Commit Graph

6003 Commits

Author SHA1 Message Date
Fangrui Song b498d99338 [ELF] Start a new PT_LOAD if LMA region is different
GNU ld has a counterintuitive lang_propagate_lma_regions rule.

```
// .foo's LMA region is propagated to .bar because their VMA region is the same,
// and .bar does not have an explicit output section address (addr_tree).
.foo : { *(.foo) } >RAM AT> FLASH
.bar : { *(.bar) } >RAM

// An explicit output section address disables propagation.
.foo : { *(.foo) } >RAM AT> FLASH
.bar . : { *(.bar) } >RAM
```

In both cases, lld thinks .foo's LMA region is propagated and
places .bar in the same PT_LOAD, so lld diverges from GNU ld w.r.t. the
second case (lma-align.test).

This patch changes Writer<ELFT>::createPhdrs to disable propagation
(start a new PT_LOAD). A user of the first case can make linker scripts
portable by explicitly specifying `AT>`. By contrast, there was no
workaround for the old behavior.

This change uncovers another LMA related bug in assignOffsets() where
`ctx->lmaOffset = 0;` was omitted. It caused a spurious "load address
range overlaps" error for at2.test

The new PT_LOAD rule is complex. For convenience, I listed the origins of some subexpressions:

* rL323449: `sec->memRegion == load->firstSec->memRegion`; linkerscript/at3.test
* D43284: `load->lastSec == Out::programHeaders` (don't start a new PT_LOAD after program headers); linkerscript/at4.test
* D58892: `sec != relroEnd` (start a new PT_LOAD after PT_GNU_RELRO)

Reviewed By: psmith

Differential Revision: https://reviews.llvm.org/D74297
2020-02-12 08:20:14 -08:00
Fangrui Song e21b9ca751 [ELF] Respect output section alignment for AT> (non-null lmaRegion)
When lmaRegion is non-null, respect `sec->alignment`
This rule is analogous to `switchTo(sec)` which advances sh_addr (VMA).

This fixes the p_paddr misalignment issue as reported by
https://android-review.googlesource.com/c/trusty/external/trusted-firmware-a/+/1230058

Note, `sec->alignment` is the maximum of ALIGN and input section alignments. We may overalign LMA than GNU ld.

linkerscript/align-lma.s has a FIXME that demonstrates another bug:
`.bss ... >RAM` should be placed in a different PT_LOAD (GNU ld
behavior) because its lmaRegion (nullptr) is different from the previous
section's lmaRegion (ROM).

Reviewed By: psmith

Differential Revision: https://reviews.llvm.org/D74286
2020-02-12 08:19:42 -08:00
Fangrui Song 9f854c0489 [ELF][RISCV] Add R_RISCV_IRELATIVE
https://github.com/riscv/riscv-elf-psabi-doc/pull/131 assigned 58 to R_RISCV_IRELATIVE.

Differential Revision: https://reviews.llvm.org/D74022
2020-02-10 20:22:39 -08:00
Fangrui Song 5f38040359 [ELF] Simplify parsing of version dependency. NFC 2020-02-08 14:10:29 -08:00
Nico Weber c29003813a Revert "[LLD][ELF][ARM] Do not substitute BL/BLX for non STT_FUNC symbols."
There are still problems after the fix in
"[ELF][ARM] Fix regression of BL->BLX substitution after D73542"
so let's revert to get trunk back to green while we investigate.
See https://reviews.llvm.org/D73542

This reverts commit 5461fa2b1f.
This reverts commit 0b4a047bfb.
2020-02-07 08:55:52 -05:00
Russell Gallop e7cb374433 [LLD][ELF] Add time-trace to ELF LLD
This adds some of LLD specific scopes and picks up optimisation scopes
via LTO/ThinLTO. Makes use of TimeProfiler multi-thread support added in
77e6bb3c.

Differential Revision: https://reviews.llvm.org/D71060
2020-02-06 12:14:13 +00:00
Fangrui Song 5461fa2b1f [ELF][ARM] Fix regression of BL->BLX substitution after D73542
D73542 made a typo (`rel.type == R_PLT_PC`; should be `rel.expr`) and introduced a regression:
BL->BLX substitution was disabled when the target symbol is preemptible
(expr is R_PLT_PC).

