Commit Graph

5768 Commits

Author SHA1 Message Date
Rui Ueyama e6c24299d2 Use a class instead of lambda-based callbacks to organize garbage collector.
lld's mark-sweep garbage collector was written in the visitor pattern.
There are functions that traverses a given graph, and the functions calls
callback functions to dispatch according to node type.

The code was originaly pretty simple, and lambdas worked pretty
well. However, as we add more features to the garbage collector, that became
more like a callback hell. We now have a callback function that wraps
another callback function, for example. It is not easy to follow the flow of
the control.

This patch rewrites it as a regular class. What was once a lambda is now a
regular class member function. I think this change fixes the readability
issue.

No functionality change intended.

Differential Revision: https://reviews.llvm.org/D59800

llvm-svn: 356966
2019-03-25 23:28:47 +00:00
Rui Ueyama d2e0ed7755 Simplify PltSection.
Previously, `Entries` contains pairs of symbols and their indices.
The indices are always 0, x, 2x, 3x, ..., where x is the size of
relocation entry. We didn't have to store that values because we can
compute them when we consume them.

llvm-svn: 356812
2019-03-22 21:17:25 +00:00
Fangrui Song dc6c0cf94d [DWARF] Refactor RelocVisitor and fix computation of SHT_RELA-typed relocation entries
Summary: This is the lld-side change of D57939

Reviewers: echristo, dblaikie, ruiu, espindola

Reviewed By: echristo, ruiu

Subscribers: jdoerfert, emaste, arichardson, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D57940

llvm-svn: 356730
2019-03-22 02:43:17 +00:00
Peter Smith 54dab70bb7 [ELF][ARM] Revert Redesign of .ARM.exidx handling to use a SyntheticSection
There is a reproducible buildbot failure (segfault) on the 2 stage
clang-cmake-armv8-lld bot. Reverting while I investigate.

Differential Revision: https://reviews.llvm.org/D59216

llvm-svn: 356684
2019-03-21 17:17:54 +00:00
Peter Smith d3511a214e [ELF][ARM] Redesign of .ARM.exidx handling to use a SyntheticSection
Instead of creating extra Synthetic .ARM.exidx sections to account for
gaps in the table, create a single .ARM.exidx SyntheticSection that can
derive the contents of the gaps from a sorted list of the executable
InputSections. This has the benefit of moving the ARM specific code for
SyntheticSections in SHF_LINK_ORDER processing and the table merging code
into the ARM specific SyntheticSection. This also makes it easier to create
EXIDX_CANTUNWIND table entries for executable InputSections that don't
have an associated .ARM.exidx section.

Fixes pr40277

Differential Revision: https://reviews.llvm.org/D59216

llvm-svn: 356666
2019-03-21 14:06:40 +00:00
Fangrui Song 1092fc9057 [ELF] Allow sh_entsize to be unrelated to sh_addralign and not a power of 2
Summary:
This implements Rui Ueyama's idea in PR39044.
I've checked that ld.bfd and gold do not have the power-of-2 requirement
and do not require sh_entsize to be a multiple of sh_align.

Now on the updated test merge-entsize.s, all the 3 linkers happily
create .rodata that is not 3-byte aligned.

This has a use case in Linux arch/x86/crypto/sha512-avx2-asm.S
It uses sh_entsize of 640, which is not a power of 2.
See https://github.com/ClangBuiltLinux/linux/issues/417

Reviewers: ruiu, espindola

Reviewed By: ruiu

Subscribers: nickdesaulniers, E5ten, emaste, arichardson, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D59478

llvm-svn: 356428
2019-03-18 23:49:18 +00:00
Siva Chandra 1915e2be93 [ELF] Emit weak-undef symbols in .dynsym of a PIE binary only if linked against shared libs.
Reviewers: espindola

Subscribers: emaste, arichardson, MaskRay, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D59275

llvm-svn: 356374
2019-03-18 15:32:57 +00:00
Fangrui Song 6778b53e95 [ELF] De-virtualize findOrphanPos, excludeLibs and handleARMTlsRelocation
llvm-svn: 356331
2019-03-17 13:53:42 +00:00
Fangrui Song f9695e166b [ELF] Delete unused forward declarations and unused DynamicReloc::getInputSec(). NFC
llvm-svn: 356239
2019-03-15 07:16:39 +00:00
Fangrui Song 76a7ecb3ae [ELF] De-template wrapSymbols, isReserved and addGotEntry. NFC
llvm-svn: 356237
2019-03-15 06:58:23 +00:00
Fangrui Song e8710ef1fb [ELF] Split RW PT_LOAD on the PT_GNU_RELRO boundary
Summary:
Based on Peter Collingbourne's suggestion in D56828.

Before D56828: PT_LOAD(.data PT_GNU_RELRO(.data.rel.ro .bss.rel.ro) .bss)
Old:           PT_LOAD(PT_GNU_RELRO(.data.rel.ro .bss.rel.ro) .data .bss)
New:           PT_LOAD(PT_GNU_RELRO(.data.rel.ro .bss.rel.ro)) PT_LOAD(.data. .bss)

The new layout reflects the runtime memory mappings.
By having two PT_LOAD segments, we can utilize the NOBITS part of the
first PT_LOAD and save bytes for .bss.rel.ro.

.bss.rel.ro is currently small and only used by copy relocations of
symbols in read-only segments, but it can be used for other purposes in
the future, e.g. if a relro section's statically relocated data is all
zeros, we can move it to .bss.rel.ro.

Reviewers: espindola, ruiu, pcc

Reviewed By: ruiu

Subscribers: nemanjai, jvesely, nhaehnle, javed.absar, kbarton, emaste, arichardson, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D58892

llvm-svn: 356226
2019-03-15 01:29:57 +00:00
Rui Ueyama cc8e4e839f Make a hack for LTO work only when you are actually doing LTO.
We allow an archive file without symbol table as a linker input as a
workaround for a very common error in LTO build. But that logic worked
even for an archive file containing non-bitcode files, which is not
expected. This patch limits that workaround to one that contains only
bitcode files.

Differential Revision: https://reviews.llvm.org/D59373

llvm-svn: 356186
2019-03-14 18:21:32 +00:00
Fangrui Song 07f8daf05e [ELF] Simplify RelRo, TLS, NOBITS section ranks and make RW PT_LOAD start with RelRo
Old: PT_LOAD(.data | PT_GNU_RELRO(.data.rel.ro .bss.rel.ro) | .bss)
New: PT_LOAD(PT_GNU_RELRO(.data.rel.ro .bss.rel.ro) | .data .bss)

The placement of | indicates page alignment caused by PT_GNU_RELRO. The
new layout has simpler rules and saves space for many cases.

Old size: roundup(.data) + roundup(.data.rel.ro)
New size: roundup(.data.rel.ro + .bss.rel.ro) + .data

Other advantages:

* At runtime the 3 memory mappings decrease to 2.
* start(PT_TLS) = start(PT_GNU_RELRO) = start(RW PT_LOAD). This
  simplifies binary manipulation tools.
  GNU strip before 2.31 discards PT_GNU_RELRO if its
  address is not equal to the start of its associated PT_LOAD.
  This has been fixed by https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=f2731e0c374e5323ce4cdae2bcc7b7fe22da1a6f
  But with this change, we will be compatible with GNU strip before 2.31
* Before, .got.plt (non-relro by default) was placed before .got (relro
  by default), which made it impossible to have _GLOBAL_OFFSET_TABLE_
  (start of .got.plt on x86-64) equal to the end of .got (R_GOT*_FROM_END)
  (https://bugs.llvm.org/show_bug.cgi?id=36555). With the new ordering, we
  can improve on this regard if we'd like to.

Reviewers: ruiu, espindola, pcc

Subscribers: emaste, arichardson, llvm-commits, joerg, jdoerfert

Differential Revision: https://reviews.llvm.org/D56828

llvm-svn: 356117
2019-03-14 03:47:45 +00:00
Simon Atanasyan 2e855675eb [mips] Remove redundant setup of less-significant bit. NFC
The less-significant bit for microMIPS symbols configured
in the `getSymVA` function. Do not need to setup it once again.

llvm-svn: 356058
2019-03-13 16:00:35 +00:00
Francis Visoiu Mistrih dd42236c6c Reland "[Remarks] Add -foptimization-record-passes to filter remark emission"
Currently we have -Rpass for filtering the remarks that are displayed as
diagnostics, but when using -fsave-optimization-record, there is no way
to filter the remarks while generating them.

This adds support for filtering remarks by passes using a regex.
Ex: `clang -fsave-optimization-record -foptimization-record-passes=inline`

will only emit the remarks coming from the pass `inline`.

This adds:

* `-fsave-optimization-record` to the driver
* `-opt-record-passes` to cc1
* `-lto-pass-remarks-filter` to the LTOCodeGenerator
* `--opt-remarks-passes` to lld
* `-pass-remarks-filter` to llc, opt, llvm-lto, llvm-lto2
* `-opt-remarks-passes` to gold-plugin

Differential Revision: https://reviews.llvm.org/D59268

Original llvm-svn: 355964

llvm-svn: 355984
2019-03-12 21:22:27 +00:00
Peter Collingbourne 8a28673a2e ELF: Don't add .dynamic strings to .dynstr early.
This does not appear to be necessary because StringTableSection does not
need to be finalized, which also means that we can remove the call to
finalizeSynthetic on .dynstr.

Differential Revision: https://reviews.llvm.org/D59240

llvm-svn: 355977
2019-03-12 20:58:34 +00:00
Francis Visoiu Mistrih 1d6c47ad2b Revert "[Remarks] Add -foptimization-record-passes to filter remark emission"
This reverts commit 20fff32b7d.

llvm-svn: 355976
2019-03-12 20:54:18 +00:00
Peter Collingbourne e2b8c40a77 ELF: Use bump pointer allocator for uncompressed section buffers. NFCI.
This shaves another word off SectionBase and makes it possible to clone a
section using the implicit copy constructor.

This basically reverts r311056, which removed the mutex in order to
make the code easier to understand. On balance I think it's probably more
straightforward to have a mutex here than to have an unusual copy constructor
in SectionBase.

Differential Revision: https://reviews.llvm.org/D59269

llvm-svn: 355966
2019-03-12 20:32:30 +00:00
Francis Visoiu Mistrih 20fff32b7d [Remarks] Add -foptimization-record-passes to filter remark emission
Currently we have -Rpass for filtering the remarks that are displayed as
diagnostics, but when using -fsave-optimization-record, there is no way
to filter the remarks while generating them.

This adds support for filtering remarks by passes using a regex.
Ex: `clang -fsave-optimization-record -foptimization-record-passes=inline`

will only emit the remarks coming from the pass `inline`.

This adds:

* `-fsave-optimization-record` to the driver
* `-opt-record-passes` to cc1
* `-lto-pass-remarks-filter` to the LTOCodeGenerator
* `--opt-remarks-passes` to lld
* `-pass-remarks-filter` to llc, opt, llvm-lto, llvm-lto2
* `-opt-remarks-passes` to gold-plugin

Differential Revision: https://reviews.llvm.org/D59268

llvm-svn: 355964
2019-03-12 20:28:50 +00:00
Peter Collingbourne 2da7b32684 ELF: Simplify. NFCI.
We don't need to take a slice of SectionCommands in addOrphanSections()
because it is not modified until the end of the function.

Differential Revision: https://reviews.llvm.org/D59239

llvm-svn: 355954
2019-03-12 19:19:23 +00:00
George Rimar 43b6689e64 [LLD][ELF] - Show symbols visibility in "undefined symbol..." error messages.
This teaches LLD to report visibility when showing undefined symbol errors
and fixes https://bugs.llvm.org/show_bug.cgi?id=40770.

Differential revision: https://reviews.llvm.org/D58577

llvm-svn: 355909
2019-03-12 11:10:29 +00:00
Peter Collingbourne b5a307fa24 ELF: Remove dead code. NFCI.
The Live bit is already set to false by SectionBase.

Differential Revision: https://reviews.llvm.org/D59052

llvm-svn: 355893
2019-03-12 02:18:03 +00:00
Eli Friedman 3751ae4a94 [ELF] Print a better error for an archive containing a non-ELF file.
Hopefully gives a more readable error message for the most obvious
mistake.

Differential Revision: https://reviews.llvm.org/D59170

llvm-svn: 355888
2019-03-12 01:24:39 +00:00
Rui Ueyama 033c4d2126 Include an archive file name in an error message for a corrupted file.
Differential Revision: https://reviews.llvm.org/D59212

llvm-svn: 355886
2019-03-12 00:24:34 +00:00
Rong Xu f92e59cbba [PGO] Add options for context-sensitive PGO
Add lld options for CSPGO (context-sensitive PGO).

Differential Revision: https://reviews.llvm.org/D56675

llvm-svn: 355876
2019-03-11 22:51:38 +00:00
Rui Ueyama 7fd99fc475 Fail early if an output file is not writable
Fixes https://bugs.llvm.org/show_bug.cgi?id=36478

Differential Revision: https://reviews.llvm.org/D43664

llvm-svn: 355834
2019-03-11 16:30:55 +00:00
Sam Clegg d15a4154a8 [WebAssembly] Don't mark lazy symbols as `IsUsedInRegularObj`
This matches the ELF does.  Update the comment in ELF/Symbols.h and
duplicate it in wasm/Symbols.h

This a followup on rL355580 and rL355577.

Differential Revision: https://reviews.llvm.org/D59075

llvm-svn: 355737
2019-03-08 21:10:48 +00:00
Peter Collingbourne dfbb9a793e ELF: Reduce the size of InputSectionBase by two words. NFCI.
- The Assigned bit was previously taking a word on its own. Move
  it into the bit fields in SectionBase.
- NumRelocations and AreRelocsRela were previously also taking up
  a word despite only using half of it. Move them into the alignment gap
  after SectionBase's fields.

Differential Revision: https://reviews.llvm.org/D59044

llvm-svn: 355622
2019-03-07 18:48:12 +00:00
Peter Collingbourne 5ee9abd4c8 ELF: De-template OutputSection::finalize() and MipsGotSection::build(). NFCI.
Differential Revision: https://reviews.llvm.org/D58810

llvm-svn: 355479
2019-03-06 03:07:57 +00:00
Peter Collingbourne 704dfd6e28 ELF: Extract a non-ELFT base class for VersionNeedSection.
We're going to need a separate VersionNeedSection for each partition, and
the partition data structure won't be templated.

