StringRefZ is a class to represent a null-terminated string. String
length is computed lazily, so it's more efficient than StringRef to
represent strings in string table.
The motivation of defining this new class is to merge functions
that only differ in string types; we have many constructors that takes
`const char *` or `StringRef`. With StringRefZ, we can merge them.
Differential Revision: https://reviews.llvm.org/D27037
llvm-svn: 288172
We have different functions to stringize objects to construct
error messages. For InputFile, we have getFilename, and for
InputSection, we have getName. You had to memorize them.
I think this is the case where the function overloading comes in handy.
This patch defines toString() functions that are overloaded for all these
types, so that you just call it in error().
Differential Revision: https://reviews.llvm.org/D27030
llvm-svn: 287787
Previously, we stored offsets in string tables to symbols, so
you needed to pass a string table to get a symbol name. This patch
stores const char pointers instead to eliminate the need to pass
a string table.
llvm-svn: 287737
LLD's error messages contain line numbers, function names or section names.
Currently they are formatter as follows.
foo.c (32): symbol 'foo' not found
foo.c (function bar): symbol 'foo' not found
foo.c (.text+0x1234): symbol 'foo' not found
This patch changes them so that they are consistent with Clang's output.
foo.c:32: symbol 'foo' not found
foo.c:(function bar): symbol 'foo' not found
foo.c:(.text+0x1234): symbol 'foo' not found
Differential Revision: https://reviews.llvm.org/D26901
llvm-svn: 287537
Patch adds a filename to that error message.
I faced next error when debugged one of FreeBSD port:
error: relocation R_X86_64_PLT32 cannot refer to absolute symbol __tls_get_addr
error message was poor and this patch improves it to show the locations
of symbol declaration and using.
Differential revision: https://reviews.llvm.org/D26508
llvm-svn: 286940
Found this when tried to link lang/ccl FreeBSD port.
Issue is very close to D23201.
This is the reason of lang/ccl port link fail.
GNU assembler 2.17.50 [FreeBSD] 2007-07-03 could generate broken objects,
where notype symbols are associated with symtab:
...
[ 9] .symtab SYMTAB 0000000000000000 00003c78
0000000000006858 0000000000000018 10 803 8
...
192: 000000000000000d 0 NOTYPE LOCAL DEFAULT 9 _cons_org
Patch allows to handle such objects.
Differential revision: https://reviews.llvm.org/D26613
llvm-svn: 286939
The functions getBitcodeTargetTriple(), isBitcodeContainingObjCCategory(),
getBitcodeProducerString() and hasGlobalValueSummary() now return errors
via their return value rather than via the diagnostic handler.
To make this work, re-implement these functions using non-member functions
so that they can be used without the LLVMContext required by BitcodeReader.
Differential Revision: https://reviews.llvm.org/D26532
llvm-svn: 286623
The change in D26502 splits ReaderWriter.h, which contains the APIs
into both the BitReader and BitWriter libraries, into BitcodeReader.h
and BitcodeWriter.h.
Change lld uses to the appropriate split header, removing it
completely in one case where it wasn't needed.
llvm-svn: 286568
Relocations are the last thing that we wore storing a raw section
pointer to and parsing on demand.
With this patch we parse it only once and store a pointer to the
actual data.
The patch also changes where we store it. It is now in
InputSectionBase. Not all sections have relocations, but most do and
this simplifies the logic. It also means that we now only support one
relocation section per section. Given that that constraint is
maintained even with -r with gold bfd and lld, I think it is OK.
llvm-svn: 286459
Previously, we have both input and output section for .MIPS.abiflags.
Now we have only one class for .MIPS.abiflags, which is MipsAbiFlagsSection.
This class is a synthetic input section.
.MIPS.abiflags sections are handled as regular sections until
the control reaches Writer. Writer then aggregates all sections
whose type is SHT_MIPS_ABIFLAGS to create a single synthesized
input section. The synthesized section is then processed normally
as if it came from an input file.
llvm-svn: 286398
Previously, we have both input and output sections for .reginfo and
.MIPS.options. Now for each such sections we have one synthetic input
sections: MipsReginfoSection and MipsOptionsSection respectively.