The two added bl instructions in arm-thumb-interwork-shared.s check that
we patch BL to BLX.

Fixes https://bugs.chromium.org/p/chromium/issues/detail?id=1047531
2020-02-05 14:09:14 -08:00
Fangrui Song da1973a241 [ELF][Mips] Drop an unneeded config->relocatable check 2020-01-31 21:00:28 -08:00
Jonas Devlieghere 3e24242a7d [lld] Replace SmallStr.str().str() with std::string conversion operator.
Use the std::string conversion operator introduced in
d7049213d0.
2020-01-29 21:30:21 -08:00
Fangrui Song 4a4ce14eb2 [ELF] Mention symbol name in reportRangeError()
For an out-of-range relocation referencing a non-local symbol, report the symbol name and the object file that defines the symbol. As an example:
```
t.o:(function func: .text.func+0x3): relocation R_X86_64_32S out of range: -281474974609120 is not in [-2147483648, 2147483647]
```
=>
```
t.o:(function func: .text.func+0x3): relocation R_X86_64_32S out of range: -281474974609120 is not in [-2147483648, 2147483647]; references func
>>> defined in t1.o
```

Reviewed By: grimar

Differential Revision: https://reviews.llvm.org/D73518
2020-01-29 09:38:25 -08:00
Peter Smith 0b4a047bfb [LLD][ELF][ARM] Do not substitute BL/BLX for non STT_FUNC symbols.
D73474 disabled the generation of interworking thunks for branch
relocations to non STT_FUNC symbols. This patch handles the case of BL and
BLX instructions to non STT_FUNC symbols. LLD would normally look at the
state of the caller and the callee and write a BL if the states are the
same and a BLX if the states are different.

This patch disables BL/BLX substitution when the destination symbol does
not have type STT_FUNC. This brings our behavior in line with GNU ld which
may prevent difficult to diagnose runtime errors when switching to lld.

Differential Revision: https://reviews.llvm.org/D73542
2020-01-29 11:42:25 +00:00
Benjamin Kramer adcd026838 Make llvm::StringRef to std::string conversions explicit.
This is how it should've been and brings it more in line with
std::string_view. There should be no functional change here.

This is mostly mechanical from a custom clang-tidy check, with a lot of
manual fixups. It uncovers a lot of minor inefficiencies.

This doesn't actually modify StringRef yet, I'll do that in a follow-up.
2020-01-28 23:25:25 +01:00
Fangrui Song e11b709b19 [ELF][PPC32] Support --emit-relocs link of R_PPC_PLTREL24
Similar to R_MIPS_GPREL16 and R_MIPS_GPREL32 (D45972).

If the addend of an R_PPC_PLTREL24 is >= 0x8000, it indicates that r30
is relative to the input section .got2.

```
addis 30, 30, .got2+0x8000-.L1$pb@ha
addi 30, 30, .got2+0x8000-.L1$pb@l
...
bl foo+0x8000@PLT
```

After linking, the relocation will be relative to the output section .got2.
To compensate for the shift `address(input section .got2) - address(output section .got2) = ppc32Got2OutSecOff`, adjust by `ppc32Got2OutSecOff`:

```
addis 30, 30, .got2+0x8000-.L1+ppc32Got2OutSecOff$pb@ha
addi 30, 30, .got2+0x8000-.L1+ppc32Got2OutSecOff$pb@ha$pb@l
...
bl foo+0x8000+ppc32Got2OutSecOff@PLT
```

This rule applys to a relocatable link or a non-relocatable link with --emit-relocs.

Reviewed By: Bdragon28

Differential Revision: https://reviews.llvm.org/D73532
2020-01-28 11:04:04 -08:00
Peter Smith 4f38ab250f [LLD][ELF][ARM] Do not insert interworking thunks for non STT_FUNC symbols
ELF for the ARM architecture requires linkers to provide
interworking for symbols that are of type STT_FUNC. Interworking for
other symbols must be encoded directly in the object file. LLD was always
providing interworking, regardless of the symbol type, this breaks some
programs that have branches from Thumb state targeting STT_NOTYPE symbols
that have bit 0 clear, but they are in fact internal labels in a Thumb
function. LLD treats these symbols as ARM and inserts a transition to Arm.