With this the VersionTableSection class no longer needs ELFT, so detemplate it.

Differential Revision: https://reviews.llvm.org/D58808

llvm-svn: 355478
2019-03-06 03:07:48 +00:00
Peter Collingbourne 16d9a0acfd ELF: Change FileSize back to a uint64_t.
This lets us detect file size overflows when creating a 64-bit binary on
a 32-bit machine.

Differential Revision: https://reviews.llvm.org/D58840

llvm-svn: 355218
2019-03-01 18:53:41 +00:00
Peter Smith 22ce712c19 [ELF][ARM] Fix clang-armv7-linux-build-cache builds of LLD [NFC]
r355153 introduced a build failure on a build bot that uses clang natively
on an armv7-a machine. This a temporary fix to use size_t rather than
uint64_t.

llvm-svn: 355195
2019-03-01 10:52:25 +00:00
Peter Collingbourne 7fb9eabda5 ELF: Write .eh_frame_hdr explicitly after writing .eh_frame.
This lets us remove the special case from Writer::writeSections(), and also
fixes a bug where .eh_frame_hdr isn't necessarily written in the correct
order if a linker script moves .eh_frame and .eh_frame_hdr into the same
output section.

Differential Revision: https://reviews.llvm.org/D58795

llvm-svn: 355153
2019-02-28 23:11:35 +00:00
Alexey Lapshin 77fc1f6049 [DebugInfo] add SectionedAddress to DebugInfo interfaces.
That patch is the fix for https://bugs.llvm.org/show_bug.cgi?id=40703
   "wrong line number info for obj file compiled with -ffunction-sections"
   bug. The problem happened with only .o files. If object file contains
   several .text sections then line number information showed incorrectly.
   The reason for this is that DwarfLineTable could not detect section which
   corresponds to specified address(because address is the local to the
   section). And as the result it could not select proper sequence in the
   line table. The fix is to pass SectionIndex with the address. So that it
   would be possible to differentiate addresses from various sections. With
   this fix llvm-objdump shows correct line numbers for disassembled code.

   Differential review: https://reviews.llvm.org/D58194

llvm-svn: 354972
2019-02-27 13:17:36 +00:00
George Rimar cc19dc75fb [LLD][ELF] - Improve "sh_addralign is not a power of 2" diagnostics.
This patch removes the precompiled binary from inputs,
replacing it with a YAML. And teaches LLD to report a
section name in case of such error.

Differential revision: https://reviews.llvm.org/D58670

llvm-svn: 354959
2019-02-27 10:28:23 +00:00
Bill Wendling 01706bda5b Output ELF files after ThinLTO is run.
Summary:
The gold linker allowed you to output the ELF files after LTO was run. It did
it by using the 'obj-path' option. This replicates that behavior.

Reviewers: espindola, ruiu, MaskRay, pcc

Reviewed By: MaskRay, pcc

Subscribers: grimar, emaste, inglorion, arichardson, steven_wu, dexonsmith, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D56046

llvm-svn: 354917
2019-02-26 19:29:14 +00:00
Peter Smith 777e1cfdc3 [ELF][ARM] Accept and ignore -p and -no-pipleline-knowledge
The linux kernel uses an old flag -p/-no-pipeline-knowledge that is
accepted by bfd and gold but ignored by modern versions of them. The
original option is very old and is pre-ABI, it sometimes comes up in
code-bases that had support for pre ABI toolchains. The Linux kernel uses
it in 3 places in the ARM specific section.

Differential Revision: https://reviews.llvm.org/D58540

llvm-svn: 354769
2019-02-25 10:48:31 +00:00
Rui Ueyama c94dad9d97 Remove a function from header and move the implementation to a .cpp file. NFC.
llvm-svn: 354703
2019-02-22 23:59:43 +00:00
Peter Collingbourne f9232b0c00 ELF: Remove dead code. NFCI.
RelocationBaseSection is not used in -r links, so Config->Relocatable will
always be false.

Differential Revision: https://reviews.llvm.org/D58489

llvm-svn: 354607
2019-02-21 18:53:58 +00:00
Simon Atanasyan 14b09810a0 [mips] Put some MIPS-specific sections to separate segments
Three MIPS-specific sections `.reginfo`, `.MIPS.options`, and `.MIPS.abiflags`
are used by loader to read their contents and setup environment for running
a program. Loader looks up these data in the corresponding segments:
`PT_MIPS_REGINFO`, `PT_MIPS_OPTIONS`, and `PT_MIPS_ABIFLAGS` respectively.

This patch put these sections to separate segments like we do already
for ARM `SHT_ARM_EXIDX` section.

Differential Revision: http://reviews.llvm.org/D58381

llvm-svn: 354468
2019-02-20 14:47:02 +00:00
Fangrui Song eaa0db021f ELF: Remove field for .gdb_index in InStruct. NFC.
Summary: This field is unreferenced outside of createSyntheticSections.

Reviewers: ruiu, pcc, espindola, grimar

Reviewed By: grimar

Subscribers: grimar, emaste, arichardson, arphaman, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D58423

llvm-svn: 354449
2019-02-20 11:34:18 +00:00
Peter Collingbourne 7cca803d4c ELF: Remove field for .interp in InStruct. NFC.
This field is unreferenced outside of createSyntheticSections.

Differential Revision: https://reviews.llvm.org/D58422

llvm-svn: 354428
2019-02-20 02:32:53 +00:00
Tom Stellard 20332cc4ea ELF: Fix typo in --build-id option description
Reviewers: ruiu, espindola

Reviewed By: ruiu

Subscribers: emaste, arichardson, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D58265

llvm-svn: 354422
2019-02-20 01:40:35 +00:00
Rui Ueyama 3ae6726234 Sort enum members so that arch-dependent members are at the right place. NFC.
llvm-svn: 354405
2019-02-20 00:01:21 +00:00
Simon Atanasyan a6012a65d4 [mips] Create LA25 thunks for MIPS R6 code
MIPS R6 code uses the `R_MIPS_PC26_S2` relocation for calls which might
cross boundaries of non-PIC-to-PIC code. We need to create a LA25 thunks
for that case.

llvm-svn: 354312
2019-02-19 11:11:12 +00:00
Simon Atanasyan fae2a509fa [MIPS] Handle cross-mode (regular <-> microMIPS) jumps
The patch solves two tasks:

1. MIPS ABI allows to mix regular and microMIPS code and perform
cross-mode jumps. Linker needs to detect such cases and replace
jump/branch instructions by their cross-mode equivalents.

2. Other tools like dunamic linkers need to recognize cases when dynamic
table entries, e_entry field of an ELF header etc point to microMIPS
symbol. Linker should provide such information.

The first task is implemented in the `MIPS<ELFT>::relocateOne()` method.
New routine `fixupCrossModeJump` detects ISA mode change, checks and
replaces an instruction.

The main problem is how to recognize that relocation target is microMIPS
symbol. For absolute and section symbols compiler or assembler set the
less-significant bit of the symbol's value or sum of the symbol's value
and addend. And this bit signals to linker about microMIPS code. For
global symbols compiler cannot do the same trick because other tools like,
for example, disassembler wants to know an actual position of the symbol.
So compiler sets STO_MIPS_MICROMIPS flag in the `st_other` field.

In `MIPS<ELFT>::relocateOne()` method we have a symbol's value only and
cannot access any symbol's attributes. To pass type of the symbol
(regular/microMIPS) to that routine as well as other places where we
write a symbol value as-is (.dynamic section, `Elf_Ehdr::e_entry` field
etc) we set when necessary a less-significant bit in the `getSymVA`
function.

Differential revision: https://reviews.llvm.org/D40147

llvm-svn: 354311
2019-02-19 10:36:58 +00:00
Rui Ueyama 9efdd7ac5e [PPC64] Preserve LocalEntry when linking
On PowerPC64, it is necessary to keep the LocalEntry bits in st_other,
especially when -r is used. Otherwise, when the resulting object is used
in a posterior linking, LocalEntry info will be unavailable and
functions may be called through the wrong entrypoint.

Patch by Leandro Lupori.

Differential Revision: https://reviews.llvm.org/D56782

llvm-svn: 354184
2019-02-15 23:11:18 +00:00
Konstantin Zhuravlyov 87498153aa LLD/AMDGPU: Preserve ABI version during linking ELF for AMDGPU
Differential Revision: https://reviews.llvm.org/D58026

llvm-svn: 354086
2019-02-14 23:59:44 +00:00
Nick Desaulniers c3aefedc46 Revert "Revert "[lld] Fix elf::unlinkAsync detached thread""
This reverts commit 9934f2ff02dba9fdabe6e27a83f9f95388bf4132.

llvm-svn: 354081
2019-02-14 23:41:23 +00:00
Nick Desaulniers 289d70cf58 Revert "[lld] Fix elf::unlinkAsync detached thread"
This reverts commit 2694810153.

llvm-svn: 354080
2019-02-14 23:39:32 +00:00
Vitaly Buka 2694810153 [lld] Fix elf::unlinkAsync detached thread
Summary:
So this patch just make sure that the thread is at least stated
before we return from main.

If we just detach then the thread may be actually be stated just after
the process returned from main and it's calling atexit handers. Then the thread may try to create own function static variable and it will
add new at exit handlers confusing libc.

GLIBC before 2.27 had race in that case which corrupted atexit handlers
list. Support for this use-case for other implementation is also unclear,
so we can try just avoid that.

PR40162

Reviewers: ruiu, espindola

Subscribers: emaste, arichardson, jfb, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D58246

llvm-svn: 354078
2019-02-14 23:29:28 +00:00
Rui Ueyama 72c3b1ed1d Move a function from .h to .cpp and use a shorter name. NFC.
llvm-svn: 354054
2019-02-14 19:33:26 +00:00
Rui Ueyama 980fb790c1 Remove a comparator from header and instead use lambdas for simplicity. NFC.
llvm-svn: 354052
2019-02-14 19:21:10 +00:00
Rui Ueyama f69bbbbdd2 Add a comment. NFC.
llvm-svn: 354049
2019-02-14 18:50:59 +00:00
Rui Ueyama b8b81e9b43 Improve error message for unknown relocations.
Previously, we showed the following message for an unknown relocation:

  foo.o: unrecognized reloc 256

This patch improves it so that the error message includes a symbol name:

  foo.o: unknown relocation (256) against symbol bar

llvm-svn: 354040
2019-02-14 18:02:20 +00:00
Peter Collingbourne 8331f61a51 ELF: Allow GOT relocs pointing to non-preemptable ifunc to resolve to an IRELATIVE where possible.
Non-GOT non-PLT relocations to non-preemptible ifuncs result in the
creation of a canonical PLT, which now takes the identity of the IFUNC
in the symbol table. This (a) ensures address consistency inside and
outside the module, and (b) fixes a bug where some of these relocations
end up pointing to the resolver.

Fixes (at least) PR40474 and PR40501.

Differential Revision: https://reviews.llvm.org/D57371

llvm-svn: 353981
2019-02-13 21:49:55 +00:00
Rui Ueyama 4134143cf5 Recover elf32-bigmips and elf32-powerpc support in LLD
This fixes a 7.0 -> 8.0 regression when parsing
OUTPUT_FORMAT("elf32-powerpc"); or elf32-bigmips directive in ldscripts
as well as an unknown emulation error when lld is invoked by clang due
to missed elf32ppclinux case.

Patch by vit9696

Differential Revision: https://reviews.llvm.org/D58005

llvm-svn: 353968
2019-02-13 18:51:15 +00:00
Rui Ueyama 265e8e8252 Show "Unknown -z option" error message even if --version or --help are given.
Previously, we validated -z options after we process --version or --help flags.
So, if one of these flags is given, we wouldn't show an "unknown -z option"
error. This patch fixes that behavior.

Differential Revision: https://reviews.llvm.org/D55446

llvm-svn: 353967
2019-02-13 18:48:39 +00:00
Sean Fertile d694160e66 [PPC64] Sort .toc sections accessed with small code model relocs.
A follow up to the intial patch that unblocked linking against libgcc.
For lld we don't need to bother tracking which objects have got based small
code model relocations. This is due to the fact that the compilers on
powerpc64 use the .toc section to generate indirections to symbols (rather then
using got relocations) which keeps the got small. This makes overflowing a
small code model got relocation very unlikely.

Differential Revision: https://reviews.llvm.org/D57245

llvm-svn: 353849
2019-02-12 15:35:49 +00:00
Rui Ueyama 016833bac2 lld: unquote possibly quoted `EXTERN("symbol")` entry in linker script.
gold accepts quoted strings. binutils requires quoted strings for some
kinds of symbols, e.g.:

  it accepts quoted symbols with @ in name:

  $ echo 'EXTERN("__libc_start_main@@GLIBC_2.2.5")' > a.script
  $ g++ a.script
  /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../../lib64/crt1.o: In function `_start':
  (.text+0x20): undefined reference to `main'
  collect2: error: ld returned 1 exit status

  but rejects them if unquoted:

  $ echo 'EXTERN(__libc_start_main@@GLIBC_2.2.5)' > a.script
  $ g++ a.script
  a.script: file not recognized: File format not recognized
  collect2: error: ld returned 1 exit status

To maintain compatibility with existing linker scripts support quoted
strings in lld as well.

Patch by Lucian Adrian Grijincu.

Differential Revision: https://reviews.llvm.org/D57987

llvm-svn: 353756
2019-02-11 22:01:32 +00:00
Fangrui Song 283d103bde [ELF] Delete a comment that is no longer correct. Fix a typo. NFC
llvm-svn: 353605
2019-02-09 10:09:40 +00:00
Rui Ueyama 04cbd988f9 Fix a bug in R_X86_64_PC{8,16} relocation handling.
R_X86_64_PC{8,16} relocations are sign-extended, so when we check
for relocation overflow, we had to use checkInt instead of checkUInt.
I confirmed that GNU linkers create the same output for the test case.

llvm-svn: 353437
2019-02-07 18:12:57 +00:00
George Rimar 55f7c72bea [LLD][ELF] - Set DF_STATIC_TLS flag for X64 target
This is the same as D57749, but for x64 target.

"ELF Handling For Thread-Local Storage" p41 says (https://www.akkadia.org/drepper/tls.pdf):
R_X86_64_GOTTPOFF relocation is used for IE TLS models.
Hence if linker sees this relocation we should add DF_STATIC_TLS flag.

