I found this while trying to build u-boot. It uses -Ttext in
combination with linker scripts.
My first reaction was to change the linker scripts to have the correct
value, but I found that it is actually quite convenient to have -Ttext
take precedence.
By having just
.text : { *(.text) }
In the script, they can define the text address in a single Makefile
and pass it to ld with -Ttext and for the C code with
-DFoo=value. Doing the same with linker scripts would require them to
be generated during the build.
llvm-svn: 305766
Summary:
This is a first step towards getting line info to show up in VS and
windbg. So far, only llvm-pdbutil can parse the PDBs that we produce.
cvdump doesn't like something about our file checksum tables. I'll have
to dig into that next.
This patch adds a new DebugSubsectionRecordBuilder which takes bytes
directly from some other producer, such as a linker, and sticks it into
the PDB. Line tables only need to be relocated. No data needs to be
rewritten.
File checksums and string tables, on the other hand, need to be re-done.
Reviewers: zturner, ruiu
Subscribers: llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D34257
llvm-svn: 305713
This patch adds support for segment NONE in linker scripts which enables the
specification that a section should not be assigned to any segment.
Note that GNU ld does not disallow the definition of a segment named NONE, which
if defined, effectively overrides the behaviour described above. This feature
has been copied.
Differential Revision: https://reviews.llvm.org/D34203
llvm-svn: 305700
In r305364, Rui changed the mechanism that parses -z option values slightly.
This caused a bug, as demonstrated by this test, which now fails:
---
# REQUIRES: x86
# RUN: llvm-mc -filetype=obj -triple=x86_64-pc-linux %s -o %t.o
# RUN: ld.lld %t.o -o %t -z max-page-size
.global _start
_start:
nop
---
Before, the link succeeded and set the max-page-size to the target default.
After we get the following two error messages:
"invalid max-page-size: "
"max-page-size: value isn't a power of 2"
The latter error is because an uninitialised variable ends up being passed back
to getMaxPageSize).
This change ensures we only get the first error.
Reviewers: ruiu
Differential Revision: https://reviews.llvm.org/D34234
llvm-svn: 305679
The get{ARM,AArch64}UndefinedRelativeWeakVA() functions should only be
called for PC-relative relocations. Complete the supported pc-relative
relocations in the switch statement and make the default case unreachable.
The R_ARM_TARGET relocation can be evaluated as R_ARM_REL32 but it is only
used in the context of exception tables, and is never output with respect
to a weak reference so it does not appear in the switch statement.
Differential Revision: https://reviews.llvm.org/D34138
llvm-svn: 305673
In this patch, I flip the switch in DriverUtils from using the external
cvtres.exe tool to using the Windows Resource library in llvm.
I also fixed a bug where .rsrc sections were marked as discardable
memory and therefore were placed in the wrong order in the final PE.
Furthermore, I modified WindowsResource to write the coff directly to a
memory buffer instead of to file, also had it use the machine types
already declared in COFF.h instead creating my own enum.
Finally, I flipped the switch to allow all unit tests that had
previously run only on windows due to a winres dependency to run
cross-platform.
Reviewers: zturner, ruiu
Subscribers: llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D34265
llvm-svn: 305592
Summary:
Adds a "Discarded" bool to SectionChunk to indicate if the section was
discarded by COMDAT deduplication. The Writer still just checks
`isLive()`.
Fixes PR33446
Reviewers: ruiu
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D34288
llvm-svn: 305582
AVR support is somewhat exotic as generated ELF executables are not
directly consumed but objcopy'ed to write it to on-chip flush memory.
This comment describes it for those why a full-fledged ELF linker is
used to link programs for the 8-bit microcontroller.
llvm-svn: 305567
Target.cpp contains code for all the targets that LLD supports. It was
simple and easy, but as the number of supported targets increased,
it got messy.
This patch splits the file into per-target files under ELF/arch directory.
Differential Revision: https://reviews.llvm.org/D34222
llvm-svn: 305565
In preparation for supporting range extension thunks we now continually
call createThunks() until no more thunks are added. This requires us to
record the thunks we add on each pass and only merge the new ones into the
OutputSection. We also need to check if a Relocation is targeting a thunk
to prevent us from infinitely creating more thunks.
Differential Revision: https://reviews.llvm.org/D34034
llvm-svn: 305555
This resubmits commit c0c249e9f2ef83e1d1e5f166b50673d92f3579d7.
It was broken due to some weird template issues, which have
since been fixed.
llvm-svn: 305517
This is probably the main patch left in unifying our intermediary
representation.
