2016-12-01 Cary Coutant <ccoutant@gmail.com>
Igor Kudrin <ikudrin@accesssoftek.com>
PR gold/20717
* script-sections.cc (Script_sections): Set *keep to false when
no match.
2016-12-11 Igor Kudrin <ikudrin@accesssoftek.com>
PR gold/20717
* testsuite/Makefile.am (pr20717): New test.
* testsuite/Makefile.in: Regenerate.
* testsuite/pr20717.c: New test source file.
* testsuite/pr20717.sh: New test script.
* testsuite/pr20717.t: New test linker script.
PR 16711 noted that gold allocates file space for BSS sections when using
a linker script. I've fixed that by rewriting set_section_addresses and
set_section_list_addresses to track the file offset separate from the
current virtual address, so that BSS sections do not move the file offset.
Now, if a series of BSS sections come at the end of a segment, we do not
allocate file space; but if a script forces them into the middle of a
segment, we will still allocate file space (matching Gnu ld behavior).
I've also added a warning when that happens.
That exposed another problem where orphan .bss sections were sometimes
placed in the middle of a segment. For example, if the script mentions
the .got section, but both .data and .bss are orphans, gold would put
both .data and .bss in front of .got. I've fixed that by ensuring that
orphan BSS sections are always placed after all other allocated sections.
It also exposed a problem where the SUBALIGN property is not handled
properly. The ld manual clearly states that it should override input section
alignment, whether greater or less than the given alignment, but gold would
only increase an input section's alignment. Gold would also place the output
section based on its original alignment before the SUBALIGN property took
effect, leading to a misaligned output section (where the input section
was properly aligned in memory, but was not aligned relative to the start
of the section), in violation of the ELF/gABI spec. I've fixed that by
making sure that the SUBALIGN property overrides the internal alignment of
the input sections as well as the external alignment of the output section.
This affected the behavior of script_test_2, which was written to expect
a misaligned section.
The net effect is, I think, improved compatibility with the BFD linker.
There are still cases where orphan placement differs, but the differences
should be rarer and less important. ALIGN and SUBALIGN behavior is closer,
but still not an exact match -- I still found cases where ld would create
a misaligned output section, and where gold will properly align it.
gold/
PR gold/16711
* output.cc (Output_section::set_final_data_size): Calculate data size
based on relative offset rather than file offset.
(Output_segment::set_section_addresses): Track file offset separately
from address offset.
(Output_segment::set_section_list_addresses): Add pfoff parameter.
Track file offset separately. Don't move file offset for BSS
sections.
* output.h (Output_segment::set_section_list_addresses): Add pfoff
parameter.
* script-sections.cc (Orphan_section_placement): Add PLACE_LAST_ALLOC.
(Orphan_section_placement::Orphan_section_placement): Initialize it.
(Orphan_section_placement::output_section_init): Track last allocated
section.
(Orphan_section_placement::find_place): Place BSS after last allocated
section.
(Output_section_element_input::set_section_addresses): Always override
input section alignment when SUBALIGN is specified.
(Output_section_definition::set_section_addresses): Override alignment
of output section when SUBALIGN is specified.
* testsuite/Makefile.am (script_test_15a, script_test_15b)
(script_test_15c): New test cases.
* testsuite/Makefile.in: Regenerate.
* testsuite/script_test_2.cc: Adjust expected layout.
* testsuite/script_test_15.c: New source file.
* testsuite/script_test_15a.sh: New shell script.
* testsuite/script_test_15a.t: New linker script.
* testsuite/script_test_15b.sh: New shell script.
* testsuite/script_test_15b.t: New linker script.
* testsuite/script_test_15c.sh: New shell script.
* testsuite/script_test_15c.t: New linker script.
A branch in a non-exec section that needs a stub can lead to this
assertion.
* powerpc.cc (Powerpc_relobj::stub_table): Return NULL rather
then asserting.
Adds a new option, defaulting to off, that allows a group of stubs to
serve multiple output sections. Prior to this patch powerpc gold
allowed this unconditionally, which is a little unsafe with clever
code that discards/reuses sections at runtime.
* options.h (--stub-group-multi): New PowerPC option.
* powerpc.cc (Stub_control): Add multi_os_ var and param
to constructor. Sort start_ var later. Comment State.
(Stub_control::can_add_to_stub_group): Heed multi_os_.
(Target_powerpc::group_sections): Update.
Gold attaches stubs to an existing section in contrast to ld.bfd which
inserts a new section for stubs. If we want stubs before branches,
then the stubs must be added to the previous section. Adding to the
previous section is a disaster if there is a large gap between the
previous section and the group.
PR gold/20878
* powerpc.cc (Stub_control): Replace stubs_always_before_branch_
with stubs_always_after_branch_, group_end_addr_ with
group_start_addr_.