Both sections are handled as regular sections until the control reaches
Writer. Writer then aggregates all sections whose type is SHT_MIPS_REGINFO
or SHT_MIPS_OPTIONS to create a single synthesized input section. In that
moment Writer also save GP0 value to the MipsGp0 field of the corresponding
ObjectFile. This value required for R_MIPS_GPREL16 and R_MIPS_GPREL32
relocations calculation.
Differential revision: https://reviews.llvm.org/D26444
llvm-svn: 286397
Instead of remembering a raw Elf_Shdr, store the symbol table proper
and the index of the first non local.
This moves error handling upfront and simplifies it.
llvm-svn: 285933
DIHelper is a class having only one member, and ObjectFile has
a unique pointer to a DIHelper. So we can directly have ObjectFile
have the member.
Differential Revision: https://reviews.llvm.org/D26223
llvm-svn: 285850
When we have SHT_GNU_versym section, it is should be associated with symbol table
section. Usually (and in out implementation) it is .dynsym.
In case when .dynsym is absent (due to broken object for example),
lld crashes in parseVerdefs() when accesses null pointer:
Versym = reinterpret_cast<const Elf_Versym *>(this->ELFObj.base() +
VersymSec->sh_offset) +
this->Symtab->sh_info;
DIfferential revision: https://reviews.llvm.org/D25553
llvm-svn: 285796
Previously, we have a lot of BumpPtrAllocators, but all these
allocators virtually have the same lifetime because they are
not freed until the linker finishes its job. This patch aggregates
them into a single allocator.
Differential revision: https://reviews.llvm.org/D26042
llvm-svn: 285452
Instead of having 3 section allocators per file, have 3 for all files.
This is a substantial performance improvement for some cases. Linking
chromium without gc speeds up by 1.065x.
This requires using _exit in fatal since we have to avoid destructing
an InputSection if fatal is called from the constructor.
Thanks to Rui for the suggestion.
llvm-svn: 285290
We used to have one allocator per file, which reduces the advantage of
using an allocator in the first place.
This is a small speed up is most cases. The largest speedup was in
1.014X in chromium no-gc. The largest slowdown was scylla at 1.003X.
llvm-svn: 285205
This patch make lld show following details for undefined symbol errors:
- file (line)
- file (function name)
- file (section name + offset)
Differential revision: https://reviews.llvm.org/D25826
llvm-svn: 285186
This is not particularly efficient, but does avoid exposing Comdat*
out of LTO.h.
I will send a patch for review with a more efficient interface that
should map nicely to a bitcode symbol table.
llvm-svn: 284413
In continue of D25555, this patch fixes possible crash when
we have multiple SHT_MIPS_REGINFO or SHT_MIPS_ABIFLAGS sections.
yaml2obj was used to produce such objects.
Differential revision: https://reviews.llvm.org/D25609
llvm-svn: 284376
Issue was revealed by AFl and I was able to generate such object using yaml2obj.
When object has more than one SHT_MIPS_OPTIONS,
each except the last one is destroyed after placing into Sections array.
Sections array contains dead pointers finally. LLD may crash then.
Differential revision: https://reviews.llvm.org/D25555
llvm-svn: 284227
With fix: commit changes from InputFiles.cpp too.
Original commit message:
We have following code in lld, that truncates the alignment value to 32 bit. Big alignment in this case
may give result 0 and crash later.
template <class ELFT>
CommonInputSection<ELFT>::CommonInputSection(std::vector<DefinedCommon *> Syms)
: InputSection<ELFT>(nullptr, &Hdr, "") {
....
for (DefinedCommon *Sym : Syms) {
this->Alignment = std::max<uintX_t>(this->Alignment, Sym->Alignment);
...
}
}
Patch fixes the issue.
Differential revision: https://reviews.llvm.org/D25235
llvm-svn: 283738
.ARM.exidx sections have a reverse dependency on the section they have
a SHF_LINK_ORDER dependency on. In other words a .ARM.exidx section is
live only if the executable section it describes is live. We implement
this with a reverse dependency field in InputSection.
Adding the dependency to InputSection is the simplest implementation
but it could be moved out to a separate map if it were found to decrease
performance for non ARM targets.
Differential revision: https://reviews.llvm.org/D25234
llvm-svn: 283734
When sh_info of sumbol table value was set to zero, lld was asserting.