It moves the creation of default commands before section sorting. This
has the nice effect that we now have one location where we decide
where an orphan section should be placed.
Before this patch sortSections would decide the relative location of
orphan sections to other sections, but it was up to placeOrphanSection
to decide on the exact location.
We now only sort sections we created since the linker script is
already in the correct order.
llvm-svn: 305512
This reverts commit 83ea17ebf2106859a51fbc2a86031b44d33696ad.
This is failing due to some strange template problems, so reverting
until it can be straightened out.
llvm-svn: 305505
When link is invoked with `/def:` and no input files, it behaves as if
`lib.exe` was invoked. Emulate this behaviour, generating the import
library from the def file that was passed. Because there is no input to
actually generate the dll, we simply process the def file early and exit
once we have created the import library.
llvm-svn: 305502
After some internal discussions, we agreed that the raw output style had
outlived its usefulness. It was originally created before we had even
thought of dumping to YAML, and it was intended to give us some insight
into the internals of a PDB file. Now we have YAML mode which does
almost exactly this but is more powerful in that it can round-trip back
to a PDB, which the raw mode could not do. So the raw mode had become
purely a maintenance burden.
One option was to just delete it. However, its original goal was to be
as readable as possible while staying close to the "metal" - i.e.
presenting the output in a way that maps directly to the underlying file
format. We don't actually need that last requirement anymore since it's
covered by the yaml mode, so we could repurpose "raw" mode to actually
just be as readable as possible.
This patch implements about 80% of the functionality previously in raw
mode, but in a completely different style that is more akin to what
cvdump outputs. Records are very compressed, often times appearing on
just one line. One nice thing about this is that it makes full record
matching easier, because you can grep for indices, names, and leaf types
on a single line often.
See the tests for some examples of what the new output looks like.
Note that this patch actually regresses the functionality of raw mode in
a few areas, but only because the patch was already unreasonably large
and going 100% would have been even worse. Specifically, this patch is
missing:
The ability to dump module debug subsections (checksums, lines, etc)
The ability to dump section headers
Aside from that everything is here. While goign through the tests fixing
them all up, I found many duplicate tests. They've been deleted. In
subsequent patches I will go through and re-add the missing
functionality.
Differential Revision: https://reviews.llvm.org/D34191
llvm-svn: 305495
This is necessary to ensure that sections containing symbols referenced
from linker scripts (e.g. in data commands) don't get GC'd.
Differential Revision: https://reviews.llvm.org/D34195
llvm-svn: 305452
This was originally reverted because of some non-deterministic
failures on certain buildbots. Luckily ASAN eventually caught
this as a stack-use-after-scope, so the fix is included in
this patch.
llvm-svn: 305393
Summary: I found that getString defined in the LLD's Driver.cpp is
exactly the same as Arg::getLastArgValue defined in
llvm/Option/ArgLIst.h. This patch removes that local function and use
the function that the Arg class provides.
Reviewers: ruiu
Subscribers: emaste, llvm-commits
Differential Revision: https://reviews.llvm.org/D34196
llvm-svn: 305374
This is causing failures on linux bots with an invalid stream
read. It doesn't repro in any configuration on Windows, so
reverting until I have a chance to investigate on Linux.
llvm-svn: 305371
This allows us to use yaml2obj and obj2yaml to round-trip CodeView
symbol and type information without having to manually specify the bytes
of the section. This makes for much easier to maintain tests. See the
tests under lld/COFF in this patch for example. Before they just said
SectionData: <blob> whereas now we can use meaningful record
descriptions. Note that it still supports the SectionData yaml field,
which could be useful for initializing a section to invalid bytes for
testing, for example.
Differential Revision: https://reviews.llvm.org/D34127
llvm-svn: 305366
Currently we do layout as if non alloc sections had an actual address
and then set it to zero. This produces a few odd results where a
symbol has an address that is inconsistent with the section address.
The simplest way to fix it is probably to just set the address earlier.
The behavior of bfd seems to be similar, but it only sets the non
alloc section address is missing from the linker script or if the
script has an explicit " : 0" setting the address of the output
section (which the default script does).
llvm-svn: 305323
This means that 'llvm-pdbutil' in test commands will resolve
to the absolute path to the tool, in line with what happens
already for other tools.
This works either way because the bin directory is also
prepended to the PATH. I'm not sure why both methods are
used.
Differential Revision: https://reviews.llvm.org/D34128
llvm-svn: 305297
Summary:
Expose the module descriptor index and fill it in for section
contributions.