This fixes the problem for in range branches, R_ARM_JUMP24,
R_ARM_THM_JUMP24 and R_ARM_THM_JUMP19. This is expected to be the vast
majority of problem cases as branching to an internal label close to the
function.

There is at least one follow up patch required.
- R_ARM_CALL and R_ARM_THM_CALL may do interworking via BL/BLX
  substitution.

In theory range-extension thunks can be altered to not change state when
the symbol type is not STT_FUNC. I will need to check with ld.bfd to see if
this is the case in practice.

Fixes (part of) https://github.com/ClangBuiltLinux/linux/issues/773

Differential Revision: https://reviews.llvm.org/D73474
2020-01-28 11:54:18 +00:00
Peter Smith 3238b03c19 [LLD][ELF][ARM] clang-format function signature [NFC]
ARM::needsThunk had gone over 80 characters, run clang-format over it to
prevent it wrapping.
2020-01-28 11:54:18 +00:00
Teresa Johnson 2f63d549f1 Restore "[LTO/WPD] Enable aggressive WPD under LTO option"
This restores 59733525d3 (D71913), along
with bot fix 19c76989bb.

The bot failure should be fixed by D73418, committed as
af954e441a.

I also added a fix for non-x86 bot failures by requiring x86 in new test
lld/test/ELF/lto/devirt_vcall_vis_public.ll.
2020-01-27 07:55:05 -08:00
Fangrui Song 70389be7a0 [ELF][PPC32] Support range extension thunks with addends
* Generalize the code added in D70637 and D70937. We should eventually remove the EM_MIPS special case.
* Handle R_PPC_LOCAL24PC the same way as R_PPC_REL24.

Reviewed By: Bdragon28

Differential Revision: https://reviews.llvm.org/D73424
2020-01-25 22:32:42 -08:00
Fangrui Song 837e8a9c0c [ELF][PPC32] Support canonical PLT
-fno-pie produces a pair of non-GOT-non-PLT relocations R_PPC_ADDR16_{HA,LO} (R_ABS) referencing external
functions.

```
lis 3, func@ha
la 3, func@l(3)
```

In a -no-pie/-pie link, if func is not defined in the executable, a canonical PLT entry (st_value>0, st_shndx=0) will be needed.
References to func in shared objects will be resolved to this address.
-fno-pie -pie should fail with "can't create dynamic relocation ... against ...", so we just need to think about -no-pie.

On x86, the PLT entry passes the JMP_SLOT offset to the rtld PLT resolver.
On x86-64: the PLT entry passes the JUMP_SLOT index to the rtld PLT resolver.
On ARM/AArch64: the PLT entry passes &.got.plt[n]. The PLT header passes &.got.plt[fixed-index]. The rtld PLT resolver can compute the JUMP_SLOT index from the two addresses.

For these targets, the canonical PLT entry can just reuse the regular PLT entry (in PltSection).

On PPC32: PltSection (.glink) consists of `b PLTresolve` instructions and `PLTresolve`. The rtld PLT resolver depends on r11 having been set up to the .plt (GotPltSection) entry.
On PPC64 ELFv2: PltSection (.glink) consists of `__glink_PLTresolve` and `bl __glink_PLTresolve`. The rtld PLT resolver depends on r12 having been set up to the .plt (GotPltSection) entry.

We cannot reuse a `b PLTresolve`/`bl __glink_PLTresolve` in PltSection as a canonical PLT entry. PPC64 ELFv2 avoids the problem by using TOC for any external reference, even in non-pic code, so the canonical PLT entry scenario should not happen in the first place.
For PPC32, we have to create a PLT call stub as the canonical PLT entry. The code sequence sets up r11.

Reviewed By: Bdragon28

Differential Revision: https://reviews.llvm.org/D73399
2020-01-25 17:56:37 -08:00
Fangrui Song deb5819d62 [ELF] Rename relocateOne() to relocate() and pass `Relocation` to it
Symbol information can be used to improve out-of-range/misalignment diagnostics.
It also helps R_ARM_CALL/R_ARM_THM_CALL which has different behaviors with different symbol types.

