: New test.
* testsuite/ld-elf/retain1b.d: New test.
* testsuite/ld-elf/retain2.d: New test.
* testsuite/ld-elf/retain2.ld: New test.
* testsuite/ld-elf/retain2.map: New test.
* testsuite/ld-elf/retain3.d: New test.
* testsuite/ld-elf/retain3.s: New test.
* testsuite/ld-elf/retain4.d: New test.
* testsuite/ld-elf/retain4.s: New test.
* testsuite/ld-elf/retain5.d: New test.
* testsuite/ld-elf/retain5.map: New test.
* testsuite/ld-elf/retain5lib.s: New test.
* testsuite/ld-elf/retain5main.s: New test.
* testsuite/ld-elf/retain6a.d: New test.
* testsuite/ld-elf/retain6b.d: New test.
* testsuite/ld-elf/retain6lib.s: New test.
* testsuite/ld-elf/retain6main.s: New test.
The purpose of the test is to look for entries of "foo" and
"bar" in the ".got". The old "objdump -s ..." has been replaced
with "objdump -D ..." to inspect the final executable. A sample
output looks like:
------------------------8<------------------------
$ arc-elf32-objdump -D -j .got tls_ie-01
tls_ie-01: file format elf32-littlearc
Disassembly of section .got:
00002110 <_GLOBAL_OFFSET_TABLE_>:
...
211c: 08 00 00 00 .word 0x00000008
2120: 0c 00 00 00 .word 0x0000000c
------------------------>8------------------------
ld/
* testsuite/ld-arc/tls_ie-01.d: Update the dump command.
bfdlink.h has
/* Symbol will be converted from absolute to section-relative. Set for
symbols defined by a script from "dot" (also SEGMENT_START or ORIGIN)
outside of an output section statement. */
unsigned int rel_from_abs : 1;
linker.c has
.{* Return TRUE if the symbol described by a linker hash entry H
. is going to be absolute. Linker-script defined symbols can be
. converted from absolute to section-relative ones late in the
. link. Use this macro to correctly determine whether the symbol
. will actually end up absolute in output. *}
.#define bfd_is_abs_symbol(H) \
. (((H)->type == bfd_link_hash_defined \
. || (H)->type == bfd_link_hash_defweak) \
. && bfd_is_abs_section ((H)->u.def.section) \
. && !(H)->rel_from_abs)
.
Set rel_from_abs to 1 for __ehdr_start which will be converted from
absolute to section-relative in assign_file_positions_for_load_sections.
PR ld/26869
* ldelf.c (ldelf_before_allocation): Set rel_from_abs to 1 for
__ehdr_start.
* testsuite/ld-i386/i386.exp: Run pr26869.
* testsuite/ld-i386/pr26869.d: New file.
* testsuite/ld-i386/pr26869.s: Likewise.
* lexsup.c (parse_args): Add more checks of the mapfile. If it is
a directory use the basename of the output file as the file
component. If the % character is present, replace it with the
full output filepath.
* testsuite/ld-scripts/map-address.exp: Add test of %
functionality.
* ld.texi: Document the new behaviour.
This patch reverts most of git commit 1e3b96fd6c, so IR symbols are
again not marked def_regular or ref_regular. That should be enough to
stop IR symbols from becoming dynamic. To mark as-needed shared
libraries referenced by IR symbols, use the referencing BFD rather
than the ref flags.
bfd/
PR 15146
PR 26314
PR 26530
PR 26806
* elflink.c (elf_link_add_object_symbols): Don't set def/ref flags
for plugin syms. Do allow plugin syms to mark as-needed libs.
ld/
PR 26806
* testsuite/ld-plugin/lto-19.h,
* testsuite/ld-plugin/lto-19a.c,
* testsuite/ld-plugin/lto-19b.c,
* testsuite/ld-plugin/lto-19c.c: New test.
* testsuite/ld-plugin/pr26806.c,
* testsuite/ld-plugin/pr26806.d: New test.
* testsuite/ld-plugin/lto.exp: Run them.
PR 26626
* ldmain.c (undefined_symbol): Use the keyword undefined-symbol
when invoking the error handling script for undefined symbols.
* ld.texi: Update documentation.
Make plug_opt available to all linker tests. Skip bootstrap tests when
linker is compiled with -fprofile-generate=.
* testsuite/ld-plugin/lto.exp (plug_opt): Moved to ...
* testsuite/config/default.exp (plug_opt): Here. New.
* testsuite/ld-bootstrap/bootstrap.exp: Skip when linker is
compiled with -fprofile-generate=.
alpha-dec-vms always loads a number of libraries, -limagelib,
-lstarlet, and -lsys$public_vectors. When running the ld testsuite
without a full cross-build environment, those libraries are missing
and cause fails. This patch provides dummies, and tidies default.exp
a little.
* testsuite/config/default.exp: Provide dummy libraries for
alpha-dec-vms.
(compiler_supports): New proc. Use it for compiler tests.
* testsuite/lib/ld-lib.exp (default_ld_assemble): Don't die if
subdir not set.
(run_ld_link_tests): Pass LDFLAGS to ld.
* testsuite/ld-checks/checks.exp (section_check): Likewise.
* testsuite/ld-scripts/assert.exp: Likewise.
* testsuite/ld-scripts/extern.exp: Likewise.
* testsuite/ld-scripts/log2.exp: Likewise.