Differential revision: https://reviews.llvm.org/D57821

llvm-svn: 353378
2019-02-07 07:59:43 +00:00
Ed Maste e9932c103b Correct "varaible" typo in comment
llvm-svn: 353340
2019-02-06 20:36:02 +00:00
Rui Ueyama 1b11e9e8a4 Remove a small header that is used only by one file. NFC.
llvm-svn: 353331
2019-02-06 19:28:23 +00:00
Rui Ueyama 7c77044a38 Add comment.
llvm-svn: 353323
2019-02-06 18:53:17 +00:00
Rui Ueyama 33dbcbb2bc Support R_X86_64_PC8 and R_X86_64_PC16.
They are defined by the x86-64 ELF ABI standard.

Differential Revision: https://reviews.llvm.org/D57799

llvm-svn: 353314
2019-02-06 16:50:09 +00:00
George Rimar ae54e58b90 Recommit r353293 "[LLD][ELF] - Set DF_STATIC_TLS flag for i386 target."
With the following changes:
1) Compilation fix:
std::atomic<bool> HasStaticTlsModel = false; ->
std::atomic<bool> HasStaticTlsModel{false};

2) Adjusted the comment in code.

Initial commit message:

DF_STATIC_TLS flag indicates that the shared object or executable
contains code using a static thread-local storage scheme.

Patch checks if IE/LE relocations were used to check if the code uses
a static model. If so it sets the DF_STATIC_TLS flag.

Differential revision: https://reviews.llvm.org/D57749
----
Modified : /lld/trunk/ELF/Arch/X86.cpp
Modified : /lld/trunk/ELF/Config.h
Modified : /lld/trunk/ELF/SyntheticSections.cpp
Added : /lld/trunk/test/ELF/Inputs/i386-static-tls-model1.s
Added : /lld/trunk/test/ELF/Inputs/i386-static-tls-model2.s
Added : /lld/trunk/test/ELF/Inputs/i386-static-tls-model3.s
Added : /lld/trunk/test/ELF/Inputs/i386-static-tls-model4.s
Added : /lld/trunk/test/ELF/i386-static-tls-model.s
Modified : /lld/trunk/test/ELF/i386-tls-ie-shared.s
Modified : /lld/trunk/test/ELF/tls-dynamic-i686.s
Modified : /lld/trunk/test/ELF/tls-opt-iele-i686-nopic.s

llvm-svn: 353299
2019-02-06 14:43:30 +00:00
George Rimar 52fafcb919 Revert r353293 "[LLD][ELF] - Set DF_STATIC_TLS flag for i386 target."
It broke BB:
http://lab.llvm.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/builds/43450
http://lab.llvm.org:8011/builders/lld-x86_64-freebsd/builds/27891

Error is:
tools/lld/ELF/Config.h:84:41: error: copying member subobject of type
'std::atomic<bool>' invokes deleted constructor std::atomic<bool> HasStaticTlsModel = false;

llvm-svn: 353297
2019-02-06 13:53:32 +00:00
George Rimar da60ad220b [LLD][ELF] - Set DF_STATIC_TLS flag for i386 target.
DF_STATIC_TLS flag indicates that the shared object or executable
contains code using a static thread-local storage scheme.

Patch checks if IE/LE relocations were used to check if the code uses
a static model. If so it sets the DF_STATIC_TLS flag.

Differential revision: https://reviews.llvm.org/D57749

llvm-svn: 353293
2019-02-06 13:38:10 +00:00
Fangrui Song f55e9a2d2e [PPC64] Set the number of relocations processed for R_PPC64_TLS[GL]D to 2
Summary:
R_PPC64_TLSGD and R_PPC64_TLSLD are used as markers on TLS code sequences. After GD-to-IE or GD-to-LE relaxation, the next relocation R_PPC64_REL24 should be skipped to not create a false dependency on __tls_get_addr. When linking statically, the false dependency may cause an "undefined symbol: __tls_get_addr" error.

R_PPC64_GOT_TLSGD16_HA
R_PPC64_GOT_TLSGD16_LO
R_PPC64_TLSGD R_TLSDESC_CALL
R_PPC64_REL24 __tls_get_addr

Reviewers: ruiu, sfertile, syzaara, espindola

Reviewed By: sfertile

Subscribers: emaste, nemanjai, arichardson, kbarton, jsji, llvm-commits, tamur

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D57673

llvm-svn: 353262
2019-02-06 02:00:24 +00:00
Rui Ueyama 3fdb07258b Inline a trivial function and update comment. NFC.
llvm-svn: 353200
2019-02-05 19:19:45 +00:00
Fangrui Song ae0294375f [ELF] Default to --no-allow-shlib-undefined for executables
Summary:
This follows the ld.bfd/gold behavior.

The error check is useful as it captures a common type of ld.so undefined symbol errors as link-time errors:

    // a.cc => a.so (not linked with -z defs)
    void f(); // f is undefined
    void g() { f(); }

    // b.cc => executable with a DT_NEEDED entry on a.so
    void g();
    int main() { g(); }

    // ld.so errors when g() is executed (lazy binding) or when the program is started (-z now)
    // symbol lookup error: ... undefined symbol: f

Reviewers: ruiu, grimar, pcc, espindola

Reviewed By: ruiu

Subscribers: llvm-commits, emaste, arichardson

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D57569

llvm-svn: 352943
2019-02-02 00:34:28 +00:00
Fangrui Song b4744d306c [ELF] Support --{,no-}allow-shlib-undefined
Summary:
In ld.bfd/gold, --no-allow-shlib-undefined is the default when linking
an executable. This patch implements a check to error on undefined
symbols in a shared object, if all of its DT_NEEDED entries are seen.

Our approach resembles the one used in gold, achieves a good balance to
be useful but not too smart (ld.bfd traces all DSOs and emulates the
behavior of a dynamic linker to catch more cases).

The error is issued based on the symbol table, different from undefined
reference errors issued for relocations. It is most effective when there
are DSOs that were not linked with -z defs (e.g. when static sanitizers
runtime is used).

gold has a comment that some system libraries on GNU/Linux may have
spurious undefined references and thus system libraries should be
excluded (https://sourceware.org/bugzilla/show_bug.cgi?id=6811). The
story may have changed now but we make --allow-shlib-undefined the
default for now. Its interaction with -shared can be discussed in the
future.

Reviewers: ruiu, grimar, pcc, espindola

Reviewed By: ruiu

Subscribers: joerg, emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D57385

llvm-svn: 352826
2019-02-01 02:25:05 +00:00
Sam Clegg dfbd19033b Fix names of functions in TargetOptionsCommandFlags.h. NFC.
Differential Revision: https://reviews.llvm.org/D57555

llvm-svn: 352825
2019-02-01 02:24:50 +00:00
Sam Clegg 5cdc91d003 [LTO] Set CGOptLevel in LTO config.
Previously we were never setting this which means it was always being
set to Default (-O2/-Os).

Differential Revision: https://reviews.llvm.org/D57422

llvm-svn: 352667
2019-01-30 20:46:18 +00:00
Dimitry Andric 19b134cc44 Recognize FreeBSD specific BFD names in OUTPUT_FORMAT
Summary:
After rLLD344952 ("Add OUTPUT_FORMAT linker script directive support"),
using BFD names such as `elf64-x86-64-freebsd` the `OUTPUT_FORMAT`
linker script command does not work anymore, resulting in errors like:

```
ld: error: /home/dim/src/clang800-import/stand/efi/loader/arch/amd64/ldscript.amd64:2: unknown output format name: elf64-x86-64-freebsd
>>> OUTPUT_FORMAT("elf64-x86-64-freebsd", "elf64-x86-64-freebsd", "elf64-x86-64-freebsd")
>>>               ^
```

To fix this, recognize a `-freebsd` suffix in BFD names, and also set
`Configuration::OSABI` to `ELFOSABI_FREEBSD` for those cases.

Add and/or update several test cases to check for the correct results of
these new `OUTPUT_FORMAT` arguments.

Reviewers: ruiu, atanasyan, grimar, hokein, emaste, espindola

Reviewed By: ruiu

Subscribers: nemanjai, javed.absar, arichardson, krytarowski, kristof.beyls, kbarton, llvm-commits

Differential Revision: https://reviews.llvm.org/D57283

llvm-svn: 352606
2019-01-30 06:31:52 +00:00
Fangrui Song ae82599a30 [ELF] Simplify. NFC
llvm-svn: 352499
2019-01-29 14:24:35 +00:00
George Rimar ff193c473e [ELF] - Remove dead `readBfdName` declaration. NFC.
`readBfdName` was removed recently.

llvm-svn: 352482
2019-01-29 11:46:00 +00:00
Rui Ueyama 0068d223ee Attempt to fix build failure with GCC 5.4.
llvm-svn: 352435
2019-01-28 21:45:50 +00:00
Peter Collingbourne ff35dbac47 ELF: Set sh_info on RelaIplt to point to the IgotPlt output section.
Previously we were setting it to the GotPlt output section, which is
incorrect on ARM where this section is in .got. In static binaries
this can lead to sh_info being set to -1 (because there is no .got.plt)
which results in various tools rejecting the output file.

Differential Revision: https://reviews.llvm.org/D57274

llvm-svn: 352413
2019-01-28 19:29:41 +00:00
Rui Ueyama 4f8c82281d Refactoring. NFC.
llvm-svn: 352407
2019-01-28 19:11:52 +00:00
Nick Desaulniers 6cff0cb35a lld: elf: discard more specific .gnu.linkonce section
Summary:
lld discards .gnu.linonce.* sections work around a bug in glibc.
https://sourceware.org/bugzilla/show_bug.cgi?id=20543

Unfortunately, the Linux kernel uses a section named
.gnu.linkonce.this_module to store infomation about kernel modules. The
kernel reads data from this section when loading kernel modules, and
errors if it fails to find this section. The current behavior of lld
discards this section when kernel modules are linked, so kernel modules
linked with lld are unloadable by the linux kernel.

The Linux kernel should use a comdat section instead of .gnu.linkonce.
The minimum version of binutils supported by the kernel supports comdat
sections. The kernel is also not relying on the old linkonce behavior;
it seems to have chosen a name that contains a deprecated GNU feature.

Changing the section name now in the kernel would require all kernel
modules to be recompiled to make use of the new section name. Instead,
rather than discarding .gnu.linkonce.*, let's discard the more specific
section name to continue working around the glibc issue while supporting
linking Linux kernel modules.

Link: https://github.com/ClangBuiltLinux/linux/issues/329

Reviewers: pcc, espindola

Reviewed By: pcc

Subscribers: nathanchance, emaste, arichardson, void, srhines

Differential Revision: https://reviews.llvm.org/D57294

llvm-svn: 352302
2019-01-27 02:54:23 +00:00
Rui Ueyama 3749befe74 Remove dead declaration.
llvm-svn: 352257
2019-01-26 00:31:49 +00:00
Rui Ueyama e14e46b3f1 Simplify. NFC.
llvm-svn: 352242
2019-01-25 21:25:25 +00:00
Sean Fertile 83cb252876 [PPC64] Reland r351978 'Sort .toc sections accessed with small code model ...'
Guessing that the slashes used in the scripts SECTION command was causing the
windows related failures in the added test.

Original commit message:
Small code model global variable access on PPC64 has a very limited range of
addressing. The instructions the relocations are used on add an offset in the
range [-0x8000, 0x7FFC] to the toc pointer which points to .got +0x8000, giving
an addressable range of [.got, .got + 0xFFFC]. While user code can be recompiled
with medium and large code models when the binary grows too large for small code
model, there are small code model relocations in the crt files and libgcc.a
which are typically shipped with the distros, and the ABI dictates that linkers
must allow linking of relocatable object files using different code models.

To minimze the chance of relocation overflow, any file that contains a small
code model relocation should have its .toc section placed closer to the .got
then any .toc from a file without small code model relocations.

Differential Revision: https://reviews.llvm.org/D56920

llvm-svn: 352071
2019-01-24 18:17:40 +00:00
Rui Ueyama ec33fd6dd5 Untabify.
llvm-svn: 352070
2019-01-24 18:17:17 +00:00
Serge Guelton 1fa239f500 Partial support of SHT_GROUP without flag
This does *not* implement full SHT_GROUP semantic, yet it is a simple step forward:
Sections within a group are still considered valid, but they do not behave as
specified by the standard in case of garbage collection.

Differential Revision: https://reviews.llvm.org/D56437

llvm-svn: 352068
2019-01-24 17:56:08 +00:00
Sean Fertile 72679cff12 Revert "[PPC64] Sort .toc sections accessed with small code model ..."
This reverts commit ca87c57a3aa4770c9cf0defd4b2feccbc342ee93.
Added test fails on several windows buildbots.

llvm-svn: 351985
2019-01-23 21:46:28 +00:00
Sean Fertile a010cf615a [PPC64] Sort .toc sections accessed with small code model relocs close to .got.
Small code model global variable access on PPC64 has a very limited range of
addressing. The instructions the relocations are used on add an offset in the
range [-0x8000, 0x7FFC] to the toc pointer which points to .got +0x8000, giving
an addressable range of [.got, .got + 0xFFFC]. While user code can be recompiled
with medium and large code models when the binary grows too large for small code
model, there are small code model relocations in the crt files and libgcc.a
which are typically shipped with the distros, and the ABI dictates that linkers
must allow linking of relocatable object files using different code models.

To minimze the chance of relocation overflow, any file that contains a small
code model relocation should have its .toc section placed closer to the .got
then any .toc from a file without small code model relocations.

Differential Revision: https://reviews.llvm.org/D56920

llvm-svn: 351978
2019-01-23 21:04:03 +00:00
Peter Collingbourne bcd08c16bb COFF, ELF: ICF: Perform 2 rounds of relocation hash propagation.
LLD's performance on PGO instrumented Windows binaries was still not
great even with the fix in D56955; out of the 2m41s linker runtime,
around 2 minutes were still being spent in ICF. I looked into this more
closely and discovered that the vast majority of the runtime was being
spent segregating .pdata sections with the following relocation chain:

.pdata -> identical .text -> unique PGO counter (not eligible for ICF)

This patch causes us to perform 2 rounds of relocation hash
propagation, which allows the hash for the .pdata sections to
incorporate the identifier from the PGO counter. With that, the amount
of time spent in ICF was reduced to about 2 seconds. I also found that
the same change led to a significant ICF performance improvement in a
regular release build of Chromium's chrome_child.dll, where ICF time
was reduced from around 1s to around 700ms.