There are many (67) relocateOne() call sites used in thunks, {Arm,AArch64}errata, PLT, etc.
Rename them to `relocateNoSym()` to be clearer that there is no symbol information.

Reviewed By: grimar, peter.smith

Differential Revision: https://reviews.llvm.org/D73254
2020-01-25 12:00:18 -08:00
Fangrui Song f1dab29908 [ELF][PowerPC] Support R_PPC_COPY and R_PPC64_COPY
Reviewed By: Bdragon28, jhenderson, grimar, sfertile

Differential Revision: https://reviews.llvm.org/D73255
2020-01-24 09:06:20 -08:00
Teresa Johnson 90e630a95e Revert "[LTO/WPD] Enable aggressive WPD under LTO option"
This reverts commit 59733525d3.

There is a windows sanitizer bot failure in one of the cfi tests
that I will need some time to figure out:
http://lab.llvm.org:8011/builders/sanitizer-windows/builds/57155/steps/stage%201%20check/logs/stdio
2020-01-23 17:29:24 -08:00
Teresa Johnson 59733525d3 [LTO/WPD] Enable aggressive WPD under LTO option
Summary:
Third part in series to support Safe Whole Program Devirtualization
Enablement, see RFC here:
http://lists.llvm.org/pipermail/llvm-dev/2019-December/137543.html

This patch adds type test metadata under -fwhole-program-vtables,
even for classes without hidden visibility. It then changes WPD to skip
devirtualization for a virtual function call when any of the compatible
vtables has public vcall visibility.

Additionally, internal LLVM options as well as lld and gold-plugin
options are added which enable upgrading all public vcall visibility
to linkage unit (hidden) visibility during LTO. This enables the more
aggressive WPD to kick in based on LTO time knowledge of the visibility
guarantees.

Support was added to all flavors of LTO WPD (regular, hybrid and
index-only), and to both the new and old LTO APIs.

Unfortunately it was not simple to split the first and second parts of
this part of the change (the unconditional emission of type tests and
the upgrading of the vcall visiblity) as I needed a way to upgrade the
public visibility on legacy WPD llvm assembly tests that don't include
linkage unit vcall visibility specifiers, to avoid a lot of test churn.

I also added a mechanism to LowerTypeTests that allows dropping type
test assume sequences we now aggressively insert when we invoke
distributed ThinLTO backends with null indexes, which is used in testing
mode, and which doesn't invoke the normal ThinLTO backend pipeline.

Depends on D71907 and D71911.

Reviewers: pcc, evgeny777, steven_wu, espindola

Subscribers: emaste, Prazek, inglorion, arichardson, hiraditya, MaskRay, dexonsmith, dang, davidxl, cfe-commits, llvm-commits

Tags: #clang, #llvm

Differential Revision: https://reviews.llvm.org/D71913
2020-01-23 16:09:44 -08:00
Fangrui Song 0fbf28f7aa [ELF] --no-dynamic-linker: don't emit undefined weak symbols to .dynsym
I felt really sad to push this commit for my selfish purpose to make
glibc -static-pie build with lld. Some code constructs in glibc require
R_X86_64_GOTPCREL/R_X86_64_REX_GOTPCRELX referencing undefined weak to
be resolved to a GOT entry not relocated by R_X86_64_GLOB_DAT (GNU ld
behavior), e.g.

csu/libc-start.c
  if (__pthread_initialize_minimal != NULL)
    __pthread_initialize_minimal ();

elf/dl-object.c
  void
  _dl_add_to_namespace_list (struct link_map *new, Lmid_t nsid)
  {
    /* We modify the list of loaded objects.  */
    __rtld_lock_lock_recursive (GL(dl_load_write_lock));

Emitting a GLOB_DAT will make the address equal &__ehdr_start (true
value) and cause elf/ldconfig to segfault. glibc really should move away
from weak references, which do not have defined semantics.

Temporarily special case --no-dynamic-linker.
2020-01-23 12:25:15 -08:00
Fangrui Song 1e57038bf2 [ELF] Pass `Relocation` to relaxGot and relaxTls{GdToIe,GdToLe,LdToLe,IeToLe}
These functions call relocateOne(). This patch is a prerequisite for
making relocateOne() aware of `Symbol` (D73254).