Reviewers: zturner
Subscribers: llvm-commits, ruiu, hiraditya
Differential Revision: https://reviews.llvm.org/D34126
llvm-svn: 305296
This shows an oddity of this output. While the section address is 0,
the the symbol address is computed as if the section was allocatable.
llvm-svn: 305250
The last fix required the user to manually add the required
feature. This caused an LLD test to fail because I failed to
update LLD. In practice we can hide this logic so it can just
be transparently added when we write the PDB.
llvm-svn: 305236
The ELF standard defines that the SHT_GROUP section as follows:
- its sh_link has the symbol index, and
- the symbol name is used to uniquify section groups.
Object files created by GNU gold does not seem to comply with the
standard. They have this additional rule:
- if the symbol has no name and a STT_SECTION symbol, a section
name is used instead of a symbol name.
If we don't do anything for this, the linker fails with a mysterious
error message if input files are generated by gas. It is unfortunate
but I think we need to support it.
Differential Revision: https://reviews.llvm.org/D34064
llvm-svn: 305218
Given
.weak target
.global _start
_start:
b target
The intention is that the branch goes to the instruction after the
branch, effectively turning it on a nop. The branch adds the runtime
PC, but we were adding it statically too.
I noticed the oddity by inspection, but llvm-objdump seems to agree,
since it now prints things like:
b #-4 <_start+0x4>
llvm-svn: 305212
Relocations referring to merge sections are considered equal if they
resolve to the same offset in the same output section.
Differential Revision: https://reviews.llvm.org/D34094
llvm-svn: 305177
Currently the freebsd early boot code fails to link. The issue reduces
to 16 bit code at position 0x7000 wanting to jump to position
0x9000. That is represented in the .o file as a relocation with no
symbol and an addend of 0x9000 - 2 (The -2 is because i386 uses the ip
after the current instruction for jumps).
If the addend is interpreted as signed (it should), it is -28674. In a
32 bit architecture, that is the address 0xffff8ffe. To get there from
0x7000 we have to add 4294909950 (too big) or subtract 57346 (too
small). We then produce an error.
What lld is missing is the fact that at runtime this will actually be
a 16 bit architecture and adding 0x1ffe produces 0x8ffe which is the
correct result in 16 bits (-28674).
Since we have a 16 bit addend and a 16 bit PC, the relocation can move
the PC to any 16 bit address and that is the only thing we really need
to check: if the address we are pointing to fits in 16 bits. This is
unfortunately hard to do since we have to delay subtracting the PC and
if we want to do that outside of Target.cpp, we have to move the
overflow check out too. An incomplete patch that tries to do that is
at https://reviews.llvm.org/D34070
This patch instead just relaxes the check. Since the value we have is
the destination minus the PC and the PC is 16 bits, it should fit in
17 bits if the destination fits in 16 too.
bfd had a similar issue for some time and got a similar fix:
https://sourceware.org/ml/binutils/2005-08/msg00001.html
llvm-svn: 305135
SHF_GROUP bit doesn't make sense in executables or DSOs, so linkers are
expected to remove that bit from section flags. We did that when we create
output sections.
This patch is to do that earlier than before. Now the flag is dropped when
we instantiate input section objects.
This change improves ICF. Previously, two sections that differ only in
SHF_GROUP flag were not merged, because when the control reached ICF,
the flag was still there. Now the flag is dropped before reaching to ICF,
so the difference is ignored naturally.
This issue was found by pcc.
Differential Revision: https://reviews.llvm.org/D34074
llvm-svn: 305134
This is to reflect the evolving nature of the tool as being
useful for more than just dumping PDBs, as it can do many other
things.
Differential Revision: https://reviews.llvm.org/D34062
llvm-svn: 305106
This is used by linux kernel build system.
(https://www.kernel.org/doc/Documentation/kbuild/makefiles.txt "3.2 Built-in object goals")
It has for example next configuration for linking built-in.o files:
drivers-y := $(patsubst %/, %/built-in.o, $(drivers-y))
drivers-$(CONFIG_PCI) += arch/ia64/pci/
...
drivers-$(CONFIG_OPROFILE) += arch/ia64/oprofile/
Im most simple case all CONFIG_* options are off. That means linker is called with empty input archive,
emulation option and no inputs and expected to generate some relocatable output.
ld.bfd is able to do that, we dont.
Patch allows to support this case.
Differential revision: https://reviews.llvm.org/D33937
llvm-svn: 305069
The symbols generated for Thunks have type STT_FUNC, to permit a thunk to
be reused via a blx instruction the Thumb bit (0) needs to be set properly.