With the same change applied to the ELF linker, median of 100 runs
for lld-speed-test/chrome reduced from 4.53s to 4.45s on my machine.

I also experimented with increasing the number of propagation rounds
further, but I did not observe any further significant performance
improvements linking Chromium or Firefox.

Differential Revision: https://reviews.llvm.org/D56986

llvm-svn: 351899
2019-01-22 23:54:49 +00:00
Peter Collingbourne 3426111145 COFF, ELF: Adjust ICF hash computation to account for self relocations.
It turns out that sections in PGO instrumented object files on Windows
contain a large number of relocations pointing to themselves. With
r347429 this can cause many sections to receive the same hash (usually
zero) as a result of a section's hash being xor'ed with itself.

This patch causes the COFF and ELF linkers to avoid this problem
by adding the hash of the relocated section instead of xor'ing it.
On my machine this causes the regressing test case
provided by Mozilla to terminate in 2m41s.

Differential Revision: https://reviews.llvm.org/D56955

llvm-svn: 351898
2019-01-22 23:51:35 +00:00
Serge Guelton 52d777fc4f Tentative fix for r351701 and gcc 6.2 build on ubuntu
llvm-svn: 351728
2019-01-21 07:44:52 +00:00
Chandler Carruth 2946cd7010 Update the file headers across all of the LLVM projects in the monorepo
to reflect the new license.

We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.

Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.

llvm-svn: 351636
2019-01-19 08:50:56 +00:00
Mandeep Singh Grang f9d76dc354 [lld] Use range-based llvm::sort
llvm-svn: 351612
2019-01-18 23:41:34 +00:00
Peter Wu 5391489674 [ELF][X86_64] Fix corrupted LD -> LE optimization for TLS without PLT
The LD -> LE optimization for Thread-Local Storage without PLT requires
an additional "66" prefix, otherwise the next instruction will be
corrupted, causing runtime misbehavior (crashes) of the linked object.

The other (GD -> IE/LD) optimizations are the same with or without PLT,
but add tests for completeness. The instructions are copied from
https://raw.githubusercontent.com/wiki/hjl-tools/x86-psABI/x86-64-psABI-1.0.pdf#subsection.11.1.2

This does not try to address ILP32 (x32) support.

Fixes https://bugs.llvm.org/show_bug.cgi?id=37303

Reviewed By: ruiu

Differential Revision: https://reviews.llvm.org/D56779

llvm-svn: 351396
2019-01-16 23:28:51 +00:00
Peter Smith e8374614c9 [ELF][AArch64] Add R_AARCH64_PLT_PAGE_PC to isRelExpr
As a follow on to D56666 (r351186) there is a case when taking the address
of an ifunc when linking -pie that can generate a spurious can't create
dynamic relocation R_AARCH64_ADR_PREL_PG_HI21 against symbol in readonly
segment. Specifically the case is where the ifunc is in the same
translation unit as the address taker, so given -fpie the compiler knows
the ifunc is defined in the executable so it can use a non-got-generating
relocation.

The error message is due to R_AARCH64_PLT_PAGE_PC not being added to
isRelExpr, its non PLT equivalent R_AARCH64_PAGE_PC is already in
isRelExpr.

Differential Revision: https://reviews.llvm.org/D56724

llvm-svn: 351335
2019-01-16 13:24:02 +00:00
Peter Smith 13d134684f [ELF] Implement option to force PIC compatible Thunks
By default LLD will generate position independent Thunks when the --pie or
--shared option is used. Reference to absolute addresses is permitted in
other cases. For some embedded systems position independent thunks are
needed for code that executes before the MMU has been set up. The option
--pic-veneer is used by ld.bfd to force position independent thunks.
    
The patch adds --pic-veneer as the option is needed for the Linux kernel
on Arm.
    
fixes pr39886
    
Differential Revision: https://reviews.llvm.org/D55505

llvm-svn: 351326
2019-01-16 12:09:13 +00:00
Rui Ueyama 6f9d49cdde Do not emit a corrupt symbol table entry for .rela_iplt_{start,end}.
If .rela.iplt does not exist, we used to emit a corrupt symbol table
that contains two symbols, .rela_iplt_{start,end}, pointing to a
nonexisting section.

This patch fixes the issue by setting section index 0 to the symbols
if .rel.iplt section does not exist.

Differential Revision: https://reviews.llvm.org/D56623

llvm-svn: 351218
2019-01-15 18:30:23 +00:00
Peter Smith b20f993df8 [ELF][AArch64] Add missing PLT relocations to isStaticLinkTimeConstant
r347650 fixed pr38074 for AArch64 for static linking. It added two new
RelExpr instances R_AARCH64_GOT_PAGE_PC_PLT and R_GOT_PLT. These need to be
added to isStaticLinkTimeConstant so that the address of an ifunc can be
taken when building a shared library.

fixes pr40250

Differential Revision: https://reviews.llvm.org/D56666

llvm-svn: 351186
2019-01-15 11:17:03 +00:00
Rui Ueyama b41a6966e0 TrapInstr must be 4 bytes long. Currently we write two zeros on every two bytes.
llvm-svn: 350853
2019-01-10 17:45:56 +00:00
Peter Smith 1811e48b7b [ELF] Fix ARM and Thumb V7PILongThunk overflow behavior.
When the range between the source and target of a V7PILongThunk exceeded an
int32 we would trigger a relocation out of range error for the
R_ARM_MOVT_PREL or R_ARM_THM_MOVT_PREL relocation. This case can happen when
linking the linux kernel as it is loaded above 0xf0000000.

There are two parts to the fix.
- Remove the overflow check for R_ARM_MOVT_PREL or R_ARM_THM_MOVT_PREL. The
ELF for the ARM Architecture document defines these relocations as having no
overflow checking so the check was spurious.
- Use int64_t for the offset calculation, in line with similar thunks so
that PC + (S - P) < 32-bits. This results in less surprising disassembly.

Differential Revision: https://reviews.llvm.org/D56396

llvm-svn: 350836
2019-01-10 16:08:23 +00:00
Sean Fertile 3ca494b2ee Modify InputSectionBase::getLocation to add section and offset to every loc.
The section and offset can be very helpful in diagnosing certian errors.
For example on a relocation overflow or misalignment diagnostic:

test.c:(function  foo): relocation R_PPC64_ADDR16_DS out of range: ...

The function foo can have many R_PPC64_ADDR16_DS relocations. Adding the offset
and section will identify exactly which relocation is causing the failure.

Differential Revision: https://reviews.llvm.org/D56453

llvm-svn: 350828
2019-01-10 15:08:06 +00:00
Sean Fertile 461725b585 [PPC64] Fix RelType in checkInt and checkAlignment diagnsotics.
In the PPC64 target we map toc-relative relocations, dynamic thread pointer
relative relocations, and got relocations into a corresponding ADDR16 relocation
type for handling in relocateOne. This patch saves the orignal RelType before
mapping to an ADDR16 relocation so that any diagnostic messages will not
mistakenly use the mapped type.

Differential Revision: https://reviews.llvm.org/D56448

llvm-svn: 350827
2019-01-10 15:08:02 +00:00
George Rimar 73af3d4060 [LLD][ELF] - Support MSP430.
Patch by Michael Skvortsov!

This change adds a basic support for linking static MSP430 ELF code.
Implemented relocation types are intended to correspond to the BFD.

Differential revision: https://reviews.llvm.org/D56535

llvm-svn: 350819
2019-01-10 13:43:06 +00:00
Rui Ueyama c58fbe3c71 Fix formatting. NFC.
llvm-svn: 350767
2019-01-09 22:24:27 +00:00
Ryan Prichard d7d2369c09 [ARM][AArch64] Increase TLS alignment to reserve space for Android's TCB
ARM and AArch64 use TLS variant 1, where the first two words after the
thread pointer are reserved for the TCB, followed by the executable's TLS
segment. Both the thread pointer and the TLS segment are aligned to at
least the TLS segment's alignment.

Android/Bionic historically has not supported ELF TLS, and it has
allocated memory after the thread pointer for several Bionic TLS slots
(currently 9 but soon only 8). At least one of these allocations
(TLS_SLOT_STACK_GUARD == 5) is widespread throughout Android/AArch64
binaries and can't be changed.

To reconcile this disagreement about TLS memory layout, set the minimum
alignment for executable TLS segments to 8 words on ARM/AArch64, which
reserves at least 8 words of memory after the TP (2 for the ABI-specified
TCB and 6 for alignment padding). For simplicity, and because lld doesn't
know when it's targeting Android, increase the alignment regardless of
operating system.

Differential Revision: https://reviews.llvm.org/D53906

llvm-svn: 350681
2019-01-09 00:09:59 +00:00
Thomas Anderson 4a401e9479 Fix linker-defined symbols possibly not being defined when -wrap is used
Fixes https://bugs.llvm.org/show_bug.cgi?id=40134

addWrappedSymbols() must be called before addReservedSymbols() because the
latter only defines reserved symbols when they are undefined in the symbol
table. If addWrappedSymbols() is called after, then addUndefined() is called
which may lazily pull in more object files that could reference reserved
symbols.

Differential Revision: https://reviews.llvm.org/D56110

llvm-svn: 350251
2019-01-02 19:28:00 +00:00
Fangrui Song 50394f6e01 [ELF] A shared object is needed if any of its occurrences is needed
Summary:
If a DSO appears more than once with and without --as-needed, ld.bfd and gold consider --no-as-needed to takes precedence over --as-needed. lld didn't and this patch makes it do so.

This makes it a bit away from the position-dependent behavior (how
different occurrences of the same DSO interact) and protects us from
some mysterious runtime errors: if some interceptor libraries add their
own --no-as-needed dependencies (e.g. librt.so), and the user
application specifies -Wl,--as-needed -lrt , the absence of the
DT_NEEDED entry would make dlsym(RTLD_NEXT, "clock_gettime") return NULL
and would break at runtime.

Reviewers: ruiu, espindola

Reviewed By: ruiu

Subscribers: emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D56089

llvm-svn: 350105
2018-12-27 22:24:45 +00:00
Rui Ueyama 56c5343e5a Use error() instead of fatal() to report an invalid address range.
In this patch we also use toString() to stringize a section.

llvm-svn: 350070
2018-12-26 19:15:04 +00:00
David Blaikie 059b1c5e01 gdb-index: Handle errors when parsing ranges
When parsing CU ranges for gdb-index, handle the error (now propagated
up though the API lld is calling here - previously the error was
printed within the libDebugInfo API, not allowing lld to format or
handle the message at all) - including information about the object and
archive name, as well as failing the link.

llvm-svn: 349979
2018-12-22 00:31:05 +00:00
Fangrui Song f747971bf1 keymethod -> keyfunction
Pointed out by ruiu in rLLD349969

llvm-svn: 349974
2018-12-21 22:57:11 +00:00
Fangrui Song fe36417f6e [ELF] .gnu.hash bloom filter: use Shift2 = 26 instead of 6
Summary:
For the 2-bit bloom filter, we currently pick the bits Hash%64 and Hash>>6%64 (Shift2=6), but bits [6:...] are also used to select a word, causing a loss of precision.

In this patch, we choose Shift2=26, with is suggested by Ambrose Feinstein.

Note, Shift2 is computed as maskbitslog2 in bfd/elflink.c and gold/dynobj.cc
It is varying with the number of dynamic symbols but we don't
necessarily copy its rule.

Reviewers: ruiu, espindola

Reviewed By: ruiu

Subscribers: emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D55971

llvm-svn: 349966
2018-12-21 21:59:34 +00:00
Rui Ueyama afe9673df2 Add a doc for missing key function and an error message referencing the doc.
Summary:
This is a common error, and because many people don't know what the key
function is, it is sometimes very confusing.

The doc was originally written by Brooks Moses and slightly edited by me.

Reviewers: MaskRay, espindola

Subscribers: emaste, llvm-commits, arichardson

Differential Revision: https://reviews.llvm.org/D55968

llvm-svn: 349941
2018-12-21 19:28:49 +00:00
Rui Ueyama 23b28703f9 Remove dead code.
This code is no-op because of r349849.

Differential Revision: https://reviews.llvm.org/D55962

llvm-svn: 349859
2018-12-20 23:47:39 +00:00
Rui Ueyama fb81437638 Simplify. NFC.
llvm-svn: 349850
2018-12-20 22:54:41 +00:00
Fangrui Song 51fe635f81 [ELF] Move IsNeeded logic from SymbolTable::addShared to MarkLive, and check IsUsedInRegularObj
Summary:
In glibc, libc.so is a linker script with an as-needed dependency on ld-linux-x86-64.so.2

    GROUP ( /lib/x86_64-linux-gnu/libc.so.6 /usr/lib/x86_64-linux-gnu/libc_nonshared.a  AS_NEEDED ( /lib/x86_64-linux-gnu/ld-linux-x86-64.so.2 ) )

ld-linux-x86-64.so.2 (as-needed) defines some symbols which resolve undefined references in libc.so.6, it will therefore be added as a DT_NEEDED entry, which isn't necessary.

The test case as-needed-not-in-regular.s emulates the libc.so scenario, where ld.bfd and gold don't add DT_NEEDED for a.so

The relevant code in gold/resolve.cc:

  // If we have a non-WEAK reference from a regular object to a
  // dynamic object, mark the dynamic object as needed.
  if (to->is_from_dynobj() && to->in_reg() && !to->is_undef_binding_weak())
    to->object()->set_is_needed();

in_reg() appears to do something similar to IsUsedInRegularObj.

This patch makes lld do the similar thing, but moves the check from
addShared to a later stage MarkLive where all symbols are scanned.

Reviewers: ruiu, pcc, espindola

Reviewed By: ruiu

Subscribers: emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D55902

llvm-svn: 349849
2018-12-20 22:46:01 +00:00
Sean Fertile 367a7b7d66 [PPC64] Add toc-optimizations for got based relocations.
Differential Revision: https://reviews.llvm.org/D54907

llvm-svn: 349772
2018-12-20 17:00:33 +00:00
Rui Ueyama 392f0b2b71 Simplify. NFC.
Differential Revision: https://reviews.llvm.org/D55903

llvm-svn: 349697
2018-12-19 23:25:02 +00:00
George Rimar 751e6e1cf5 [LLD][ELF] - Report a location for symbols from the linker script when reporting an error.
When we report an error for symbols defined in the linker script,
we do not report the location properly.