Patch fixes the issue.
Differential revision: https://reviews.llvm.org/D25016
llvm-svn: 283562
Previously if sh_size of dynamic section was broken,
lld may crash. Or even may not crash if used 32 bits host.
(then value may be truncated to 32 bits when doing pointer arithmetic
and could be just zero).
Patch fixes the issue.
Differential revision: https://reviews.llvm.org/D25327
llvm-svn: 283533
createELFObj() may call error(...), for example when file is too short.
In that case header is not set and following line lead to crash:
EMachine = ELFObj.getHeader()->e_machine;
Patch fixes the issue.
Differential revision: https://reviews.llvm.org/D25233
llvm-svn: 283532
This patch makes the check for null section stricter,
so it is only allowed for STT_SECTION symbols now.
Differential revision: https://reviews.llvm.org/D25231
llvm-svn: 283426
Do not merge sections if generating a relocatable object. It makes
the code simpler because we do not need to update relocations addends
to reflect changes introduced by merging. Instead of that we write
such "merge" sections into separate OutputSections and keep SHF_MERGE
/ SHF_STRINGS flags and sh_entsize value to be able to perform merging
later during a final linking.
Differential Revision: http://reviews.llvm.org/D25066
llvm-svn: 283300
Testcase contains a common symbol with zero alignment,
previously lld would crash, patch fixes that.
Differential revision: https://reviews.llvm.org/D25085
llvm-svn: 283197
This was intended to warn on a usage error of an ar command.
I was thinking that false positive would be rare because it
has two guards: it warns only when an archive file has no symbols
and contains at least one file.
Turned out that false positive would probably be not that rare.
You wouldn't intentionally create an object file that contains no
exported symbols, but with conditional compile guards, you could
create such file. If it is for a LTO build, you could create an
archive file containing such file. That means there's no way to
detect the usage error in a reliable manner.
This patch simply removes the warning.
llvm-svn: 282916
Previously, it warned on any archive file that has no symbol.
It turned out that that is too noisy.
With this patch, it warns on such archive file that contains no file.
Differential Revision: https://reviews.llvm.org/D25111
llvm-svn: 282885
This surfaced again with Rust. As per bug 30435, rustc creates a
mergeable section with a sh_entsize zero. It bit us before, too.
I think we should relax the input check rather than being too picky.
Differential Revision: https://reviews.llvm.org/D24789
llvm-svn: 282049
Previously, all input files were owned by the symbol table.
Files were created at various places, such as the Driver, the lazy
symbols, or the bitcode compiler, and the ownership of new files
was transferred to the symbol table using std::unique_ptr.
All input files were then free'd when the symbol table is freed
which is on program exit.
I think we don't have to transfer ownership just to free all
instance at once on exit.
In this patch, all instances are automatically collected to a
vector and freed on exit. In this way, we no longer have to
use std::unique_ptr.
Differential Revision: https://reviews.llvm.org/D24493
llvm-svn: 281425
Implemented by building an ELF file in memory.
elf, default, and binary match gold behavior.
Differential Revision: https://reviews.llvm.org/D24060
llvm-svn: 281108
This simplifies error handling as there is now only one place in the
code that needs to consider the possibility that the name is
corrupted. Before we would do it in every access.
llvm-svn: 280937
This approach is not only consistent with UnresolvedPolicy,
but also should help to solve a problem
of options with opposing meanings, mentioned in PR28843
Differential revision: https://reviews.llvm.org/D23869
llvm-svn: 280206
After latest changes we combine input sections with
different attributes into single output section.
Problem here is that regular output sections does not
support adding mergeable input sections (and vise versa).
Patch just temporarily disables merging for now at
the same way we do for -O0 for example.
This change helps for linking FreeBSD kernel.
Differential revision: https://reviews.llvm.org/D23447
llvm-svn: 278555
We found that GNU assembler 2.17.50 [FreeBSD] 2007-07-03
could generate broken objects. STT_SECTION symbols can be
associated with SHT_REL[A]/SHT_SYMTAB/SHT_STRTAB sections.
This is PR28868, patch fixes handling of such files.