Reviewed By: grimar, nickdesaulniers

Differential Revision: https://reviews.llvm.org/D73250
2020-01-23 10:39:25 -08:00
Thomas Preud'homme c42fe24754 [lld/ELF] PR44498: Support input filename in double quote
Summary:
Linker scripts allow filenames to be put in double quotes to prevent
characters in filenames that are part of the linker script syntax from
having their special meaning. Case in point the * wildcard character.

Availability of double quoting filenames also allows to fix a failure in
ELF/linkerscript/filename-spec.s when the path contain a @ which the
lexer consider as a special characters and thus break up a filename
containing it. This may happens under Jenkins which createspath such as
pipeline@2.

To avoid the need for escaping GlobPattern metacharacters in filename
in double quotes, GlobPattern::create is augmented with a new parameter
to request literal matching instead of relying on the presence of a
wildcard character in the pattern.

Reviewers: jhenderson, MaskRay, evgeny777, espindola, alexshap

Reviewed By: MaskRay

Subscribers: peter.smith, grimar, ruiu, emaste, arichardson, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72517
2020-01-22 12:03:10 +00:00
Peter Smith e727f39ec0 [LLD][ELF][ARM] Don't apply --fix-cortex-a8 to relocatable links.
The --fix-cortex-a8 is sensitive to alignment and the precise destination
of branch instructions. These are not knowable at relocatable link time. We
follow GNU ld and the --fix-cortex-a53-843419 (D72968) by not patching the
code when there is a relocatable link.

Differential Revision: https://reviews.llvm.org/D73100
2020-01-22 11:03:40 +00:00
Sid Manning 6b9a5e6f05 [lld][Hexagon] Add General Dynamic relocations (GD)
Differential revision: https://reviews.llvm.org/D72522
2020-01-21 14:10:03 -06:00
Andrew Ng 4e8116f469 [ELF] Refactor uses of getInputSections to improve efficiency NFC
Add new method getFirstInputSection and use instead of getInputSections
where appropriate to avoid creation of an unneeded vector of input
sections.

Differential Revision: https://reviews.llvm.org/D73047
2020-01-21 12:27:52 +00:00
Peter Smith dbd0ad3366 [LLD][ELF] Add support for INPUT_SECTION_FLAGS
The INPUT_SECTION_FLAGS linker script command is used to constrain the
section pattern matching to sections that match certain combinations of
flags.

There are two ways to express the constraint.
withFlags: Section must have these flags.
withoutFlags: Section must not have these flags.

The syntax of the command is:
INPUT_SECTION_FLAGS '(' sect_flag_list ')'
sect_flag_list: NAME
| sect_flag_list '&' NAME

Where NAME matches a section flag name such as SHF_EXECINSTR, or the
integer value of a section flag. If the first character of NAME is ! then
it means must not contain flag.

We do not support the rare case of { INPUT_SECTION_FLAGS(flags) filespec }
where filespec has no input section description like (.text).

As an example from the ld man page:
SECTIONS {
  .text : { INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) }
  .text2 :  { INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) }
}
.text will match sections called .text that have both the SHF_MERGE and
SHF_STRINGS flag.
.text2 will match sections called .text that don't have the SHF_WRITE flag.

The flag names accepted are the generic to all targets and SHF_ARM_PURECODE
as it is very useful to filter all the pure code sections into a single
program header that can be marked execute never.

fixes PR44265

Differential Revision: https://reviews.llvm.org/D72756
2020-01-21 10:05:26 +00:00
James Clarke d1da63664f [lld][RISCV] Print error when encountering R_RISCV_ALIGN
Summary:
Unlike R_RISCV_RELAX, which is a linker hint, R_RISCV_ALIGN requires the
support of the linker even when ignoring all R_RISCV_RELAX relocations.
This is because the compiler emits as many NOPs as may be required for
the requested alignment, more than may be required pre-relaxation, to
allow for the target becoming more unaligned after relaxing earlier
sequences. This means that the target is often not initially aligned in
the object files, and so the R_RISCV_ALIGN relocations cannot just be
ignored. Since we do not support linker relaxation, we must turn these
into errors.