For example:

ld.lld: error: relocation R_AARCH64_CALL26 cannot refer to absolute symbol: aliasto__text
>>> defined in <internal>
>>> referenced by rtoabs.o:(.text+0x4)

This patch fixes that.

Differential revision: https://reviews.llvm.org/D55360

llvm-svn: 349612
2018-12-19 10:19:40 +00:00
Rui Ueyama 2d7d128117 Use unique_ptr to manage a TarWriter instance. NFC.
llvm-svn: 349581
2018-12-18 23:50:37 +00:00
Rui Ueyama 5961809640 Fix error message.
Previously, this code printed out an error message like this

  ld.lld: error: --reproduce: failed to open /foo: cannot open /foo

Apparently "failed to open /foo:" part is redundant.

llvm-svn: 349571
2018-12-18 23:33:10 +00:00
Rui Ueyama a01246b7a7 Reject .so files if -static is given.
Previously, if you pass -static to lld, lld searches for only foo.a
and skips foo.so for -lfoo option. However, it didn't reject .so files
if you directly pass their pathnames via the command line, which is a bug.

Differential Revision: https://reviews.llvm.org/D55845

llvm-svn: 349557
2018-12-18 22:30:23 +00:00
Fangrui Song 41031d97f2 [ELF] Place .note in the first page to ensure they are available in core files
Summary:
Other large sections (e.g. .rela.dyn .dynstr) may push .note.* off the
first page. They won't be available in core files if RLIMIT_CORE is
limited.

This patch gives priority to alloctable SHT_NOTE sections so that they
are assuredly in the first page and will be available in core files.
They are small and contain important information (e.g. .note.gnu.build-id
identifies the origin of the core, .note.tag stores NT_FREEBSD_ABI_TAG).

Note: gold Output_section_order has a similar rule:

  // Loadable read-only note sections come next so that the PT_NOTE
  // segment is on the first page of the executable.
  ORDER_RO_NOTE,

Reviewers: ruiu, pcc, espindola

Subscribers: emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D55800

llvm-svn: 349524
2018-12-18 19:16:37 +00:00
Sean Fertile 09a5bc0107 [PPC64] Support got-based relocations.
Differential Revison: https://reviews.llvm.org/D54859

llvm-svn: 349511
2018-12-18 17:34:26 +00:00
Peter Collingbourne 4ca86d289f ELF: AArch64: Fix errata patch address calculation.
The code here wants the output section offset of the instruction
requiring the errata patch, not the virtual address. Without this
change we can end up placing a patch out of range if the virtual
address of the code section is large enough.

Differential Revision: https://reviews.llvm.org/D55732

llvm-svn: 349386
2018-12-17 20:15:12 +00:00
Peter Smith 6ece0ad429 [ELF][ARM] Add support for architecture v6m thunks
ARM Architecture v6m is used by the smallest microcontrollers such as the
cortex-m0. It is Thumb only (no Thumb 2) which prevents it from using the
existing Thumb 2 range extension thunks as these use the Thumb 2 movt/movw
instructions. Range extension thunks are not usually needed for
microcontrollers due to the small amount of flash and ram on the device,
however if code is copied from flash into ram then a range extension thunk
is required to call that code.

This change adds support for v6m range extension thunks. The procedure call
standard APCS permits a thunk to corrupt the intra-procedural scratch
register r12 (referred to as ip in the APCS). Most Thumb instructions do
not permit access to high registers (r8 - r15) so the thunks must spill
some low registers (r0 - r7) to perform the control transfer.

Fixes pr39922

Differential Revision: https://reviews.llvm.org/D55555

llvm-svn: 349337
2018-12-17 10:33:47 +00:00
Peter Collingbourne ae15e7232a ELF: Handle R_ARM_V4BX correctly in PIC output files.
Previously we considered R_ARM_V4BX to be an absolute relocation,
which meant that we rejected it in read-only sections in PIC output
files. Instead, treat it as a hint relocation so that relocation
processing ignores it entirely.

Also fix a problem with the test case where it was never being run
because it has a .yaml extension and we don't run tests with that
extension.

Differential Revision: https://reviews.llvm.org/D55728

llvm-svn: 349216
2018-12-15 00:20:17 +00:00
Rui Ueyama 9f49990976 Add --plugin-opt=emit-llvm option.
`--plugin-opt=emit-llvm` is an option for LTO. It makes the linker to
combine all bitcode files and write the result to an output file without
doing codegen. Gold LTO plugin has this option.

This option is being used for some post-link code analysis tools that
have to see a whole program but don't need to see them in the native
machine code.

Differential Revision: https://reviews.llvm.org/D55717

llvm-svn: 349198
2018-12-14 21:58:49 +00:00
Peter Smith fe3015d164 [ELF][AArch64] Fix adrp to undefined weak reference.
In the ABI for the 64-bit Arm architecture the section on weak references
states:
During linking, the symbol value of an undefined weak reference is:
- Zero if the relocation type is absolute
- The address of the place if the relocation type is pc-relative.

The relocations associated with an ADRP are relative so we should resolve
the undefined weak reference to the place instead of 0. This matches GNU
ld.bfd behaviour.

fixes pr34928

Differential Revision: https://reviews.llvm.org/D55599

llvm-svn: 349024
2018-12-13 11:13:01 +00:00
George Rimar 1f958ed269 [LLD][ELF] - Support discarding the .dynamic section.
This is a part of https://bugs.llvm.org/show_bug.cgi?id=39810.

Seems it turns out that supporting /DISCARD/ for the .dynamic section with the
linker script is something we can do easily. The patch does this.

Differential revision: https://reviews.llvm.org/D55211

llvm-svn: 348749
2018-12-10 09:24:49 +00:00
George Rimar ad667661c4 [ELF] - Allow discarding .dynsym from the linker script.
This is a part of https://bugs.llvm.org/show_bug.cgi?id=39810.
The patch allows discarding the .dynsym section using linker script.

Differential revision: https://reviews.llvm.org/D55218

llvm-svn: 348748
2018-12-10 09:13:36 +00:00
George Rimar 4af28e46ca [LLD][ELF] - Support discarding .dynstr section.
This is a part of https://bugs.llvm.org/show_bug.cgi?id=39810.
The patch allows discarding the .dynstr section using linker script.

Differential revision: https://reviews.llvm.org/D55215

llvm-svn: 348746
2018-12-10 09:07:30 +00:00
George Rimar e1fd3f940b [LLD][ELF] - Use auto. NFC.
This addresses the missed review comment.

llvm-svn: 348480
2018-12-06 10:56:11 +00:00
George Rimar b3be390f94 [ELF] - (-Map file) Implement printing of LMA for assignments outside of section declarations.
This was a missing piece.
We started to print LMAs and information about assignments,
but did not do that for assignments outside of section declarations yet.
The patch implements it.

Differential revision: https://reviews.llvm.org/D45314

llvm-svn: 348468
2018-12-06 09:04:52 +00:00
George Rimar f49fe218c2 [LLD][ELF] - Linker script: accept using a file name without a list of sections.
This is a part of
https://bugs.llvm.org/show_bug.cgi?id=39885

Linker script specification says:
"You can specify a file name to include sections from a particular file. You would
do this if one or more of your files contain special data that needs to be at a
particular location in memory."

LLD did not accept this syntax. The patch implements it.

Differential revision: https://reviews.llvm.org/D55324

llvm-svn: 348463
2018-12-06 08:34:52 +00:00
Rui Ueyama c9c34bdc1a Do not use a hash table to uniquify mergeable strings.
Previously, we have a hash table containing strings and their offsets
to manage mergeable strings. Technically we can live without that, because
we can do binary search on a vector of mergeable strings to find a mergeable
strings.

We did have both the hash table and the binary search because we thought
that that is faster.

We recently observed that lld tend to consume more memory than gold when
building an output with debug info. A few percent of memory is consumed by
the hash table. So, we needed to reevaluate whether or not having the extra
hash table is a good CPU/memory tradeoff. I run a few benchmarks with and
without the hash table.

I got a mixed result for the benchmark. We observed a regression for some
programs by removing the hash table (that's what we expected), but we also
observed that performance imrpovements for some programs. This is perhaps
due to reduced memory usage.

Differential Revision: https://reviews.llvm.org/D55234

llvm-svn: 348401
2018-12-05 19:13:31 +00:00
Fangrui Song 01fbb06b12 [ELF] Simplify getSectionPiece
Reviewers: ruiu, espindola

Reviewed By: ruiu

Subscribers: grimar, emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D55248

llvm-svn: 348311
2018-12-04 22:25:05 +00:00
Rui Ueyama e84f14ee39 Remove unreachable code.
llvm-svn: 348294
2018-12-04 19:00:56 +00:00
Rui Ueyama a592aeb35a ELF: allow non allocated sections to go into allocated sections
Patch from Andrew Kelley.

For context, see https://bugs.llvm.org/show_bug.cgi?id=39862

The use case is embedded / OS programming where the kernel wants
access to its own debug info via mapped dwarf info. I have a proof of
concept of this working, using this linker script snippet:

  .rodata : ALIGN(4K) {
    *(.rodata)
    __debug_info_start = .;
    KEEP(*(.debug_info))
    __debug_info_end = .;
    __debug_abbrev_start = .;
    KEEP(*(.debug_abbrev))
    __debug_abbrev_end = .;
    __debug_str_start = .;
    KEEP(*(.debug_str))
    __debug_str_end = .;
    __debug_line_start = .;
    KEEP(*(.debug_line))
    __debug_line_end =
    .;
    __debug_ranges_start
    = .;
    KEEP(*(.debug_ranges))
    __debug_ranges_end
    = .;
  }

Differential revision: https://reviews.llvm.org/D55276

llvm-svn: 348291
2018-12-04 18:47:44 +00:00
Martell Malone 15b6c453b8 [ELF] Allow discarding of .rela.plt
When linking the linux kernel on ppc64le

ld.lld -EL -m elf64lppc -Bstatic --orphan-handling=warn --build-id -o
.tmp_vmlinux1 -T ./arch/powerpc/kernel/vmlinux.lds --whole-archive
built-in.a --no-whole-archive --start-group lib/lib.a --end-group
ld.lld: error: discarding .rela.plt section is not allowed

The linker script discards with the following matches
*(.glink .iplt .plt .rela* .comment)

Differential Revision: https://reviews.llvm.org/D54871

llvm-svn: 348258
2018-12-04 12:37:56 +00:00
Martell Malone aa6086a64c [PPC][PPC64] PPC_REL14 and PPC64_REL14 relocations
When linking the linux kernel on ppc64 and ppc
ld.lld: error: unrecognized reloc 11
11 is PPC_REL14 and PPC64_REL14

Differential revision: https://reviews.llvm.org/D54868

llvm-svn: 348255
2018-12-04 12:26:21 +00:00
Rui Ueyama 3b6cd2c602 Show a proper error message if output file is too large.
At least on Linux, if a file size given to FileOutputBuffer is greater
than 2^63, it fails with "Invalid argument" error, which is not a
user-friendly error message. With this patch, lld prints out "output
file too large" instead.

llvm-svn: 348153
2018-12-03 17:42:57 +00:00
George Rimar 89230f7bc2 [LLD][ELF] - Remove the excessive condition. NFC.
There is no need to check that In.DynSymTab != nullptr,
because `includeInDynsym` already checks for `!Config->HasDynSymTab`
and `HasDynSymTab` is the pre-condition for In.DynSymTab creation.

llvm-svn: 348143
2018-12-03 14:44:02 +00:00
George Rimar c2bea66cf2 [LLD][ELF] - Improve the DWARF v5 suport for building .gdb_index.
Now LLD might build the broken/incomplete .gdb_index when some DWARF v5
sections (like .debug_rnglists and .debug_addr) are used.

Particularly, for the case above, we emit an empty address area. 
A test case is provided and patch fixes the issue.

Differential revision: https://reviews.llvm.org/D55109

llvm-svn: 348119
2018-12-03 10:48:57 +00:00
Rui Ueyama aea706083f Inline a function template that is used only once. NFC.
llvm-svn: 348013
2018-11-30 18:19:15 +00:00
Peter Smith fd8aeb2c2a [LLD][ELF] Error if _GLOBAL_OFFSET_TABLE_ is defined in input objects
The _GLOBAL_OFFSET_TABLE_ is a linker defined symbol that is placed at
some location relative to the .got, .got.plt or .toc section. On some
targets such as Arm the correctness of some code sequences using a
relocation to _GLOBAL_OFFSET_TABLE_ depend on the value of the symbol
being in the linker defined place. Follow the ld.gold example and give
a multiple symbol definition error. The ld.bfd behaviour is to ignore the
definition in the input object and redefine it, which seems like it could
be more surprising.

fixes pr39587

Differential Revision: https://reviews.llvm.org/D54624

llvm-svn: 347854
2018-11-29 11:18:07 +00:00
Fangrui Song f2143761d6 [ELF] --gdb-index: use lower_bound to compute relative CU index in the object file
Summary:
This reinstates what I originally intended to do in D54361.
It removes the assumption that .debug_gnu_pubnames has increasing CuOffset.

Now we do better than gold here: when .debug_gnu_pubnames contains
multiple sets, gold would think every set has the same CU index as the
first set (incorrect).

Reviewed By: ruiu

Reviewers: ruiu, dblaikie, espindola

Subscribers: emaste, arichardson, arphaman, llvm-commits

Differential Revision: https://reviews.llvm.org/D54483

llvm-svn: 347820
2018-11-29 00:17:00 +00:00
Rui Ueyama 63d397ea6e Simplify Symbol::getPltVA.
This patch also makes getPltEntryOffset a non-member function because
it doesn't depend on any private members of the TargetInfo class.

I tried a few different ideas, and it seems this change fits in best to me.