(Stub_control::can_add_to_stub_group): Rewrite to suit scanning
sections by increasing address.
(Target_powerpc::group_sections): Scan that way. Delete corner
case.
* options.h (--stub-group-size): Update help string.
Some more debug output, and a little hardening.
* powerpc.cc (Stub_table_owner): Provide constructor.
(Powerpc_relobj::set_stub_table): Resize fill with -1.
(Target_powerpc::Branch_info::make_stub): Provide target debug
output on returning false.
If the default starting address is less than the new ABI page size,
we end up misaligning the file header, causing an internal error.
gold/
PR gold/20834
* target.h (Target::default_text_segment_address): Bump default
start address up to ABI page size.
gold/
PR gold/18989
* options.cc (General_options::object_format_to_string): New function.
(General_options::copy_from_posdep_options): New function.
(General_options::parse_push_state): New function.
(General_options::parse_pop_state): New function.
* options.h (--push-state, --pop-state): New options.
(General_options::object_format_to_string): New method.
(General_options::set_incremental_disposition): New method.
(General_options::copy_from_posdep_options): New method.
(General_options::options_stack_): New data member.
gold/
PR gold/20346
* options.cc (One_option::print): Print "(default)" when appropriate.
* options.h: Clean up and re-sort options.
(One_option::is_default): New data member.
(One_option::One_option): Add is_default parameter; adjust all calls.
(DEFINE_var): Add is_default__ parameter; adjust all calls.
(DEFINE_bool): Set is_default based on default_value__.
(DEFINE_bool_ignore): New macro.
(--no-eh-frame-hdr): New option.
(--enable-new-dtags): Remove mention of DT_FLAGS.
2016-11-21 Igor Kudrin <ikudrin@accesssoftek.com>
gold/
* layout.cc: Include windows.h and rpcdce.h (for MinGW32).
(Layout::create_build_id): Generate uuid using UuidCreate().
This fix keeps the section with the highest alignment when folding sections with ICF.
PR gold/17704
* icf.cc (match_sections): Add new parameter section_addraligns.
Check section alignment and keep the section with the strictest
alignment.
(find_identical_sections): New local variable section_addraligns.
Store each section's alignment.
* testsuite/pr17704a_test.s: New file.
* testsuite/Makefile.am (pr17704a_test): New test.
* testsuite/Makefile.in: Regenerate.
When we see an unversioned symbol reference in a shared library, followed
by a default definition of the symbol in another shared library, we were
treating them as separate symbols. That should only happen when both are
definitions.
gold/
PR gold/20238
* symtab.cc (Symbol_table::define_default_version): Check that
unversioned symbol is defined.
This patch adds a little more debug output, and replaces two variables
with one, tracking current max group size by group_size_ rather than
by has14_.
* powerpc.cc (class Stub_control): Delete stub14_group_size_
and has14_. Add group_size_.
(Stub_control::can_add_to_stub_group): Adjust to suit. Print
debug info when switching to adding sections before stubs.
This patch rewrites the rather obscure can_add_to_stub_group, fixing
a problem with the handling of sections containing conditional
external branches. When a section group contains any such section,
the group size needs to be limited to a much smaller size than groups
with only non-conditional external branches.
PR 20523
* powerpc.cc (class Stub_control): Add has14_. Comment owner_.
(Stub_control::can_add_to_stub_group): Correct grouping of
sections containing 14-bit external branches. When returning
false, set state_ to reflect the fact that we have one section
for the next group. Rewrite most of function for clarity.
Add and expand comments.
(Target_powerpc::do_relax): Print stub group size retry in hex.
gold/ChangeLog
2016-08-26 Han Shen <shenhan@google.com>
* powerpc.cc (Stub_table::min_size_threshold_): New member to
limit size.
(Stub_table::set_min_size_threshold): New member function.
(Stub_table::set_address_and_size): Add code to only allow size
increase.
(Target_powerpc::do_relax): Add code to record last size.
Build the test objects with the in-tree assembler. Also fix some cascading
error messages caused by not resetting the skip_call_tls_get_addr_ flag
after printing the error.
gold/
* i386.cc (Target_i386): Reset skip_call_tls_get_addr_ after printing
error message.
* testsuite/Makefile.am (pr20216a): Add missing dependencies.
(pr20308a): Add -Bgcctestdir/ to compile rules.
* testsuite/Makefile.in: Regenerate.
With some versions of gas, the call to tls_get_addr uses a GOTPCREL
relocation instead of a GOTPCRELX relocation. We should allow for that
when skip_call_tls_get_addr_ is true. We should also build the test
objects with the in-tree assembler.