Differential Revision: https://reviews.llvm.org/D34036
llvm-svn: 305065
Previously, it couldn't parse
SECTIONS .text (0x1000) : { *(.text) }
because "(" was interpreted as the begining of the "(NOLOAD)" directive.
llvm-svn: 305006
Without this, when building with shared BUILD_SHARED_LIBS=ON
I get errors such as:
lib/Core/Reader.cpp:40: error: undefined reference to
'llvm::identify_magic(llvm::StringRef)'
Differential Revision: https://reviews.llvm.org/D34004
llvm-svn: 304932
This patch reimplements .gdb_index in more natural way,
that makes proccess of switching to multithreaded index building to
be trivial.
Differential revision: https://reviews.llvm.org/D33552
llvm-svn: 304927
Approach significantly simplifies LLD .gdb_index code and makes it much faster.
Also it should resolve issues, like D33176 tries to address once and forever in a clean way.
LLC binary linking without patch and without --gdb-index: 1,599241063
LLC binary linking without patch and with --gdb-index: 6,064316262
LLC binary linking with patch and with --gdb-index: 4,116792104
Time spent for building gdbindex changes from (6,064316262 - 1,599241063 == 4,465075199)
to (4,116792104- 1,599241063 == 2,517551041).
That is 2,517551041/4,465075199 = 0,564 or about 44% speedup.
Differential revision: https://reviews.llvm.org/D33183
llvm-svn: 304895
Thunks are now generated per InputSectionDescription instead of per
OutputSection. This allows created ThunkSections to be inserted directly
into InputSectionDescription.
Changes in this patch:
- Loop over InputSectionDescriptions to find relocations to Thunks
- Generate a ThunkSection per InputSectionDescription
- Remove synchronize() as we no longer need it
- Move fabricateDefaultCommands() before createThunks
Differential Revision: https://reviews.llvm.org/D33835
llvm-svn: 304887
Previously we would merge relocation sections by name.
That did not work in some cases, like testcase shows.
Patch implements logic to merge relocation sections if their target
sections were merged into the same output section.
Differential revision: https://reviews.llvm.org/D33824
llvm-svn: 304886
When linking linux kernel LLD currently reports next errors:
ld: error: unable to evaluate expression: input section .head.text has no output section assigned
ld: error: At least one side of the expression must be absolute
ld: error: At least one side of the expression must be absolute
That does not provide file/line information and overall looks unclear.
Patch adds location information to ExprValue and that allows
to provide more clear error messages.
Differential revision: https://reviews.llvm.org/D33943
llvm-svn: 304881
This creates a new library called BinaryFormat that has all of
the headers from llvm/Support containing structure and layout
definitions for various types of binary formats like dwarf, coff,
elf, etc as well as the code for identifying a file from its
magic.
Differential Revision: https://reviews.llvm.org/D33843
llvm-svn: 304864
Before this patch in -r we compute the OutputSection sizes early in
the various calls to assignOffsets. With this change we can remove
most of those calls.
llvm-svn: 304860
In preparation for inserting Thunks into InputSectionDescription::Sections
extract the loop that finds InputSections that may have calls that need
Thunks. This isn't much benefit now but this will be useful when we have to
extract the InputSectionDescriptions::Sections from the script.
Differential Revision: https://reviews.llvm.org/D33834
llvm-svn: 304783
In preparation for inserting Thunks into InputSectionDescriptions this
simple change associates added Thunks with a vector of InputSections instead
of an OutputSection. As of now we are just using OutputSection::Sections.
Differential Revision: https://reviews.llvm.org/D33832
llvm-svn: 304782
Previously LLD would fail for case when there are multiple comdats and -r.
That happened because it merged all ".group" (SHT_GROUP) sections into single
output, producing broken result. Such sections may have similar name, alignment and flags
and other properties. We need to produce separate output section for each such input one.
Differential revision: https://reviews.llvm.org/D33643
llvm-svn: 304769
Traditionally, it has been defined in crtbegin.o, which is typically
provided by libgcc or as part of the C library on some systems. However,
but there's no principled reason for it to be there. We optionaly
define this symbol, which can be used on platforms that don't provide
__dso_handle in crtbegin.o or which don't use crtbegin.o at all.
Differential Revision: https://reviews.llvm.org/D33856
llvm-svn: 304732
procedural optimizations to prevent dropping symbols and allow the linker
to process re-directs.
PR33145: --wrap doesn't work with lto.
Differential Revision: https://reviews.llvm.org/D33621
llvm-svn: 304719
This is PR33289.
Previously LLD leaved section naming as is and that lead to wrong result,
because we decompress sections when using -r,
and hence should remove ".z" prefix.