Differential Revision: https://reviews.llvm.org/D54981

llvm-svn: 347781
2018-11-28 17:42:59 +00:00
Haojian Wu ac6e1fb89d Fix a typo.
llvm-svn: 347751
2018-11-28 13:20:39 +00:00
Haojian Wu 00e82a6f66 Fix build error.
llvm-svn: 347750
2018-11-28 13:15:06 +00:00
Simon Atanasyan 69c61200a9 [ELF][MIPS] Handle mips in the OUTPUT_FORMAT directive
Differential Revision: http://reviews.llvm.org/D54920

llvm-svn: 347742
2018-11-28 11:41:20 +00:00
Peter Smith 7dc5af75ae [ELF] Use more specific method to calculate DT_PLTRELSZ
The DT_PLTRELSZ dynamic tag is calculated using the size of the
OutputSection containing the In.RelaPlt InputSection. This will work for the
default no linker script case and the majority of linker scripts.
Unfortunately it doesn't work for some 'almost' sensible linker scripts. It
is permitted by ELF to have a single OutputSection containing both
In.RelaDyn, In.RelaPlt and In.RelaIPlt. It is also permissible for the range
of memory [DT_RELA, DT_RELA + DT_RELASZ) and the range
[DT_JMPREL, DT_JMPREL + DT_JMPRELSZ) to overlap as long as the the latter
range is at the end.

To support this type of linker script use the specific InputSection sizes.

Fixes pr39678

Differential Revision: https://reviews.llvm.org/D54759

llvm-svn: 347736
2018-11-28 10:04:55 +00:00
Rui Ueyama 418bcd0830 Make a member function non-member. NFC.
llvm-svn: 347678
2018-11-27 17:47:24 +00:00
George Rimar d2f8db827d [ELF] - Fix R_AARCH64_ADR_GOT_PAGE, R_AARCH64_LD64_GOT_LO12 handling against IFUNC symbols.
This is https://bugs.llvm.org/show_bug.cgi?id=38074.

The issue is that when calling a function, LLD generates a
.got entry that points to the IFUNC resolver function when
instead, it should use the PLT entries properly for
handling the IFUNC.

So we should create a got entry that points to PLT entry,
which itself loads the value from 
.got.plt, relocated with R_*_IRELATIVE to make things work.
Patch do that.

Differential revision: https://reviews.llvm.org/D54314

llvm-svn: 347650
2018-11-27 10:30:46 +00:00
Peter Smith 8a3e7178f6 [AArch64] Cortex-a53-843419 erratum should not apply to relaxed TLS.
The changes to the instructions performed by TLS relaxation and the errata
patching are performed with relocations. As these are applied so late the
errata scanning won't see the changes in the section data made by the TLS
relaxation. This can lead to a TLS relaxed sequence being patched when it
doesn't need to be.

The fix checks to see if there is a R_RELAX_TLS_IE_TO_LE instruction at the
same address as the ADRP as this indicates the presence of a relaxation
of a sequence that might get recognised as a patch.

Differential Revision: https://reviews.llvm.org/D54854

llvm-svn: 347649
2018-11-27 10:17:35 +00:00
George Rimar a1b3ddbfec [ELF] - Implement -z nodefaultlib
This is https://bugs.llvm.org//show_bug.cgi?id=38978

Spec says that:
"Objects may be built with the -z nodefaultlib option to
suppress any search of the default locations at runtime.
Use of this option implies that all the dependencies of an
object can be located using its runpaths.
Without this option, which is the most common case, no
matter how you augment the runtime linker's library
search path, its last element is always /usr/lib for 32-bit
objects and /usr/lib/64 for 64-bit objects."

The patch implements this option.

Differential revision: https://reviews.llvm.org/D54577

llvm-svn: 347647
2018-11-27 09:48:17 +00:00
Peter Collingbourne c81be65166 ELF: ICF: Include contents of referenced sections in initial partitioning hash. NFCI.
On my machine this reduced median link time of lld-speed-test/chrome
from 2.68s to 2.41s. It also reduces link time of Chrome for Android
with a prototype compiler change that causes the compiler to create
large numbers of identical (modulo relocations) sections from >15
minutes to a few seconds.

Differential Revision: https://reviews.llvm.org/D54773

llvm-svn: 347594
2018-11-26 21:15:47 +00:00
George Rimar c152281632 [LLD][ELF] - Do not crash when parsing the -defsym option from a error state.
When we are in a error state, script parser will not parse the -defsym
expression and hence will not tokenize it. Then ScriptLexer::Pos will be 0
and LLD will assert and crash here:

MemoryBufferRef ScriptLexer::getCurrentMB() {
  assert(!MBs.empty() && Pos > 0); // Bang !

Solution - stop parsing the defsym in a error state. That is consistent
with the regular case (when we parse the linker script).

llvm-svn: 347549
2018-11-26 12:29:56 +00:00
George Rimar a44c0f27c2 [LLD][ELF] - Remove the excessive safety return. NFC.
We explicitly call finalizeContents() only once for
DynamicSection. The code testing we do not do it twice is
just excessive.

It could be an assert, but we don't do
that for other sections, so does not seem we
should do it here too.

llvm-svn: 347543
2018-11-26 10:33:29 +00:00
George Rimar 0e0cd5be40 [LLD][ELF] - Add llvm_unreachable. NFC.
We never should call writeTo() for BSS section.

llvm-svn: 347540
2018-11-26 10:07:10 +00:00
George Rimar 0fc5dcd1c8 [LLD][ELF] - Simplify. NFCI.
This makes getRISCVPCRelHi20 to be static local helper,
and rotates the 'if' condition.

llvm-svn: 347497
2018-11-23 15:13:26 +00:00
George Rimar 94a16cb611 [ELF] - Make SymbolTable::addDefined return Defined.
Now it returns Symbol. This should be NFC that
avoids doing cast at the caller's sides.

Differential revision: https://reviews.llvm.org/D54627

llvm-svn: 347455
2018-11-22 11:40:08 +00:00
Fangrui Song f5badf4905 [ELF] Write IPLT header in -static -z retpolineplt mode
Summary:
This fixes PR39711: -static -z retpolineplt does not produce retpoline PLT header.
-z now is not relevant.

Statically linked executable does not have PLT, but may have IPLT with no header. When -z retpolineplt is specified, however, the repoline PLT header should still be emitted.

I've checked that this fixes the FreeBSD reproduce in PR39711 and a Linux program statically linked against glibc. The programm print "Hi" rather than SIGILL/SIGSEGV.

getPltEntryOffset may look dirty after this patch, but it can be cleaned up later.

Another possible improvement is that when there are non-preemptible IFUNC symbols (rare case, e.g. -Bsymbolic), both In.Plt and In.Iplt can be non-empty and we'll emit the retpoline PLT header twice.

Reviewers: espindola, emaste, chandlerc, ruiu

Reviewed By: emaste

Subscribers: emaste, arichardson, krytarowski, llvm-commits

Differential Revision: https://reviews.llvm.org/D54782

llvm-svn: 347404
2018-11-21 18:10:00 +00:00
Fangrui Song 537a5bfb7f [ELF] Allow --noinhibit-exec to produce corrupted executable with relocation overflow
Summary:
When --noinhibit-exec is used, ld.bfd/gold emit errors but allow to produce corrupted executable, which is handy for debugging purpose. lld's --noinhibit-exec has a different meaning and changes some errors to warnings. This patch replaces "error" with "errorOrWarn" to exploit that property.

We may revisit this: if we should keep them as errors (as ld.bfd/gold do) but allow to produce a (corrupted) executable.

Reviewers: ruiu, grimar, espindola

Reviewed By: grimar

Subscribers: Timmmm, jhenderson, emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D54651

llvm-svn: 347327
2018-11-20 17:04:11 +00:00
Fangrui Song 113d868ff1 Support ARM_V4BX relocation
Summary: This patch implementation the handler for ARM_V4BX. This relocation is used by GNU runtime files and other armv4 applications.

Patch by Yin Ma

Reviewers: espindola, MaskRay, ruiu, peter.smith, pcc

Reviewed By: ruiu

Subscribers: yinma, pcc, peter.smith, MaskRay, rovka, efriedma, emaste, javed.absar, arichardson, kristof.beyls, chrib, llvm-commits

Differential Revision: https://reviews.llvm.org/D53444

llvm-svn: 347077
2018-11-16 19:24:45 +00:00
Simon Atanasyan 275a5b442b [ELF][MIPS] Use MIPS R6 `sigrie` as a trap instruction
Current value using as a trap instruction (0xefefefef) is not a good choice
for MIPS because it's a valid MIPS instruction `swc3 $15,-4113(ra)`. This
patch replaces 0xefefefef by 0x04170001. For all MIPS ISA revisions before
R6, this value is just invalid instruction. Starting from MIPS R6 it's
a valid instruction `sigrie 1` which signals a Reserved Instruction exception.

mips-traps.s test case is added to test trap encoding. Other test cases
are modified to remove redundant checking.

Differential revision: https://reviews.llvm.org/D54154

llvm-svn: 347029
2018-11-16 05:30:47 +00:00
Sam Clegg d5f650631c [WebAssembly] Refactor config setting and checking. NFC.
This matches the way the ELF backend does it.

Differential Revision: https://reviews.llvm.org/D54559

llvm-svn: 346972
2018-11-15 18:09:41 +00:00
George Rimar 8329028b49 [ELF] - Renamed few more AArch64 specific relocation expressions. NFC.
They are AArch64 only, so have to have AARCH64_* prefix.

llvm-svn: 346963
2018-11-15 15:35:44 +00:00
Martin Storsjo 0db45d954b [ELF] Fix compilation with GCC 5
Remove the default initializer for TrapInstr; all subclasses overwrite
the defaults in their constructors anyway.

This fixes compilation errors like these, with GCC 5.4 on Ubuntu 16.04,
present since SVN r346893:

 In file included from ../tools/lld/ELF/Arch/AArch64.cpp:12:0:
 ../tools/lld/ELF/Target.h:125:49: error: array must be initialized with a brace-enclosed initializer
    std::array<uint8_t, 4> TrapInstr = {0, 0, 0, 0};
                                                  ^
 ../tools/lld/ELF/Target.h:125:49: error: too many initializers for ‘std::array<unsigned char, 4ul>’

Differential Revision: https://reviews.llvm.org/D54569

llvm-svn: 346934
2018-11-15 08:20:18 +00:00
Simon Atanasyan b0486051d2 [ELF] Make TrapInstr and Filler byte arrays. NFC.
The uint32_t type does not clearly convey that these fields are interpreted
in the target endianness. Converting them to byte arrays should make this
more obvious and less error-prone.

Patch by James Clarke

Differential Revision: http://reviews.llvm.org/D54207

llvm-svn: 346893
2018-11-14 21:05:20 +00:00
Sean Fertile 614dc11ca8 [PPC64] Long branch thunks.
On PowerPC64, when a function call offset is too large to encode in a call
instruction the address is stored in a table in the data segment. A thunk is
used to load the branch target address from the table relative to the
TOC-pointer and indirectly branch to the callee. When linking position-dependent
code the addresses are stored directly in the table, for position-independent
code the table is allocated and filled in at load time by the dynamic linker.

For position-independent code the branch targets could have gone in the .got.plt
but using the .branch_lt section for both position dependent and position
independent binaries keeps it consitent and helps keep this PPC64 specific logic
seperated from the target-independent code handling the .got.plt.

Differential Revision: https://reviews.llvm.org/D53408

llvm-svn: 346877
2018-11-14 17:56:43 +00:00
Peter Smith ad51cee866 [AArch64] Fix resolution of R_PLT_PAGE RelExpr
The R_AARCH64_ADR_PREL_PG_HI21 relocation type is given the R_PAGE_PC
RelExpr. This can be transformed to R_PLT_PAGE_PC via toPlt().
Unfortunately the resolution is identical to R_PAGE_PC so instead of
getting the address of the PLT entry we get the address of the symbol
which may not be correct in the case of static ifuncs. The fix is to
handle the cases separately and use getPltVA() + A with R_PLT_PAGE_PC.

Differential Revision: https://reviews.llvm.org/D54474

llvm-svn: 346863
2018-11-14 13:53:47 +00:00
Fangrui Song 0736461b24 [ELF] Rename NameTypeEntry to NameAttrEntry and its field "Type" to CuIndexAndAttrs
Summary:
NameTypeEntry::Type is a bit-packed value of CU index+attributes (https://sourceware.org/gdb//onlinedocs/gdb/Index-Section-Format.html), which is named cu_index_and_attrs in a local variable in gdb/dwarf2read.c:dw2_symtab_iter_next

The new name CuIndexAndAttrs is more meaningful.

Reviewers: ruiu, dblaikie, espindola

Reviewed By: dblaikie

Subscribers: emaste, aprantl, arichardson, JDevlieghere, arphaman, llvm-commits

Differential Revision: https://reviews.llvm.org/D54481

llvm-svn: 346794
2018-11-13 20:25:51 +00:00
George Rimar 8ef9babb67 [ELF] - Renamed AArch64 specific relocations expressions. NFC.
They did not have AArch64 prefix. Now they do.

llvm-svn: 346749
2018-11-13 10:16:36 +00:00
Fangrui Song 9596848037 [ELF] .gdb_index: fix CuOff when a .debug_info section contains more than 1 DW_TAG_compile_unit
Summary:
Idx passed to readPubNamesAndTypes was an index into Chunks, not an
index into the CU list. This would be incorrect if some .debug_info
section contained more than 1 DW_TAG_compile_unit.

In real world, glibc Scrt1.o is a partial link of start.os abi-note.o init.o and contains 2 CUs in debug builds.
Without this patch, any application linking such Scrt1.o would have invalid .gdb_index
The issue could be demonstrated by:

    (gdb) py print(gdb.lookup_global_symbol('main'))
    None

Reviewers: espindola, ruiu

Reviewed By: ruiu

Subscribers: Higuoxing, grimar, dblaikie, emaste, aprantl, arichardson, JDevlieghere, arphaman, llvm-commits

Differential Revision: https://reviews.llvm.org/D54361

llvm-svn: 346747
2018-11-13 08:43:07 +00:00
Fangrui Song 31be2f15a6 [ELF] Change GnuPub{Names,Types}Section from StringRef to LLDDWARFSection
Summary:
The debug_info_offset value may be relocated.

This is lld side change of D54375.

Reviewers: ruiu, dblaikie, grimar, espindola

Subscribers: emaste, arichardson, JDevlieghere, llvm-commits

Differential Revision: https://reviews.llvm.org/D54376

llvm-svn: 346616
2018-11-11 18:57:35 +00:00
Paul Robinson 746c22389c [DWARFv5] Read and dump multiple .debug_info sections.
Type units go in .debug_info comdats, not .debug_types, in v5.