This patch also fixes some cascading error messages caused by not
resetting the skip_call_tls_get_addr_ flag after printing the error.
gold/
PR gold/20216
* x86_64.cc (Target_x86_64::Relocate::relocate): Add check for
R_X86_64_GOTPCREL. Reset skip_call_tls_get_addr_ after printing
error message.
* testsuite/Makefile.am (pr20216_gd.o): Add -Bgcctestdir/.
(pr20216_ld.o): Likewise.
* testsuite/Makefile.in: Regenerate.
2016-08-10 James Clarke <jrtc27@jrtc27.com>
gold/
PR gold/20443
* symtab.cc (Symbol_table::add_from_relobj): Handle NULL symbols,
which will be present for STT_SPARC_REGISTER.
(Symbol_table::add_from_pluginobj): Likewise.
(Symbol_table::add_from_dynobj): Likewise.
(Symbol_table::add_from_incrobj): Removed dead code.
The fall-through in Target_sparc::Relocate::relocate for
R_SPARC_GOTDATA_OP_LOX10 is currently R_SPARC_GOT13, but should
clearly be R_SPARC_GOT10. GCC has been seen to emit a sethi/xor
rather than a sethi/or sequence to load a 32-bit immediate, but
if R_SPARC_GOT13 is used then bits 10-12 get zeroed out as both
the sethi and xor immediates contain them.
gold/
PR gold/20442
* sparc.cc (Target_sparc::Relocate::relocate): R_SPARC_GOTDATA_OP_LOX10
should fall back on R_SPARC_GOT10, not R_SPARC_GOT13.
With GCC 6 when not using -static-libstdc++, the operator delete(void*)
function is defined in the shared C++ support library, rather than in
the main program. The test script is too aggressive in checking for
this symbol's presence among the exported symbols. This patch removes
the check for that symbol.
gold/
PR gold/20310
* testsuite/dynamic_list.sh: Remove check for _ZdlPv.
There are extensions to 386 psABI:
https://groups.google.com/forum/#!topic/ia32-abi/awsRSvJOJfs
to call tls_get_addr via GOT:
call *___tls_get_addr@GOT(%reg)
where EBX register isn't required as GOT base.
Since direct call is 4-byte long and indirect call, is 5-byte long, the
extra one byte must be handled properly.
For general dynamic model, 7-byte lea instruction before call
instruction is replaced by 6-byte one to make room for indirect call.
For local dynamic model, we simply use 5-byte indirect call.
TLS linker optimization is updated to recognize new instruction
patterns. For local dynamic model to local exec model transition,
we generate a 6-byte lea instruction as nop, instead of a 1-byte nop
plus a 4-byte lea instruction.
PR gold/20308
* i386.cc (Target_i386::Relocate::relocate): Allow
R_386_GOT32X relocation against ___tls_get_addr.
(Target_i386::Relocate::tls_gd_to_ie): Support indirect
call to __tls_get_addr.
(Target_i386::Relocate::tls_gd_to_le): Likewise.
(Target_i386::Relocate::tls_ld_to_le): Likewise.
* testsuite/Makefile.am (check_PROGRAMS): Add pr20308a_test,
pr20308b_test, pr20308c_test, pr20308d_test, pr20308e_test.
(pr20308a_test_SOURCES): New.
(pr20308a_test_DEPENDENCIES): Likewise.
(pr20308a_test_CFLAGS): Likewise.
(pr20308a_test_LDFLAGS): Likewise.
(pr20308a_test_LDADD): Likewise.
(pr20308b_test_SOURCES): Likewise.
(pr20308b_test_DEPENDENCIES): Likewise.
(pr20308b_test_CFLAGS): Likewise.
(pr20308b_test_LDFLAGS): Likewise.
(pr20308b_test_LDADD): Likewise.
(pr20308c_test_SOURCES): Likewise.
(pr20308c_test_DEPENDENCIES): Likewise.
(pr20308c_test_CFLAGS): Likewise.
(pr20308c_test_LDFLAGS): Likewise.
(pr20308c_test_LDADD): Likewise.
(pr20308d_test_SOURCES): Likewise.
(pr20308d_test_DEPENDENCIES): Likewise.
(pr20308d_test_CFLAGS): Likewise.
(pr20308d_test_LDFLAGS): Likewise.
(pr20308d_test_LDADD): Likewise.
(pr20308e_test_SOURCES): Likewise.
(pr20308e_test_DEPENDENCIES): Likewise.
(pr20308e_test_CFLAGS): Likewise.
(pr20308e_test_LDFLAGS): Likewise.
(pr20308e_test_LDADD): Likewise.
(pr20308a.so): Likewise.
(pr20308b.so): Likewise.
(pr20308_gd.o): Likewise.
(pr20308_ld.o): Likewise.
(MOSTLYCLEANFILES): Add pr20308a.so pr20308b.so.