Differential revision: https://reviews.llvm.org/D33885
llvm-svn: 304711
This change alters the sorting for OutputSections with the SHF_LINK_ORDER
flag in OutputSection::finalize() to use the InputSectionDescription
representation and not the OutputSection::Sections representation.
Differential revision: https://reviews.llvm.org/D33772
llvm-svn: 304700
This reverts commit r304561 and re-lands r303490 & co.
The fix was to use "SymbolName" when translating LLD's internal export
list to lib/Object's short export struct. The SymbolName reflects the
actual symbol name, which may include fastcall and stdcall mangling bits
not included in the /EXPORT or .def file EXPORTS name:
@@ -434,8 +434,7 @@ std::vector<COFFShortExport> createCOFFShortExportFromConfig() {
std::vector<COFFShortExport> Exports;
for (Export &E1 : Config->Exports) {
COFFShortExport E2;
- E2.Name = E1.Name;
+ // Use SymbolName, which will have any stdcall or fastcall qualifiers.
+ E2.Name = E1.SymbolName;
E2.ExtName = E1.ExtName;
E2.Ordinal = E1.Ordinal;
E2.Noname = E1.Noname;
llvm-svn: 304573
The .def file parser changes I reverted broke this test case, and
exported "__imp__foo" instead of "__imp__foo@8". This was
http://crbug.com/728726.
llvm-svn: 304572
This reverts commits r303490, r303491, r303493, and r303494.
This caused http://crbug.com/728726. Essentially, exporting stdcall
functions doesn't appear to work after this change. Reduced test case
soon.
llvm-svn: 304561
Spec says: (http://www.sco.com/developers/gabi/latest/ch4.sheader.html)
sh_info
This member holds extra information, whose interpretation depends on the section type.
If the sh_flags field for this section header includes the attribute SHF_INFO_LINK,
then this member represents a section header table index.
SHF_INFO_LINK
The sh_info field of this section header holds a section header table index.
Since sh_info for SHT_REL[A] sections should contain the section header index of the
section to which the relocation applies, this is
consistent with spec to put this flag. Behavior matches both bfd and gold as well.
Differential revision: https://reviews.llvm.org/D33763
llvm-svn: 304531
Object files have symbol records not aligned to any particular
boundary (e.g. 1-byte aligned), while PDB files have symbol
records padded to 4-byte aligned boundaries. Since they share
the same reading / writing code, we have to provide an option to
specify the alignment and propagate it up to the producer or
consumer who knows what the alignment is supposed to be for the
given container type.
Added a test for this by modifying the existing PDB -> YAML -> PDB
round-tripping code to round trip symbol records as well as types.
Differential Revision: https://reviews.llvm.org/D33785
llvm-svn: 304484
This is probably the correct location for it: next to
fabricateDefaultCommands. If we don't have a linker script, we
fabricate one. If we have one, we patch it.
llvm-svn: 304436
This is PR33243. R_GOTONLY_PC_FROM_END was not in a list of link time constant
expressions and that was a result of confusiing messages like PR shows:
/usr/bin/ld.lld: error: /usr/lib/go/src/runtime/alg.go:47:
can't create dynamic relocation R_386_GOTPC against local symbol in readonly segment defined in /tmp/nice/go-link-597453838/go.o
Though in reality we just should not have try to create a dynamic relocation for this case at all.
Patch fixes the issue.
Differential revision: https://reviews.llvm.org/D33717
llvm-svn: 304393
We were looking up sections by name during expression evaluation. By
keeping track of forward declarations we can do the lookup during
script parsing.
Doing the lookup earlier will be more efficient when assignAddresses
is run twice and removes two uses of OutputSections.
llvm-svn: 304381
Before InputSectionBase had an OutputSection pointer, but that was not
always valid. For example, if it was a merge section one actually had
to look at MergeSec->OutSec.
This was brittle and caused bugs like the one fixed by r304260.
We now have a single Parent pointer that points to an OutputSection
for InputSection, but to a SyntheticSection for merge sections and
.eh_frame. This makes it impossible to accidentally access an invalid
OutSec.
llvm-svn: 304338
We would crash if a SHF_LINK_ORDER section pointed to a non
InputSection section. Since those sections are not merged in order,
SHF_LINK_ORDER is pretty meaningless and we can error on that case.
llvm-svn: 304327
This change converts the writing of the .ARM.exidx sentinel section to use
the InputSectionDescriptions instead of OutputSection::Sections this is in
preparation for the retirement of OutputSection::Sections.
Differential Revision: https://reviews.llvm.org/D33500
llvm-svn: 304289
This happens when attempting to link shared libraries using exceptions on
MIPS. It requires -z notext because clang generates R_MIPS_64 relocations
inside .eh_frame.