Differential Revision: https://reviews.llvm.org/D53907

llvm-svn: 346360
2018-11-07 21:39:09 +00:00
Fangrui Song db22af0f45 [PPC64] Use INT32_MIN instead of std::numeric_limits<int32_t>::min()
Summary:
D53821 fixed the bogus MSVC (at least 2017) C4146 warning (unary minus applied on unsigned type)
by using std::numeric_limits<int32_t>::min().
The warning was because -2147483648 is incorrectly treated as unsigned long instead of long long)

Let's use INT32_MIN which is arguably more readable.
Note, on GCC or clang, -0x80000000 works fine (ILP64: long, LP64: long long).

Reviewers: ruiu, jhenderson, sfertile, espindola

Reviewed By: sfertile

Subscribers: emaste, nemanjai, arichardson, kbarton, jsji, llvm-commits

Differential Revision: https://reviews.llvm.org/D54200

llvm-svn: 346356
2018-11-07 21:14:54 +00:00
Galina Kistanova 4ec5d67e49 Handle OUTPUT_FORMAT arguments in quotation marks
llvm-svn: 346068
2018-11-03 05:25:49 +00:00
Fangrui Song 27d036e995 [ELF] Change sh_link of .rel{,a}.plt to make GNU strip happy
Summary:
D52830 sets sh_link to .symtab in static link, which breaks executable stripped by GNU strip.
It may also be odd that .rela.plt (SHF_ALLOC) points to .symtab (non-SHF_ALLOC).

Change the logic on pcc's suggestion.

Before:

% clang -fuse-ld=lld -static -xc =(printf 'int main(){}') # or gcc
% strip a.out; ./a.out
unexpected reloc type in static binary[1]    61634 segmentation fault  ./a.out

Reviewers: ruiu, grimar, emaste, espindola

Reviewed By: ruiu

Subscribers: pcc, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D53993

llvm-svn: 345899
2018-11-01 22:28:58 +00:00
Rui Ueyama 4d0059c09b Omit "virtual" if overridden.
llvm-svn: 345886
2018-11-01 20:08:39 +00:00
Rui Ueyama cace50c628 Use llvm_unreachable for unreachable code.
llvm-svn: 345885
2018-11-01 20:08:29 +00:00
Fangrui Song ccfc8415c2 Set MAttrs in LTO mode
Summary: Without this patch, MAttrs are not set.

Patch by Yin Ma

Reviewers: espindola, MaskRay, ruiu, pcc

Reviewed By: MaskRay, pcc

Subscribers: pcc, emaste, sbc100, inglorion, arichardson, aheejin, steven_wu, llvm-commits

Differential Revision: https://reviews.llvm.org/D53446

llvm-svn: 345884
2018-11-01 20:02:49 +00:00
George Rimar 3608decaa5 [ELF] - Do not crash when -r output uses linker script with `/DISCARD/`
This is https://bugs.llvm.org/show_bug.cgi?id=39493.

We crashed previously because did not handle /DISCARD/ properly
when -r was used. I think it is uncommon to use scripts with -r, though I see
nothing wrong to handle the /DISCARD/ so that we will not crash at least.

Differential revision: https://reviews.llvm.org/D53864

llvm-svn: 345819
2018-11-01 09:20:06 +00:00
Ryan Prichard e7cb0225a0 [ELF] Refactor per-target TLS layout configuration. NFC.
Summary:
There are really three different kinds of TLS layouts:

 * A fixed TLS-to-TP offset. On architectures like PowerPC, MIPS, and
   RISC-V, the thread pointer points to a fixed offset from the start
   of the executable's TLS segment. The offset is 0x7000 for PowerPC
   and MIPS, which allows a signed 16-bit offset to reach 0x1000 of
   per-thread implementation data and 0xf000 of the application's TLS
   segment. The size and layout of the TCB isn't relevant to the static
   linker and might not be known.

 * A fixed TCB size. This is the format documented as "variant 1" in
   Ulrich Drepper's TLS spec. The thread pointer points to a 2-word TCB
   followed by the executable's TLS segment. The first word is always
   the DTV pointer. Used on ARM. The thread pointer must be aligned to
   the TLS segment's alignment, possibly creating alignment padding.

 * Variant 2. This format predates variant 1 and is also documented in
   Drepper's TLS spec. It allocates the executable's TLS segment before
   the thread pointer, apparently for backwards-compatibility. It's
   used on x86 and SPARC.

Factor out an lld:🧝:getTlsTpOffset() function for use in a
follow-up patch for Android. The TcbSize/TlsTpOffset fields are only used
in getTlsTpOffset, so replace them with a switch on Config->EMachine.

Reviewers: espindola, ruiu, PkmX, jrtc27

Reviewed By: ruiu, PkmX, jrtc27

Subscribers: jyknight, emaste, sdardis, nemanjai, javed.absar, arichardson, kristof.beyls, kbarton, fedor.sergeev, atanasyan, PkmX, jsji, llvm-commits

Differential Revision: https://reviews.llvm.org/D53905

llvm-svn: 345775
2018-10-31 20:53:17 +00:00
Fangrui Song e0799a7268 [ELF] Fallback to sh_link=0 if neither .dynsym nor .symtab exists
Summary: .rela.plt may only contain R_*_{,I}RELATIVE relocations and not need a symbol table link. bfd/gold fallbacks to sh_link=0 in this case. Without this patch, ld.lld --strip-all caused lld to dereference a null pointer.

Reviewers: ruiu, grimar, espindola

Reviewed By: ruiu

Subscribers: emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D53881

llvm-svn: 345648
2018-10-30 20:54:54 +00:00
Sean Fertile 92964e74a5 [PPC64] Handle powerpc64 in OUTPUT_FORMAT.
Differential Revision: https://reviews.llvm.org/D53652

llvm-svn: 345604
2018-10-30 14:37:17 +00:00
James Henderson 1e74ecab17 [ELF][PPC64]Workaround bogus Visual Studio build warning
Visual Studio has a bug where it converts the integer literal 2147483648
into an unsigned int instead of a long long (i.e. it follows C89 rules).
The bug has been reported as:
https://developercommunity.visualstudio.com/content/problem/141813/-2147483648-c4146-error.html.

Because of this bug, we were getting a signed/unsigned comparison
warning in VS2015 from the old code (the subsequent unary negation had
no effect on the type).

Reviewed by: sfertile

Differential Revision: https://reviews.llvm.org/D53821

llvm-svn: 345579
2018-10-30 10:55:14 +00:00
Rui Ueyama 4c06a6cc90 Rename warnUnorderableSymbol maybeWarnUnorderableSymbol because the function doesn't always emit a warning.
llvm-svn: 345393
2018-10-26 15:07:12 +00:00
Rui Ueyama 2b39ea477c Refactor readCallGraph() and readCallGraphFromObjectFiles(). NFC.
llvm-svn: 345392
2018-10-26 15:07:02 +00:00
Fangrui Song cc18f8aa0f [ELF] Add --{,no-}call-graph-profile-sort (enabled by default)
Summary: Add an option to disable sorting sections with call graph profile

Reviewers: ruiu, Bigcheese, espindola

Reviewed By: Bigcheese

Subscribers: grimar, emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D53683

llvm-svn: 345332
2018-10-25 23:15:23 +00:00
Fangrui Song 018b0634b4 [ELF] Don't check if symbols in .llvm.call-graph-profile are unorderable
Summary: There are too many reasonable cases that would be considered unorderable.

Reviewers: ruiu, espindola, Bigcheese

Reviewed By: ruiu

Subscribers: grimar, emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D53669

llvm-svn: 345322
2018-10-25 22:15:42 +00:00
Rui Ueyama 9db06423d1 Remove full stops from error messages for consistency.
llvm-svn: 345294
2018-10-25 18:07:55 +00:00
Rui Ueyama 444363e931 Do not call computeIsPreemptible() if its result is discarded. NFC.
llvm-svn: 345208
2018-10-24 22:15:32 +00:00
Rui Ueyama 4664280a9c Add more blank lines so that code doesn't look too dense. NFC.
llvm-svn: 345205
2018-10-24 21:59:58 +00:00
Rui Ueyama 6e33f7ac66 Add a comment for PPC64 .toc and GNU relro.
llvm-svn: 345204
2018-10-24 21:59:48 +00:00
Rui Ueyama aa2948776a Include input section name and output section name in an error message.
Differential Revision: https://reviews.llvm.org/D53645

llvm-svn: 345172
2018-10-24 18:15:40 +00:00
Rui Ueyama 8d76024916 Refactor assignFileOffsets. NFC.
llvm-svn: 345154
2018-10-24 15:47:46 +00:00
Rui Ueyama 6d26ed92cc Split a function. NFC.
llvm-svn: 345143
2018-10-24 14:24:01 +00:00
Rui Ueyama 6e763a36c6 Add parenthese around a bitand.
llvm-svn: 345139
2018-10-24 13:44:51 +00:00
Rui Ueyama ee2b61cfd3 Make a local variable scope narrower. NFC.
llvm-svn: 345138
2018-10-24 13:42:33 +00:00
Martin Storsjo c3095ace86 [ELF] Remove a superfluous semicolon, fixing warnings. NFC.
llvm-svn: 345116
2018-10-24 07:42:06 +00:00
Rui Ueyama 02c7fae348 Move forward declarations to the top of the file and sort.
llvm-svn: 345094
2018-10-23 22:37:14 +00:00
Rui Ueyama b0de6c742a Move a function out of a class because it doesn't depend on any class member. NFC.
llvm-svn: 345093
2018-10-23 22:31:08 +00:00
Rui Ueyama 5a30177ef6 Factor out code to a new function. NFC.
llvm-svn: 345088
2018-10-23 22:03:33 +00:00
Rui Ueyama edc3201b11 Simplify. NFC.
A higher order function `applySyntehtic` can be replaced with a simpler function.

llvm-svn: 345081
2018-10-23 21:17:00 +00:00
Rui Ueyama 660e8721a9 Remove a global variable that is set but not used.
llvm-svn: 345080
2018-10-23 21:00:28 +00:00
Rui Ueyama 29d8639732 Add a comment.
llvm-svn: 345062
2018-10-23 17:52:44 +00:00
Rui Ueyama 42ab6c53f8 Remove a global variable that we can live without.
Out::DebugInfo was used only by GdbIndex class to determine if
we need to create a .gdb_index section, but we can do the same
check without it.

Added a test that this patch doesn't change the existing behavior.

llvm-svn: 345058
2018-10-23 17:39:43 +00:00
Dimitry Andric ad5aab8ddb Don't mess up RelIplt symbols during relocatable processing
Summary:
During upgrading of the FreeBSD source tree with lld 7.0.0, I noticed
that it started complaining about `crt1.o` having an "index past the
end of the symbol table".

Such a symbol table looks approximately like this, viewed with `readelf
-s` (note the `Ndx` field being messed up):

```
Symbol table '.symtab' contains 4 entries:
   Num:    Value  Size Type    Bind   Vis      Ndx Name
     0: 00000000     0 NOTYPE  LOCAL  DEFAULT  UND
     1: 00000000     0 SECTION LOCAL  DEFAULT    1
     2: 00000000     0 NOTYPE  WEAK   HIDDEN  RSV[0xffff] __rel_iplt_end
     3: 00000000     0 NOTYPE  WEAK   HIDDEN  RSV[0xffff] __rel_iplt_start
```

At first, it seemed that recent ifunc relocation work had caused this:
<https://reviews.freebsd.org/rS339351>, but it turned out that it was
due to incorrect processing of the object files by lld, when using `-r`
(a.k.a. --relocatable).

Bisecting showed that rL324421 ("Convert a use of Config->Static") was
the commit where this new behavior began.  Simply reverting it solved
the issue, and the `__rel_iplt` symbols had an index of `UND` again.

Looking at Rafael's commit message, I think he simply missed the
possibility of `--relocatable` being in effect, so I have added an
additional check for it.

I also added a simple regression test case.

Reviewers: grimar, ruiu, emaste, espindola

Reviewed By: ruiu

Subscribers: arichardson, krytarowski, llvm-commits

Differential Revision: https://reviews.llvm.org/D53515

llvm-svn: 345002
2018-10-23 05:53:15 +00:00
Shoaib Meenai c8e9d2b525 [ELF] Handle elf32-littlearm in OUTPUT_FORMAT
We need this to support 32-bit ARM. Add test cases for emulation
handling for this architecture as well.

Differential Revision: https://reviews.llvm.org/D53539

llvm-svn: 344976
2018-10-22 23:56:10 +00:00
Fangrui Song efc0fe5a72 [ELF] Don't warn on two legitimate cases when reading .llvm.call-graph-profile
Summary:
Before, superfluous warnings were emitted for the following two cases:

1) When from symbol was in a discarded section.
  The profile should be thought of as affiliated to the section.
  It makes sense to ignore the profile if the section is discarded.

2) When to symbol was in a shared object.
  The object file containing the profile may not know about the to
  symbol, which can reside in another object file (useful profile) or a
  shared object (not useful as symbols in the shared object are fixed
  and unorderable). It makes sense to ignore the profile from the object
  file.

  Note, the warning when to symbol was undefined was suppressed in
  D53044, which is still useful for --symbol-ordering-file=

This patch silences the warnings. The check is actually more relaxed (no
warnings if either From or To is not Defined) for simplicity and I don't
see a compelling reason to warn on more cases.

Reviewers: ruiu, davidxl, espindola, Bigcheese

Reviewed By: ruiu

Subscribers: emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D53470

llvm-svn: 344974
2018-10-22 23:43:53 +00:00
Rui Ueyama ea8cd00a1d Add OUTPUT_FORMAT linker script directive support.
This patch adds a support for OUTPUT_FORMAT linker script directive.
Since I'm not 100% confident with BFD names you can use in the directive
for all architectures, I added only a few in this patch. We can add
other names for other archtiectures later.

We still do not support triple-style OUTPUT_FORMAT directive, namely,
OUTPUT_FORMAT(bfdname, big, little). If you pass -EL (little endian)
or -EB (big endian) to the linker, GNU linkers pick up big or little
as a BFD name, correspondingly, so that you can use a single linker
script for bi-endian processor. I'm not sure if we really need to
support that, so I'll leave it alone for now.

Note that -m takes precedence over OUTPUT_FORAMT, but we always parse
a BFD name given to OUTPUT_FORMAT for error checking. You cannot write
an invalid name in the OUTPUT_FORMAT directive.