Differential revision: https://reviews.llvm.org/D23201
llvm-svn: 278550
This section supersedes .reginfo and .MIPS.options sections. But for now
we have to support all three sections for ABI transition period.
llvm-svn: 278482
Mergeable sections with size zero are useless because they don't
actually contain data, and therefore there's no merit ot merge them.
However, in reality, there are object files in the wild containing
such sections. Currently, LLD can't handle them proerply.
This patch makes LLD to handle such sections as if they are non-
mergeable to fix the issue.
Fixes bug 28822.
llvm-svn: 277568
We already have code for ARM in initializeSections, so this
is more consistent.
Differential Revision: https://reviews.llvm.org/D22833
llvm-svn: 276811
--trace-symbol is a command line option to watch a symbol.
Previosly, we looked up a hash table for a new symbol if the
option is given. Any code that looks up a hash table for each
symbol is expensive because the linker handles a lot of symbols.
In our design, we look up a hash table strictly only once
for a symbol, so --trace-symbol was an exception.
This patch improves efficiency of the option by merging the
hash table into the symbol table.
Instead of looking up a separate hash table with a string,
this patch sets `Traced` flag to symbols specified by --trace-symbol.
So, if you insert a symbol and get a symbol with `Traced` flag on,
you know that you need to print out a log message for the symbol.
This is nearly zero cost.
llvm-svn: 275716
Previously, each subclass of SymbolBody had a pointer to a source
file from which it was created. So, there was no single way to get
a source file for a symbol. We had getSourceFile<ELFT>(), but the
function was a bit inconvenient as it's a template.
This patch makes SymbolBody have a pointer to a source file.
If a symbol is not created from a file, the pointer has a nullptr.
llvm-svn: 275701
Symbol's dtors are not called because they are allocated using
BumpPtrAllocators. So, members of std::unique_ptr type are not
freed when symbols are deallocated.
This patch is to allocate Thunks using BumpPtrAllocators.
llvm-svn: 274896
So that users are not forced to pass `-m` on the command line
when the inputs are all bitcode.
PR: 28268
Differential Revision: http://reviews.llvm.org/D21779
llvm-svn: 274107
Previously, we initialized Config->EKind and Config->EMachine when
we instantiate ELF objects. That was not an ideal location to do that
because the logic was buried too deep inside a concrete logic.
This patch moves the code to the driver so that the initialization
becomes explicit.
Differential Revision: http://reviews.llvm.org/D21784
llvm-svn: 274089
Patch by Shridhar Joshi.
This option provides names of all the link time modules which define and
reference symbols requested by user. This helps to speed up application
development by detecting references causing undefined symbols.
It also helps in detecting symbols being resolved to wrong (unintended)
definitions in case of applications containing multiple definitions for
same symbols with different types, bindings.
Implements PR28226.
llvm-svn: 273536
This should never happen with correct programs, but it is trivial
write a testcase where lld would crash or report duplicated
symbols. We now behave like when an archive is used and include the
file only once.
llvm-svn: 272724
Add support for an ARM Target and the initial set of relocations
and PLT entries that are necessary for an ARM only hello world to
link. This has been tested against an ARM only sysroot from the
4.2.0 CodeSourcery Lite release.
Tests have been added to test/ELF for the support that has been
implemented.
Main limitations:
- No Thumb support
- Relocations incomplete
- No C++ exceptions support
- No TLS support
- No range extension or interworking veneer (thunk) support
- No Build Attribute support
- No Big-endian support
The deprecated relocations R_ARM_PLT32 and R_ARM_PC24 have been
implemented as these are used by the 4.2.0 CodeSourcery Lite release.
llvm-svn: 271993
MIPS N64 ABI introduces .MIPS.options section which specifies miscellaneous
options to be applied to an object/shared/executable file. LLVM as well as
modern versions of GNU tools read and write the only type of the options -
ODK_REGINFO. It is exact copy of .reginfo section used by O32 ABI.
llvm-svn: 268485
We want --reproduce to
* not rewrite scripts and thin archives
* work with absolute paths
Given that, it pretty much has to create a full directory tree. On windows that
is problematic because of the very short maximum path limit. On most cases
users can still work around it with "--repro c:\r", but that is annoying and
not viable for automated testing.