The crash happened because for EhInputSection the OutSec member is null.
Patch by Alexander Richardson!
llvm-svn: 304260
When there is a linker script with .ARM.exidx in the SECTIONS
command we must add the .ARM.exidx sentinel section to the
InputSectionDescriptions as well as to OutputSection::Sections.
Differential Revision: https://reviews.llvm.org/D33496
llvm-svn: 304206
I found that during visual inspection of code while wrote different patch.
Script in testcase probably have nothing common with real life, but
we segfault currently using it.
If output section is known NOBITS, there is no need to create
writers threads for doing nothing or proccess any filler logic that
is useless here. We can just early return, that is what this patch do.
DIfferential revision: https://reviews.llvm.org/D33646
llvm-svn: 304192
InputSections may contain MergeInputSection members which trigger
a segmentation fault when trying to cast them to InputSection.
Differential Revision: https://reviews.llvm.org/D33628
llvm-svn: 304189
While the following expression is handled fine:
PROVIDE_HIDDEN(newsym = oldsym + address);
The following expression triggers an error because the expression
is evaluated as absolute:
PROVIDE_HIDDEN(newsym = ALIGN(oldsym, CONSTANT(MAXPAGESIZE)) + address);
To avoid this error, we use late evaluation for ALIGN by making the
alignment an attribute of the expression itself.
Differential Revision: https://reviews.llvm.org/D33629
llvm-svn: 304185
Now that we are trying to use the linker script representation as the
canonycal one, there are a few loops looking for just OutputSectionCommands.
Create a vector with just the OutputSectionCommands once that is
stable to simplify the rest of the code.
llvm-svn: 304181
This is PR33052, "Bug 33052 - -r eats comdats ".
To fix it I stop removing group section from out when -r is given
and fixing SHT_GROUP content when writing it just like we do some
other fixup, e.g. for Rel[a]. (it needs fix for section indices that
are in group).
Differential revision: https://reviews.llvm.org/D33485
llvm-svn: 304140
The .dynamic section of an ELF almost doesn't need to be written to with
the exception of the DT_DEBUG entry. For several reasons having a read
only .dynamic section would be useful. This change adds the -z keyword
"rodynamic" which forces .dynamic to be read-only. In this case DT_DEBUG
will not be emited.
Patch by Jake Ehrlich
Differential Revision: https://reviews.llvm.org/D33251
llvm-svn: 304024
On SPARC, .plt is both writeable and executable. The current way
sections are sorted means that lld puts it after .data/.bss. but it
really needs to be close to .test to make sure branches into .plt
don't overflow. I'd argue that because .bss is supposed to come last
on all architectures, we should change the default sort order such
that writable and executable sections come before sections that are
just writeable. read-only executable sections should still come after
sections that are just read-only of course. This diff makes this
change.
llvm-svn: 304008
I found this when builded llc binary using gcc 5.4.1 + LLD.
gcc produces duplicate entries in .debug_gnu_pubtypes section, ex:
UnifyFunctionExitNodes.cpp.o has:
0x0000ac07 EXTERNAL TYPE "std::success_type<void*>"
0x0000ac07 EXTERNAL TYPE "std::success_type<void*>"
clang produces single entry here:
0x0000d291 EXTERNAL TYPE "std::__success_type<void *>"
If we link output from gcc with LLD, that would produce excessive duplicate
entries in .gdb_index constant pool area. That does not seem affect gdb work,
but makes .gdb_index larger than it can be.
I also checked that gold filters out such duplicates too. Patch fixes it.
Differential revision: https://reviews.llvm.org/D32647
llvm-svn: 303975
https://sourceware.org/gdb/onlinedocs/gdb/Index-Section-Format.html says:
"A CU vector in the constant pool is a sequence of offset_type values.
The first value is the number of CU indices in the vector.
Each subsequent value is the index and symbol attributes of a CU in the CU list."
Previously we keeped 2 values until the end, what was useless.
Initially was a part of D32647, though it is possible to split out.
Patch do that.
Differential revision: https://reviews.llvm.org/D33551
llvm-svn: 303973
In this way, the content and the flag is always consistent, which I
think better than removing the bit when input sections reaches the Writer.
llvm-svn: 303926
Originally this was intended to be set up so that when linking
a PDB which refers to a type server, it would only visit the
PDB once, and on subsequent visitations it would just skip it
since all the records had already been added.
Due to some C++ scoping issues, this was not occurring and it
was revisiting the type server every time, which caused every
record to end up being thrown away on all subsequent visitations.