* testsuite/ld-scripts/map-address.exp: Likewise.
* testsuite/ld-scripts/script.exp: Likewise.
* testsuite/ld-scripts/sizeof.exp: Likewise.
I wanted to write a linker script like this:
PROVIDE(mem_origin = 0x1000);
PROVIDE(mem_length = 0x1000);
MEMORY
{
REGION : ORIGIN = mem_origin, LENGTH = mem_length
}
....
Then when I link using this script I can optionally supply:
--defsym=mem_origin=..... --defsym=mem_length=....
to override the defaults.
And though passing `--defsym' does work, if I remove the use of
`--defsym' and just rely on the defaults I get an error:
ld-new: invalid origin for memory region REGION
Interestingly, if I make the above error non-fatal and dump a linker
map file I see that (a) REGION has origin 0x0, and length 0xffff...,
and (b) the symbol from the PROVIDE is provided.
An examination of ldlang.c:lang_process shows us what the issue is,
the origin and length of all memory regions are set as a result of a
single call to lang_do_memory_regions, this call is done after calling
open_input_bfds.
During the open_input_bfds call provide statements can be converted to
provided statements if we know that the assigned symbol is needed, but
for symbols that are only used in the memory regions we are unaware
that we need these symbols.
What I propose in this patch is to make two calls to
lang_do_memory_regions, in the first call we process the expressions
for the origin and length fields of each region, however, errors,
especially undefined symbols, will be ignored. The origin and length
values are not updated. However, by evaluating the expressions any
symbols we need will be added to the symbol table.
Now when we call open_input_bfds, when we process the provide
statements, we will see that the assigned symbol is needed add its new
value to the symbol table.
Finally we reach the original call to lang_do_memory_regions, in
this (now second) call we again process the expressions, and this time
update the origin and length values. Any errors encountered now are
reported to the user.
ld/ChangeLog:
* ldlang.c (lang_process): Add extra call to
lang_do_memory_regions, and pass parameter.
(lang_do_memory_regions): Add parameter, only define origin and
length when requested. Reindent.
* testsuite/ld-scripts/provide-10.d: New file.
* testsuite/ld-scripts/provide-10.map: New file.
* testsuite/ld-scripts/provide-11.d: New file.
* testsuite/ld-scripts/provide-11.map: New file.
* testsuite/ld-scripts/provide-12.d: New file.
* testsuite/ld-scripts/provide-12.map: New file.
* testsuite/ld-scripts/provide-9.d: New file.
* testsuite/ld-scripts/provide-9.map: New file.
* testsuite/ld-scripts/provide-9.t: New file.
The ordering of command line options --defsym and -T is important,
however, the description of --defsym in the manual doesn't mention
this.
This commit adds more text to the description of --defsym to try and
explain this ordering requirement.
ld/ChangeLog:
* ld.texi (Options): Extend the description of --defsym.
PR 26626
* ldmain.c (undefined_symbol): If an error handlign script is
available, call it.
* ldfile.c (error_handling_script): Declare.
(ldfile_open_file): If a library cannot be found and an error
handling script is available, call it.
* ldmain.h (error_handling_script): Prototype.
* ldlex.h (OPTION_ERROR_HANDLING_SCRIPT): Define.
* lexsup.c (ld_options): Add --error-handling-script.
(parse_args): Add support for --errror-handling-script.
* ld.texi: Document the new feature.
* configure.ac: Add --error-handling-script option to disable
support for the new feature.
* NEWS: Mention the new feature.
* config.in: Regenerate.
* configure: Regenerate.
For the ifunc symbol, which is referenced by GOT rather than PLT relocs,
we should add the dynamic reloc (usually IRELATIVE) into the .rel.iplt
when generating the static executable. But if we use riscv_elf_append_rela
to add the dynamic relocs into .rela.iplt, this may cause the overwrite
problem.
The reason is that we don't handle the `reloc_index` of .rela.iplt, but
the riscv_elf_append_rela adds the relocs to the place that are calculated
from the reloc_index (in seqential). Therefore, we may overwrite the
dynamic relocs when the `reloc_index` of .rela.iplt isn't handled correctly.
One solution is that we can add these dynamic relocs (GOT ifunc) from
the last of .rela.iplt section. But I'm not sure if it is the best way.
bfd/
* elfnn-riscv.c (riscv_elf_link_hash_table): Add last_iplt_index.
(riscv_elf_size_dynamic_sections): Initialize the last_iplt_index.
(riscv_elf_relocate_section): Use riscv_elf_append_rela.
(riscv_elf_finish_dynamic_symbol): If the use_elf_append_rela is
false, then we should add the dynamic relocs from the last of
the .rela.iplt, and don't use the riscv_elf_append_rela to add.
ld/
* testsuite/ld-riscv-elf/ifunc-plt-got-overwrite.s: New testcase.
* testsuite/ld-riscv-elf/ifunc-plt-got-overwrite.d: Likewise.
* testsuite/ld-riscv-elf/ifunc-plt-got-overwrite-exe.rd: Likewise.
* testsuite/ld-riscv-elf/ifunc-plt-got-overwrite-pic.rd: Likewise.
* testsuite/ld-riscv-elf/ifunc-plt-got-overwrite-pie.rd: Likewise.
* testsuite/ld-riscv-elf/ld-riscv-elf.exp: Updated.