* testsuite/Makefile.in: Regenerated.
* testsuite/pr20308_def.c: New file.
* testsuite/pr20308_gd.S: Likewise.
* testsuite/pr20308_ld.S: Likewise.
* testsuite/pr20308_main.c: Likewise.
There are extensions to x86-64 psABI:
https://groups.google.com/forum/#!topic/x86-64-abi/de5_KnLHxtI
to call tls_get_addr via GOT:
call *__tls_get_addr@GOTPCREL(%rip)
Since direct call is 4-byte long and indirect call, is 5-byte long, the
extra one byte must be handled properly.
For general dynamic model, one 0x66 prefix before call instruction is
removed to make room for indirect call. For local dynamic model, we
simply use 5-byte indirect call.
TLS linker optimization is updated to recognize new instruction
patterns. For local dynamic model to local exec model transition, we
generate 4 0x66 prefixes, instead of 3, before mov instruction in 64-bit
and generate a 5-byte nop, instead of 4-byte, before mov instruction in
32-bit.
PR gold/20216
* configure.ac (DEFAULT_TARGET_X86_64_OR_X32): New
AM_CONDITIONAL.
* configure: Regenerated.
* x86_64.cc (Target_x86_64<size>::Relocate::relocate): Allow
R_X86_64_GOTPCRELX relocation against __tls_get_addr.
(Target_x86_64<size>::Relocate::tls_gd_to_ie): Support indirect
call to __tls_get_addr.
(Target_x86_64<size>::Relocate::tls_gd_to_le): Likewise.
(Target_x86_64<size>::Relocate::tls_ld_to_le): Likewise.
* testsuite/Makefile.am (check_PROGRAMS): Add pr20216a_test,
pr20216b_test, pr20216c_test, pr20216d_test, pr20216e_test.
(pr20216a_test_SOURCES): New.
(pr20216a_test_DEPENDENCIES): Likewise.
(pr20216a_test_CFLAGS): Likewise.
(pr20216a_test_LDFLAGS): Likewise.
(pr20216a_test_LDADD): Likewise.
(pr20216b_test_SOURCES): Likewise.
(pr20216b_test_DEPENDENCIES): Likewise.
(pr20216b_test_CFLAGS): Likewise.
(pr20216b_test_LDFLAGS): Likewise.
(pr20216b_test_LDADD): Likewise.
(pr20216c_test_SOURCES): Likewise.
(pr20216c_test_DEPENDENCIES): Likewise.
(pr20216c_test_CFLAGS): Likewise.
(pr20216c_test_LDFLAGS): Likewise.
(pr20216c_test_LDADD): Likewise.
(pr20216d_test_SOURCES): Likewise.
(pr20216d_test_DEPENDENCIES): Likewise.
(pr20216d_test_CFLAGS): Likewise.
(pr20216d_test_LDFLAGS): Likewise.
(pr20216d_test_LDADD): Likewise.
(pr20216e_test_SOURCES): Likewise.
(pr20216e_test_DEPENDENCIES): Likewise.
(pr20216e_test_CFLAGS): Likewise.
(pr20216e_test_LDFLAGS): Likewise.
(pr20216e_test_LDADD): Likewise.
(pr20216a.so): Likewise.
(pr20216b.so): Likewise.
(pr20216_gd.o): Likewise.
(pr20216_ld.o): Likewise.
(MOSTLYCLEANFILES): Add pr20216a.so pr20216b.so.
* testsuite/Makefile.in: Regenerated.
* testsuite/pr20216_def.c: New file.
* testsuite/pr20216_gd.S: Likewise.
* testsuite/pr20216_ld.S: Likewise.
* testsuite/pr20216_main.c: Likewise.
PowerPC gcc aligns char arrays to a word which results in the test
failing since it expects a char alignment. As the test already uses
gcc attributes to assign variables to sections, we may as well use an
attribute to align too.
* testsuite/plugin_layout_with_alignment.c: Explicitly align all
variables.
A target like powerpc64 that is PIC by default doesn't need copy relocs.
* testsuite/Makefile.am (copy_test_protected): Disable for powerpc.
* testsuite/Makefile.in: Regenerate.
According to "ELF for the ARM(r) 64-bit Architecture (AArch64)",
this relocation can be used "to prevent removal of sections that
might otherwise appear to be unused."
gold/
* aarch64-reloc.def (NONE): New relocation.
* aarch64.cc (Target_aarch64::Scan::local): Handle R_AARCH64_NONE.
(Target_aarch64::Scan::global): Likewise.
* testsuite/Makefile.am (aarch64_reloc_none): New test.
* testsuite/Makefile.in: Regenerate.
* testsuite/aarch64_reloc_none.s: New test source file.
* testsuite/aarch64_reloc_none.sh: New test script.