This doesn't affect the performance of linking clang-cl generated
object files because we don't use type servers, but when linking
object files and libraries generated with /Zi via MSVC, this means
only 1 object file has to be linked instead of N object files, so
the speedup is quite large.
llvm-svn: 303920
If you pass /delayload:<dllname> to the COFF linker, it creates thunks
so that DLLs are loaded when they are used for the first time instead of
load-time.
This mechanism do not work for data symbols as there's no way to trap
acccesses to data imported from DLLs. (Technically, I think if we do not
initially map dllimport tables in memory, we could actually trap accesses
and delay-load data symbols, but that's not what Windows do.)
This patch is to report an error when you try to delay-load data symbols.
Fixes https://bugs.llvm.org/show_bug.cgi?id=33106
Differential Revision: https://reviews.llvm.org/D33557
llvm-svn: 303890
This is a different implementation than r303225 (which was reverted
in r303270, re-submitted in r303304 and then re-reverted in r303527).
In the previous patch, I tried to add Live bit to each dllimported
symbol. It turned out that it didn't work with "oldnames.lib" which
contains a lot of weak aliases to dllimported symbols.
The way we handle weak aliases is to check if undefined symbols
can be resolved using weak aliases, and if so, memcpy the Defined
symbols to weak Undefined symbols, so that any references to weak
aliases automatically see defined symbols instead of undefined ones.
This memcpy happens before MarkLive kicks in.
That means we may have multiple copies of dllimported symbols. So
turning on one instance's Live bit is not enough.
This patch moves the Live bit to dllimport file. Since multiple
copies of dllsymbols still point to the same file, we can use it as the
central repository to keep track of liveness.
Differential Revision: https://reviews.llvm.org/D33520
llvm-svn: 303814
It is not clear why a synthetic section wants to use padding defined
in the linker script. The padding is for the space between sections.
It was also missing a test.
llvm-svn: 303812
This is the only place we use threads for ICF. The intention of this code
was to split an input vector into 256 shards and process them in parallel.
What the code was actually doing was to split an input into 257 shards,
process the first 256 shards in parallel, and the remaining one in serial.
That means this code takes ceil(256/n)+1 instead of ceil(256/n) where n
is the number of available CPU cores. The former converges to 2 while
the latter converges to 1.
This patches fixes the above issue.
llvm-svn: 303797
Summary:
This is required on some platforms, as GNU libstdc++ std::call_once is known to be buggy.
This fixes operation of LLD on at least NetBSD and perhaps OpenBSD and Linux PowerPC.
The same change has been introduced to LLVM and LLDB.
Reviewers: ruiu
Reviewed By: ruiu
Subscribers: emaste, #lld
Tags: #lld
Differential Revision: https://reviews.llvm.org/D33508
llvm-svn: 303788
A variable `ComdatGroup` is not supposed to contain a large number of
items. Even when linking clang, it ends up having only 300K strings.
It doesn't make sense to use CachedHashStringRef for this hash table.
This patch has neutral or slightly positive impact on performance while
reducing code complexity.
llvm-svn: 303787
This reduces how many times we have to map from OutputSection to
OutputSectionCommand. It is a required step to moving
clearOutputSections earlier.
In order to always use writeTo in OutputSectionCommand we have to call
fabricateDefaultCommands for -r links and move section compression
after it.
llvm-svn: 303784
LazyRandomTypeCollection is designed for random access, and in
order to provide this it lazily indexes ranges of types. In the
case of types from an object file, there is no partial index
to build off of, so it has to index the full stream up front.
However, merging types only requires sequential access, and when
that is needed, this extra work is simply wasted. Changing the
algorithm to work on sequential arrays of types rather than
random access type collections eliminates this up front scan.
llvm-svn: 303707
Once the dummy linker script is created, we want it to be used for
everything to avoid having two redundant representations that can get
out of sync.
We were already clearing OutputSections. With this patch we clear the
Sections vector of every OutputSection.
llvm-svn: 303703
By the time we get to linker scripts, all special InputSectionBase
should have been combined into synthetic sections, which are a type of
InputSection. The net result is that we can use InputSection in a few
places that were using InputSectionBase.
llvm-svn: 303702
If the compiler driver passes --build-id and the user uses -Wl to
pass --build-id= then the user's flag should take precedence.
Differential Revision: https://reviews.llvm.org/D33461
llvm-svn: 303689
Previous algotirhm assumed that types and ids are in a single
unified stream. For inputs that come from object files, this
is the case. But if the input is already a PDB, or is the result
of a previous merge, then the types and ids will already have
been split up, in which case we need an algorithm that can
accept operate on independent streams of types and ids that
refer across stream boundaries to each other.