Reviewers: ruiu, MaskRay, espindola

Reviewed By: MaskRay

Subscribers: grimar, Jim, emaste, arichardson, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, kito-cheng, shiva0217, zzheng, edward-jones, rogfer01, MartinMosbeck, brucehoult, the_o, rkruppe, PkmX, jocewei, psnobl, benna, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D71820
2020-01-21 02:49:45 +00:00
Eli Friedman c81fe34718 [lld][ELF] Don't apply --fix-cortex-a53-843419 to relocatable links.
The code doesn't apply the fix correctly to relocatable links. I could
try to fix the code that applies the fix, but it's pointless: we don't
actually know what the offset will be in the final executable. So just
ignore the flag for relocatable links.

Issue discovered building Android.

Differential Revision: https://reviews.llvm.org/D72968
2020-01-20 15:27:41 -08:00
Fangrui Song 6ab89c3c5d [ELF] Allow R_PLT_PC (R_PC) to a hidden undefined weak symbol
This essentially reverts b841e119d7.

Such code construct can be used in the following way:

  // glibc/stdlib/exit.c
  // clang -fuse-ld=lld => succeeded
  // clang -fuse-ld=lld -fpie -pie => relocation R_PLT_PC cannot refer to absolute symbol
  __attribute__((weak, visibility("hidden"))) extern void __call_tls_dtors();
  void __run_exit_handlers() {
    if (__call_tls_dtors)
        __call_tls_dtors();
  }

Since we allow R_PLT_PC in -no-pie mode, it makes sense to allow it in
-pie mode as well.

Reviewed By: pcc

Differential Revision: https://reviews.llvm.org/D72943
2020-01-17 13:06:42 -08:00
Peter Smith 01ad4c8384 [LLD][ELF][ARM][AArch64] Only round up ThunkSection Size when large OS.
In D71281 a fix was put in to round up the size of a ThunkSection to the
nearest 4KiB when performing errata patching. This fixed a problem with a
very large instrumented program that had thunks and patches mutually
trigger each other. Unfortunately it triggers an assertion failure in an
AArch64 allyesconfig build of the kernel. There is a specific assertion
preventing an InputSectionDescription being larger than 4KiB. This will
always trigger if there is at least one Thunk needed in that
InputSectionDescription, which is possible for an allyesconfig build.

Abstractly the problem case is:
.text : {
          *(.text) ;
          ...
          . = ALIGN(SZ_4K);
          __idmap_text_start = .;
          *(.idmap.text)
          __idmap_text_end = .;
          ...
        }
The assertion checks that __idmap_text_end - __idmap_start is < 4 KiB.
Note that there is more than one InputSectionDescription in the
OutputSection so we can't just restrict the fix to OutputSections smaller
than 4 KiB.

The fix presented here limits the D71281 to InputSectionDescriptions that
meet the following conditions:
1.) The OutputSection is bigger than the thunkSectionSpacing so adding
thunks will affect the addresses of following code.
2.) The InputSectionDescription is larger than 4 KiB. This will prevent
any assertion failures that an InputSectionDescription is < 4 KiB
in size.

We do this at ThunkSection creation time as at this point we know that
the addresses are stable and up to date prior to adding the thunks as
assignAddresses() will have been called immediately prior to thunk
generation.

The fix reverts the two tests affected by D71281 to their original state
as they no longer need the 4KiB size roundup. I've added simpler tests to
check for D71281 when the OutputSection size is larger than the ThunkSection
spacing.

Fixes https://github.com/ClangBuiltLinux/linux/issues/812

Differential Revision: https://reviews.llvm.org/D72344
2020-01-17 10:47:21 +00:00
Fangrui Song 2d7a8cf904 [ELF] -r: don't create .interp
`{clang,gcc} -nostdlib -r a.c` passes --dynamic-linker to the linker,
and the expected behavior is to ignore it.

If .interp is kept in the relocatable object file, a final link will get
PT_INTERP even if --dynamic-linker is not specified. glibc ld.so expects
to see PT_DYNAMIC and the executable will likely fail to run.

Ignore --dynamic-linker in -r mode as well as -shared.
2020-01-16 12:14:32 -08:00
Fangrui Song 870094decf [ELF] Decrease alignment of ThunkSection on 64-bit targets from 8 to 4
ThunkSection contains 4-byte instructions on all targets that use
thunks. Thunks should not be used in any performance sensitive places,
and locality/cache line/instruction fetching arguments should not apply.