Differential Revision: https://reviews.llvm.org/D53495

llvm-svn: 344952
2018-10-22 20:50:01 +00:00
Fangrui Song e94d8823d9 [ELF][PPC64] Fix a split-stack comment in rLLD344622
The blt- instruction (predicted not to be taken) uses cr7, not the default cr0.

llvm-svn: 344948
2018-10-22 20:30:06 +00:00
Fangrui Song 8bafbfea9d [ELF] Format PPC64.cpp, NFC
Mainly two mis-indented places. The changes are local so should not interfere with in-review revisions

llvm-svn: 344932
2018-10-22 18:20:18 +00:00
Sam Clegg ea65647254 Use llvm::arrayRefFromStringRef
Differential Revision: https://reviews.llvm.org/D53432

llvm-svn: 344888
2018-10-22 08:35:39 +00:00
Alexander Richardson 39adc6df0d Add an addAbsolute static function to Writer.cpp
Summary:
SymbolTable::addAbsolute() was removed in rL344305.
To me this is more readable than the lambda named `Add` and in our
out-of-tree CHERI target we use addAbsolute() in another function.

Reviewers: ruiu, espindola

Reviewed By: ruiu

Subscribers: kristina, emaste, llvm-commits

Differential Revision: https://reviews.llvm.org/D53393

llvm-svn: 344842
2018-10-20 11:12:50 +00:00
Sean Fertile 3acfe400a2 [PPC64] Fix offset checks on rel24 call relocations.
Adjusted the range check on a call instruction from 24 bits signed to
26 bits signed. While the instruction only encodes 24 bits, the target is
assumed to be 4 byte aligned, and the value that is encoded in the instruction
gets shifted left by 2 to form the offset. Also added a check that the offset is
indeed at least 4 byte aligned.

Differential Revision: https://reviews.llvm.org/D53401

llvm-svn: 344747
2018-10-18 15:43:41 +00:00
Sean Fertile 4b5ec7fb80 Reland "[PPC64] Add split - stack support."
Recommitting https://reviews.llvm.org/rL344544 after fixing undefined behavior
from left-shifting a negative value. Original commit message:

This support is slightly different then the X86_64 implementation in that calls
to __morestack don't need to get rewritten to calls to __moresatck_non_split
when a split-stack caller calls a non-split-stack callee. Instead the size of
the stack frame requested by the caller is adjusted prior to the call to
__morestack. The size the stack-frame will be adjusted by is tune-able through a
new --split-stack-adjust-size option.

llvm-svn: 344622
2018-10-16 17:13:01 +00:00
Sean Fertile 831a1336ff Revert "[PPC64] Add split - stack support."
This reverts commit https://reviews.llvm.org/rL344544, which causes failures on
a undefined behaviour sanitizer bot -->
lld/ELF/Arch/PPC64.cpp:849:35: runtime error: left shift of negative value -1

llvm-svn: 344551
2018-10-15 20:20:28 +00:00
Sean Fertile 795cc9332b [PPC64] Add split - stack support.
This support is slightly different then the X86_64 implementation in that calls
to __morestack don't need to get rewritten to calls to __moresatck_non_split
when a split-stack caller calls a non-split-stack callee. Instead the size of
the stack frame requested by the caller is adjusted prior to the call to
__morestack. The size the stack-frame will be adjusted by is tune-able through a
new --split-stack-adjust-size option.

Differential Revision: https://reviews.llvm.org/D52099

llvm-svn: 344544
2018-10-15 19:05:57 +00:00
Sid Manning 6df0759aad [ELF][HEXAGON] Let input determine e_flag.
Differential Revision: https://reviews.llvm.org/D53204

llvm-svn: 344518
2018-10-15 14:44:52 +00:00
George Rimar ee981860a8 [LLD][ELF] - Check options before processing the -v/-version options.
This is https://bugs.llvm.org/show_bug.cgi?id=39289.

Currently both gold and bfd report errors about invalid options values
even with -v/-versions. But LLD does not.

This makes complicated to check the options available when LLD is used.

Patch makes LLD behavior to be consistent with GNU linkers.

Differential revision: https://reviews.llvm.org/D53278

llvm-svn: 344514
2018-10-15 14:21:43 +00:00
Rui Ueyama 3d354081e0 Simplify. NFC.
- Removed redundant `llvm::`
 - Typedef a long type name
 - Initialize members by member initializers

llvm-svn: 344427
2018-10-12 22:44:06 +00:00
Rui Ueyama f3fad55787 Remove `Type` parameter from SymbolTable::insert(). NFC.
`Type` parameter was used only to check for TLS attribute mismatch,
but we can do that when we actually replace symbols, so we don't need
to type as an argument. This change should simplify the interface of
the symbol table a bit.

llvm-svn: 344394
2018-10-12 18:29:18 +00:00
Rui Ueyama 2e9d40d5f4 [lld] Add more complete support for the INCLUDE command.
Patch by Ian Tessier.

This change adds INCLUDE support to the MEMORY and SECTION commands, and
to output sections, as per:

https://sourceware.org/binutils/docs/ld/File-Commands.html#File-Commands

Differential Revision: https://reviews.llvm.org/D52951

llvm-svn: 344368
2018-10-12 17:07:32 +00:00
George Rimar 4cb05713de [ELF] - Remove unused variable. NFC.
It triggered a bot failture somehow today:
http://lab.llvm.org:8011/builders/lld-perf-testsuite/builds/7993

llvm-svn: 344344
2018-10-12 13:39:35 +00:00
George Rimar fa91327b51 [ELF] - Remove dead declaration. NFC.
MipsGotSection::FileGot::isOverflow()
was a dead declararion.

llvm-svn: 344342
2018-10-12 13:25:25 +00:00
Rui Ueyama b0d8502049 Remove SymbolTable::addAbsolute().
addAbsolute() could be implemented as a non-member function.

llvm-svn: 344305
2018-10-11 22:15:41 +00:00
Rui Ueyama 936cfa86e5 Remove `else` after `return`.
llvm-svn: 344302
2018-10-11 21:57:52 +00:00
Eli Friedman 8b44cc21c6 [ELF] Fix link failure with Android compressed relocation support.
Android uses a compressed relocation format, which means the size of the
relocation section isn't predictable based on the number of relocations,
and can vary if the layout changes in any way. To deal with this, the
linker normally runs multiple passes until the layout converges.

The layout should converge if the size of the compressed
relocation section increases monotonically: if the size of an encoded
offset increases by one byte, the larget value which can be encoded is
multiplied by 128, so the representable offsets grow much faster than
the size of the section itself.

The problem here is that there is no code to ensure the size of the
section doesn't decrease.  If the size of the relocation section
decreases, the relative offsets can increase due to alignment
restrictions, so that can force the size of the relocation section to
increase again.  The end result is an infinite loop; the loop gets cut
off after 10 iterations with the message "thunk creation not
converged".

To avoid this issue, this patch adds padding to the end of the
relocation section if its size would decrease.  The extra
padding is harmless because of the way the format is defined:
decoding stops after it reaches the number of relocations specified
in the section's header.

Differential Revision: https://reviews.llvm.org/D53003

llvm-svn: 344300
2018-10-11 21:43:06 +00:00
Rui Ueyama 2600e0946b Rename SymbolTable::addRegular -> SymbolTable::addDefined.
We have addAbsolute, addBitcode, addCommon, etc. addRegular looked a
bit inconsistent.

llvm-svn: 344294
2018-10-11 20:43:01 +00:00
Rui Ueyama cf41adaaa0 Remove unused default arguments.
llvm-svn: 344292
2018-10-11 20:38:34 +00:00
Rui Ueyama c7497d3ac5 Remove SymbolTable::addUndefined<ELF32LE>(StringRef).
Because we can implement the function as a non-member function.

llvm-svn: 344290
2018-10-11 20:34:29 +00:00
George Rimar 3368643f5e [ELF] - Set sh_info and sh_link for .rela.plt sections.
This is https://bugs.llvm.org/show_bug.cgi?id=37538,

Currently, LLD may set both sh_link and sh_info for
.rela.plt section to zero when we have only .rela.iplt section part used.

ELF spec (https://docs.oracle.com/cd/E19683-01/816-1386/chapter6-94076/index.html)
says that for SHT_REL and SHT_RELA, sh_link references the associated symbol table
and sh_info the "section to which the relocation applies."

When we set the sh_link field, for the regular case we use the .dynsym index.
For .rela.iplt sections, it is unclear what is the associated symbol table,
because R_*_RELATIVE relocations do not use symbol names and we might have no
.dynsym section at all so this patch uses .symtab section index.

Differential revision: https://reviews.llvm.org/D52830

llvm-svn: 344226
2018-10-11 08:25:35 +00:00
Rui Ueyama 3c4344810a Make a member function private and rename it to avoid function overloading.
llvm-svn: 344196
2018-10-10 22:49:29 +00:00
Fangrui Song 11ca54f49c [ELF] Don't warn on undefined symbols if UnresolvedPolicy::Ignore is used
Summary:
Add a condition UnresolvedPolicy::Ignore to elf::warnUnorderedSymbol to suppress Sym->isUndefined() warnings from both

1) --symbol-ordering-file=
2) .llvm.call-graph-profile

If --unresolved-symbols=ignore-all is used,

  no "undefined symbol" error/warning is emitted. It makes sense to not warn unorderable symbols.

Otherwise,

  If an executable is linked, the default policy UnresolvedPolicy::ErrorOrWarn will issue a "undefined symbol" error. The unorderable symbol warning is redundant.

  If a shared object is linked, it is possible that only part of object files are used and some symbols are left undefined. The warning is not very necessary.
    In particular for .llvm.call-graph-profile, when linking a shared object, a call graph profile may contain undefined symbols. This case generated a warning before but it will be suppressed by this patch.

Reviewers: ruiu, davidxl, espindola

Reviewed By: ruiu

Subscribers: grimar, emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D53044

llvm-svn: 344195
2018-10-10 22:48:57 +00:00
Rui Ueyama 9ae88c64fa Eliminate dependency to formatv(). NFC.
llvm-svn: 344177
2018-10-10 20:29:29 +00:00
Fangrui Song 2043a58abe Adapt OptTable::PrintHelp change in D51009
Summary: Before, OptTable::PrintHelp append "[options] <inputs>" to its parameter `Help`. It is more flexible to change its semantic to `Usage` and let user customize the usage line.

Reviewers: rupprecht, ruiu, espindola

Reviewed By: rupprecht

Subscribers: emaste, sbc100, arichardson, aheejin, llvm-commits

Differential Revision: https://reviews.llvm.org/D53054

llvm-svn: 344099
2018-10-10 00:15:36 +00:00
Rui Ueyama 71cdbb7fe6 Merge two overloaded functions into one function. NFC.
llvm-svn: 344089
2018-10-09 22:44:53 +00:00
Rui Ueyama e65cb4889e Return early. NFC.
llvm-svn: 344088
2018-10-09 22:44:42 +00:00
Rui Ueyama 2b53b4bea6 Attempt to fix ubsan.
Previously, we cast a pointer to Elf{32,64}_Chdr like this

  auto *Hdr = reinterpret_cast<const ELF64_Chdr>(Ptr);

and read from its members like this

  read32(&Hdr->ch_size);

I was thinking that this does not violate alignment requirement,
since &Hdr->ch_size doesn't really access memory, but seems like
it is a violation in terms of C++ spec (?)

In this patch, I use a different struct that allows unaligned access.

llvm-svn: 344083
2018-10-09 21:41:53 +00:00
Rui Ueyama e3ccd28660 Calculate an offset only once. NFC.
llvm-svn: 344076
2018-10-09 20:29:24 +00:00
Rui Ueyama fbc629702d Remove redundant `Symtab->`.
`SymbolTable` is a singleton class and is a global variable for the
unique instance, so we can always refer the symtab by `Symtab->`.
However, we don't need to use the global varaible from member functions
of SymbolTable class.

llvm-svn: 344075
2018-10-09 20:22:18 +00:00
Rui Ueyama 714abece2b Simplify. NFC.
llvm-svn: 344074
2018-10-09 20:16:16 +00:00
Rui Ueyama 659cff37d6 Remove a use of template to make code less abstracted.
Sometimes, code that is a bit longer but doesn't use template is
easier to understand than code that uses template.

llvm-svn: 344072
2018-10-09 19:54:32 +00:00
Rui Ueyama e28c146423 Avoid unnecessary buffer allocation and memcpy for compressed sections.
Previously, we uncompress all compressed sections before doing anything.
That works, and that is conceptually simple, but that could results in
a waste of CPU time and memory if uncompressed sections are then
discarded or just copied to the output buffer.

In particular, if .debug_gnu_pub{names,types} are compressed and if no
-gdb-index option is given, we wasted CPU and memory because we
uncompress them into newly allocated bufers and then memcpy the buffers
to the output buffer. That temporary buffer was redundant.

This patch changes how to uncompress sections. Now, compressed sections
are uncompressed lazily. To do that, `Data` member of `InputSectionBase`
is now hidden from outside, and `data()` accessor automatically expands
an compressed buffer if necessary.

If no one calls `data()`, then `writeTo()` directly uncompresses
compressed data into the output buffer. That eliminates the redundant
memory allocation and redundant memcpy.

This patch significantly reduces memory consumption (20 GiB max RSS to
15 Gib) for an executable whose .debug_gnu_pub{names,types} are in total
5 GiB in an uncompressed form.

Differential Revision: https://reviews.llvm.org/D52917

llvm-svn: 343979
2018-10-08 16:58:59 +00:00
Sid Manning 307c7901d0 [ELF][HEXAGON] Add R_HEX_GOT_16_X support
Differential Revision: https://reviews.llvm.org/D52909

llvm-svn: 343972
2018-10-08 15:32:46 +00:00
Sid Manning 261eec5fa5 [ELF][HEXAGON] Add support for GOT relocations.
The GOT is referenced through the symbol _GLOBAL_OFFSET_TABLE_ .

The relocation added calculates the offset into the global offset table for
the entry of a symbol.  In order to get the correct TargetVA I needed to
create an new relocation expression, HEXAGON_GOT.  It does
Sym.getGotVA() - In.GotPlt->getVA().

Differential Revision: https://reviews.llvm.org/D52744

llvm-svn: 343784
2018-10-04 14:54:17 +00:00