Differential Revision: https://reviews.llvm.org/D33417
llvm-svn: 303577
This reverts commit r303304 because it looks like the change
introduced a crash bug. At least after that change, LLD with thinlto
crashes when linking Chromium.
llvm-svn: 303527
This is split up into two commits.
This commit removes the DEF parser from LLD
See the previous commit for the creation in LLVM.
Reviewers: ruiu
Differential Revision: https://reviews.llvm.org/D32689
llvm-svn: 303491
This was originally reverted because it was a breaking a bunch
of bots and the breakage was not surfacing on Windows. After much
head-scratching this was ultimately traced back to a bug in the
lit test runner related to its pipe handling. Now that the bug
in lit is fixed, Windows correctly reports these test failures,
and as such I have finally (hopefully) fixed all of them in this
patch.
llvm-svn: 303446
This is a squash of ~5 reverts of, well, pretty much everything
I did today. Something is seriously broken with lit on Windows
right now, and as a result assertions that fire in tests are
triggering failures. I've been breaking non-Windows bots all
day which has seriously confused me because all my tests have
been passing, and after running lit with -a to view the output
even on successful runs, I find out that the tool is crashing
and yet lit is still reporting it as a success!
At this point I don't even know where to start, so rather than
leave the tree broken for who knows how long, I will get this
back to green, and then once lit is fixed on Windows, hopefully
hopefully fix the remaining set of problems for real.
llvm-svn: 303409
Right now we have multiple notions of things that represent collections of
types. Most commonly used are TypeDatabase, which is supposed to keep
mappings from TypeIndex to type name when reading a type stream, which
happens when reading PDBs. And also TypeTableBuilder, which is used to
build up a collection of types dynamically which we will later serialize
(i.e. when writing PDBs).
But often you just want to do some operation on a collection of types, and
you may want to do the same operation on any kind of collection. For
example, you might want to merge two TypeTableBuilders or you might want
to merge two type streams that you loaded from various files.
This dichotomy between reading and writing is responsible for a lot of the
existing code duplication and overlapping responsibilities in the existing
CodeView library classes. For example, after building up a
TypeTableBuilder with a bunch of type records, if we want to dump it we
have to re-invent a bunch of extra glue because our dumper takes a
TypeDatabase or a CVTypeArray, which are both incompatible with
TypeTableBuilder.
This patch introduces an abstract base class called TypeCollection which
is shared between the various type collection like things. Wherever we
previously stored a TypeDatabase& in some common class, we now store a
TypeCollection&.
The advantage of this is that all the details of how the collection are
implemented, such as lazy deserialization of partial type streams, is
completely transparent and you can just treat any collection of types the
same regardless of where it came from.
Differential Revision: https://reviews.llvm.org/D33293
llvm-svn: 303388
This converts the last (chronologically) user of OutputSections to use
the linker script commands instead.
The idea is to convert all uses after fabricateDefaultCommands, so
that we have a single representation.
llvm-svn: 303384
Our output is not compatible with the Binding feature, so make it
explicit that.
Differential Revision: https://reviews.llvm.org/D33336
llvm-svn: 303378
Previously, LLD-produced executables had IAT (Import Address Table) and
ILT (Import Lookup Table) as separate chunks of data, although their
contents are identical. My interpretation of the COFF spec when I wrote
the COFF linker is that they need to be separate tables even though they
are the same.
But Peter found that the Windows loader is fine with executables in
which IAT and ILT are merged. This is a patch to merge IAT and ILT.
I confirmed that an lld-link self-hosted with this patch works fine.
Fixes https://bugs.llvm.org/show_bug.cgi?id=33064
Differential Revision: https://reviews.llvm.org/D33326
llvm-svn: 303374
The import lists are already binned by DLL name, so there's no need to
deduplicate here.
Differential Revision: https://reviews.llvm.org/D33330
llvm-svn: 303371
We've been using make<> to allocate new objects in ELF. We have
the same function in COFF, but we didn't use it widely due to
negligence. This patch uses the function in COFF to close the gap
between ELF and COFF.
llvm-svn: 303357
GetSection is a template because write calls relocate.
relocate has two parts. The non alloc code really has to be a
template, as it is looking a raw input file data.
The alloc part is only a template because of getSize.
This patch folds the value of getSize early, detemplates
getRelocTargetVA and splits relocate into a templated non alloc case
and a regular function for the alloc case. This has the nice advantage
of making sure we collect all the information we need for relocations
before getting to InputSection::relocateNonAlloc.
Since we know got is alloc, it can just call the function directly and
avoid the template.
llvm-svn: 303355
Rafael Espindola