We use 16 bytes as preferred function alignments for modern PowerPC cores.
In any case, 8 is not optimal.

Differential Revision: https://reviews.llvm.org/D72819
2020-01-16 10:36:33 -08:00
Andrew Ng d36b2649e5 [ELF] Optimization to LinkerScript::computeInputSections NFC
Moved the section name check ahead of any filename matching or
exclusion. Firstly, this reduces the need to retrieve the filename and
secondly, reduces the amount of potentially expensive filename pattern
matching if such rules are present in the linker script.

The impact of this change is particularly significant when linking
objects built with -ffunction-sections and -fstack-size-section, using a
linker script that includes non-trivial filename patterns. In a number
of such cases, the link time is halved.

Differential Revision: https://reviews.llvm.org/D72775
2020-01-16 13:56:02 +00:00
Alex Richardson 441410be47 [ELF] Avoid false-positive assert in getErrPlace()
This assertion was added as part of D70659 but did not account for .bss
input sections. I noticed that this assert was incorrectly triggering
while building FreeBSD for MIPS64. Fixed by relaxing the assert to also
account for SHT_NOBITS input sections and adjust the test
mips-jalr-non-function.s to link a file with a .bss section first.

Reviewed By: MaskRay, grimar
Differential Revision: https://reviews.llvm.org/D72567
2020-01-15 14:32:25 +00:00
Fangrui Song bec1b55c64 [ELF] Delete the RelExpr member R_HINT. NFC
R_HINT is ignored like R_NONE. There are no strong reasons to keep
R_HINT. The largest RelExpr member R_RISCV_PC_INDIRECT is 60 now.

Differential Revision: https://reviews.llvm.org/D71822
2020-01-14 10:56:53 -08:00
Fangrui Song 40c5bd4212 [ELF] --exclude-libs: don't assign VER_NDX_LOCAL to undefined symbols
Suggested by Peter Collingbourne.

Non-VER_NDX_GLOBAL versions should not be assigned to defined symbols. --exclude-libs violates this and can cause a spurious error "cannot refer to absolute symbol" after D71795.

excludeLibs incorrectly assigns VER_NDX_LOCAL to an undefined weak symbol =>
isPreemptible is false =>
R_PLT_PC is optimized to R_PC =>
in isStaticLinkTimeConstant, an error is emitted.

Reviewed By: pcc, grimar

Differential Revision: https://reviews.llvm.org/D72681
2020-01-14 10:12:28 -08:00
Fangrui Song d9819f3662 [ELF] Delete unintended --force-bti 2020-01-13 23:57:00 -08:00
Fangrui Song 7cd429f27d [ELF] Add -z force-ibt and -z shstk for Intel Control-flow Enforcement Technology
This patch is a joint work by Rui Ueyama and me based on D58102 by Xiang Zhang.

It adds Intel CET (Control-flow Enforcement Technology) support to lld.
The implementation follows the draft version of psABI which you can
download from https://github.com/hjl-tools/x86-psABI/wiki/X86-psABI.

CET introduces a new restriction on indirect jump instructions so that
you can limit the places to which you can jump to using indirect jumps.

In order to use the feature, you need to compile source files with
-fcf-protection=full.

* IBT is enabled if all input files are compiled with the flag. To force enabling ibt, pass -z force-ibt.
* SHSTK is enabled if all input files are compiled with the flag, or if -z shstk is specified.

IBT-enabled executables/shared objects have two PLT sections, ".plt" and
".plt.sec".  For the details as to why we have two sections, please read
the comments.

Reviewed By: xiangzhangllvm

Differential Revision: https://reviews.llvm.org/D59780
2020-01-13 23:39:28 -08:00
Fangrui Song 2d077d6dfa [ELF] Make TargetInfo::writeIgotPlt a no-op
RELA targets don't read initial .got.plt entries.
REL targets (ARM, x86-32) write the address of the IFUNC resolver to the
entry (`write32le(buf, s.getVA())`).

The default writeIgotPlt() is not meaningful. Make it a no-op. AArch64
and x86-64 will have 0 as initial .got.plt entries associated with
IFUNC.