We then need to produce some form of archive with the files. The first option
that comes to mind is .a files since we already have code for writing them.
There are a few problems with them
The format has a dedicated string table, so we cannot start writing it until
all members are known.
Regular implementations don't support creating directories. We could make
llvm-ar support that, but that is probably not a good idea.
The next natural option would be tar. The problem is that to support long path
names (which is how this started) it needs a "pax extended header" making this
an annoying format to write.
The next option I looked at seems a natural fit: cpio files.
They are available on pretty much every unix, support directories and long path
names and are really easy to write. The only slightly annoying part is a
terminator, but at least gnu cpio only prints a warning if it is missing, which
is handy for crashes. This patch still makes an effort to always create it.
llvm-svn: 268404
This patch implements a new design for the symbol table that stores
SymbolBodies within a memory region of the Symbol object. Symbols are mutated
by constructing SymbolBodies in place over existing SymbolBodies, rather
than by mutating pointers. As mentioned in the initial proposal [1], this
memory layout helps reduce the cache miss rate by improving memory locality.
Performance numbers:
old(s) new(s)
Without debug info:
chrome 7.178 6.432 (-11.5%)
LLVMgold.so 0.505 0.502 (-0.5%)
clang 0.954 0.827 (-15.4%)
llvm-as 0.052 0.045 (-15.5%)
With debug info:
scylla 5.695 5.613 (-1.5%)
clang 14.396 14.143 (-1.8%)
Performance counter results show that the fewer required indirections is
indeed the cause of the improved performance. For example, when linking
chrome, stalled cycles decreases from 14,556,444,002 to 12,959,238,310, and
instructions per cycle increases from 0.78 to 0.83. We are also executing
many fewer instructions (15,516,401,933 down to 15,002,434,310), probably
because we spend less time allocating SymbolBodies.
The new mechanism by which symbols are added to the symbol table is by calling
add* functions on the SymbolTable.
In this patch, I handle local symbols by storing them inside "unparented"
SymbolBodies. This is suboptimal, but if we do want to try to avoid allocating
these SymbolBodies, we can probably do that separately.
I also removed a few members from the SymbolBody class that were only being
used to pass information from the input file to the symbol table.
This patch implements the new design for the ELF linker only. I intend to
prepare a similar patch for the COFF linker.
[1] http://lists.llvm.org/pipermail/llvm-dev/2016-April/098832.html
Differential Revision: http://reviews.llvm.org/D19752
llvm-svn: 268178
This patch redefines the default optimization level as 1 and adds
new level 0. In the command line, it is -O0. The flag disables
costly but optional features so that the linker produces semantically
correct but larger output quickly. Currently it only disables
section merging.
This flag is not intended to be used for final production linking.
It is intended to be used in compile-link-test cycle.
Time to link clang with debug info is about 2x faster with the flag.
Head:
13.24 seconds
Output size: 1227189664 bytes
With this patch:
7.41 seconds
Output size: 2490281784 bytes
Differential Revision: http://reviews.llvm.org/D19705
llvm-svn: 268056
Using multiple context used to be a really big memory saving because we
could free memory from each file while the linker proceeded with the
symbol resolution. We are getting lazier about reading data from the
bitcode, so I was curious if this was still a good tradeoff.
One thing that is a bit annoying is that we still have to copy the
symbol names. The problem is that the names are stored in the Module and
get freed when we move the module bits during linking.
Long term I think the solution is to add a symbol table to the bitcode.
That way IRObject file will not need to use a Module or a Context and we
can drop it while still keeping a StringRef to the names.
This patch is still be an interesting medium term improvement.
When linking llvm-as without debug info this patch is a small speedup:
master: 29.861877513 seconds
patch: 29.814533787 seconds
With debug info the numbers are
master: 34.765181469 seconds
patch: 34.563351584 seconds
The peak memory usage when linking llvm-as with debug info was
master: 599.10MB
patch: 600.13MB
llvm-svn: 267921
There seems to be no reason to keep st_size of undefined symbols.
This patch removes the member for it. This patch will change outputs
in cases that undefined symbols are copied to output, but I think
this is unimportant.
Differential Revision: http://reviews.llvm.org/D19574
llvm-svn: 267826
Rui Ueyama