Reviewed By: peter.smith

Differential Revision: https://reviews.llvm.org/D72474
2020-01-10 09:59:22 -08:00
Wei Mi 21a4710c67 [ThinLTO] Pass CodeGenOpts like UnrollLoops/VectorizeLoop/VectorizeSLP
down to pass builder in ltobackend.

Currently CodeGenOpts like UnrollLoops/VectorizeLoop/VectorizeSLP in clang
are not passed down to pass builder in ltobackend when new pass manager is
used. This is inconsistent with the behavior when new pass manager is used
and thinlto is not used. Such inconsistency causes slp vectorization pass
not being enabled in ltobackend for O3 + thinlto right now. This patch
fixes that.

Differential Revision: https://reviews.llvm.org/D72386
2020-01-09 21:13:11 -08:00
Fangrui Song 375371cc8b [ELF] Fix includeInDynsym() when an undefined weak is merged with a lazy definition
An undefined weak does not fetch the lazy definition. A lazy weak symbol
should be considered undefined, and thus preemptible if .dynsym exists.

D71795 is not quite an NFC. It errors on an R_X86_64_PLT32 referencing
an undefined weak symbol. isPreemptible is false (incorrect) => R_PLT_PC
is optimized to R_PC => in isStaticLinkTimeConstant, an error is emitted
when an R_PC is applied on an undefined weak (considered absolute).
2020-01-09 16:24:02 -08:00
Alex Richardson 1444e6e2e6 Re-apply "[ELF] Allow getErrPlace() to work before Out::bufferStart is set"
This time with a fix for the UBSAN failure.

Differential Revision: https://reviews.llvm.org/D70659
2020-01-09 20:26:31 +00:00
Sid Manning 0fa8f701cc [ELF][Hexagon] Add support for IE relocations
Differential Revision: https://reviews.llvm.org/D71143
2020-01-09 09:45:24 -06:00
Fangrui Song b841e119d7 [ELF] Delete an unused special rule from isStaticLinkTimeConstant. NFC
Weak undefined symbols are preemptible after D71794.

  if (sym.isPreemptible)
    return false;
  if (!config->isPic)
    return true;
  // isPic means includeInDynsym is true after D71794.

  ...

  // We can delete this if because it can never be true.
  if (sym.isUndefWeak)
    return true;

Differential Revision: https://reviews.llvm.org/D71795
2020-01-08 09:41:59 -08:00
Fangrui Song 96e2376d02 [ELF] Don't special case weak symbols for pie with no shared objects
D59275 added the following clause to Symbol::includeInDynsym()

  if (isUndefWeak() && Config->Pie && SharedFiles.empty())
    return false;

D59549 explored the possibility to generalize it for -no-pie.

GNU ld's rules are architecture dependent and partly controlled by -z
{,no-}dynamic-undefined-weak. Our attempts to mimic its rules are
actually half-baked and don't provide perceivable benefits (it can save
a few more weak undefined symbols in .dynsym in a -static-pie
executable). Let's just delete the rule for simplicity. We will expect
cosmetic inconsistencies with ld.bfd in certain -static-pie scenarios.

This permits a simplification in D71795.

Reviewed By: peter.smith

Differential Revision: https://reviews.llvm.org/D71794
2020-01-08 09:38:49 -08:00
Peter Smith 051c4d5b7b [LLD][ELF][AArch64] Do not use thunk for undefined weak symbol.
In AArch64 a branch to an undefined weak symbol that does not have a PLT
entry should resolve to the next instruction. The thunk generation code
can prevent this from happening as a range extension thunk can be generated
if the branch is sufficiently far away from 0, the value of an undefined
weak symbol.

The fix is taken from the Arm implementation of needsThunk(), we prevent a
thunk from being generated to an undefined weak symbol.

fixes pr44451

Differential Revision: https://reviews.llvm.org/D72267
2020-01-07 09:57:51 +00:00
Kazuaki Ishizaki 7ae3d33546 [lld] Fix trivial typos in comments
Reviewed By: ruiu, MaskRay

Differential Revision: https://reviews.llvm.org/D72196
2020-01-06 10:25:48 -08:00