-N (-omagic)
Set the text and data sections to be readable and writable.
Also, do not page-align the data segment.
Differential revision: https://reviews.llvm.org/D26888
llvm-svn: 288123
Previously Config->SingleRoRx was set in
createFiles() and used HasSections.
This change moves it to readConfigs at place of
common flags handling, and adds logic that sets
this flag separatelly from ScriptParser if SECTIONS present.
llvm-svn: 288021
--no-rosegment: Do not put read-only non-executable sections in their own segment
Differential revision: https://reviews.llvm.org/D26889
llvm-svn: 288020
-color-diagnostics=auto is default because that's the same as
Clang's default. When color is enabled, error or warning messages
are colored like this.
error:
<bold>ld.lld</bold> <red>error:</red> foo.o: no such file
warning:
<bold>ld.lld</bold> <magenta>warning:</magenta> foo.o: no such file
Differential Revision: https://reviews.llvm.org/D27117
llvm-svn: 287949
Uncompressing section contents and spliting mergeable section contents
into smaller chunks are heavy tasks. They scan entire section contents
and do CPU-intensive tasks such as uncompressing zlib-compressed data
or computing a hash value for each section piece.
Luckily, these tasks are independent to each other, so we can do that
in parallel_for_each. The number of input sections is large (as opposed
to the number of output sections), so there's a large parallelism here.
Actually the current design to call uncompress() and splitIntoPieces()
in batch was chosen with doing this in mind. Basically what we need to
do here is to replace `for` with `parallel_for_each`.
It seems this patch improves latency significantly if linked programs
contain debug info (which in turn contain lots of mergeable strings.)
For example, the latency to link Clang (debug build) improved by 20% on
my machine as shown below. Note that ld.gold took 19.2 seconds to do
the same thing.
Before:
30801.782712 task-clock (msec) # 3.652 CPUs utilized ( +- 2.59% )
104,084 context-switches # 0.003 M/sec ( +- 1.02% )
5,063 cpu-migrations # 0.164 K/sec ( +- 13.66% )
2,528,130 page-faults # 0.082 M/sec ( +- 0.47% )
85,317,809,130 cycles # 2.770 GHz ( +- 2.62% )
67,352,463,373 stalled-cycles-frontend # 78.94% frontend cycles idle ( +- 3.06% )
<not supported> stalled-cycles-backend
44,295,945,493 instructions # 0.52 insns per cycle
# 1.52 stalled cycles per insn ( +- 0.44% )
8,572,384,877 branches # 278.308 M/sec ( +- 0.66% )
141,806,726 branch-misses # 1.65% of all branches ( +- 0.13% )
8.433424003 seconds time elapsed ( +- 1.20% )
After:
35523.764575 task-clock (msec) # 5.265 CPUs utilized ( +- 2.67% )
159,107 context-switches # 0.004 M/sec ( +- 0.48% )
8,123 cpu-migrations # 0.229 K/sec ( +- 23.34% )
2,372,483 page-faults # 0.067 M/sec ( +- 0.36% )
98,395,342,152 cycles # 2.770 GHz ( +- 2.62% )
79,294,670,125 stalled-cycles-frontend # 80.59% frontend cycles idle ( +- 3.03% )
<not supported> stalled-cycles-backend
46,274,151,813 instructions # 0.47 insns per cycle
# 1.71 stalled cycles per insn ( +- 0.47% )
8,987,621,670 branches # 253.003 M/sec ( +- 0.60% )
148,900,624 branch-misses # 1.66% of all branches ( +- 0.27% )
6.747548004 seconds time elapsed ( +- 0.40% )
llvm-svn: 287946
This is in the context of https://llvm.org/bugs/show_bug.cgi?id=31109.
When LLD prints out errors for relocations, it tends to print out
extremely large number of errors (like millions) because it would
print out one error per relocation.
This patch makes LLD bail out if it prints out more than 20 errors.
You can configure the limitation using -error-limit argument.
-error-limit=0 means no limit.
I chose the flag name because Clang has the same feature as -ferror-limit.
"f" doesn't make sense to us, so I omitted it.
Differential Revision: https://reviews.llvm.org/D26981
llvm-svn: 287789
For now MipsGotSection class is not ready for concurrent access from
multiple threads. The problem is in the getPageEntryOffset method. It
changes state of MipsGotSection object and might be called from
different threads at the same time. So turn Threads off for this target.
It's a temporary solution. The patch fixes MipsGotSection::getPageEntryOffset
is almost ready.
Differential revision: https://reviews.llvm.org/D27035
llvm-svn: 287740
GNU linkers disagree here.
Though both -version and -v are mentioned
in help to print the version information, GNU ld just normally exits,
while gold can continue linking. We are compatible with ld.bfd here.
This fixes PR31057.
Differential revision: https://reviews.llvm.org/D26865
llvm-svn: 287448
LLD supports multi-threading, and it seems to be working well as
you can see in r287140. In short, LLD runs a few percent to 30%
faster with -threads and more than 50% faster if you are using
-build-id (your mileage may vary depending on your computer).
However, I don't think most users even don't know about that because
-threads is not a default option.
This patch enables it by default.
Discussion thread:
http://lists.llvm.org/pipermail/llvm-dev/2016-November/107160.html
llvm-svn: 287237
Our build-id is a tree hash anyway, so I'll define this as a synonym
for sha1. GNU gold takes this parameter, so this is for compatibility
with that.
llvm-svn: 287119
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
This is forcing to use Error::success(), which is in a wide majority
of cases a lot more readable.
Differential Revision: https://reviews.llvm.org/D26481
llvm-svn: 286561
The previous code didn't make sense at all. Now an error condition
is handled with fatal(). Thanks to Mehdi for pointing out the issue.
llvm-svn: 286547
Patch allows to pass a symbols file to linker.
LLD will map symbols to sections and sort sections
in output according to symbol ordering file.
That can help to reduce the startup time and/or
amount of pagefaults during startup.
Also, interesting benchmark result was produced by Rafael Espíndola.
After applying the symbols file for clang he timed compiling
X86MCTargetDesc.ii to an object file.
The page faults went from just
56,988 to 56,946 since most faults are not in the binary.
Running time went from 4.403053515 to 4.178112244.
The speedup seems to be because of better cache
locality.
Differential revision: https://reviews.llvm.org/D26130
llvm-svn: 286440
In short the patch introduces support for linking object file conform
MIPS N32 ABI [1]. This ABI is similar to N64 ABI but uses 32-bit
pointer size.
The most non-trivial requirement of this ABI is one more relocation
packing format. N64 ABI puts multiple relocation type into the single
relocation record. The N32 ABI uses series of successive relocations
with the same offset for this purpose. In this patch, new function
`mergeMipsN32RelTypes` handle this case and "convert" N32 relocation to
the N64 relocation so the rest of the code keep unchanged.
For now, linker does not support series of relocations applied to sections
without SHF_ALLOC bit. Probably later I will add the support or insert
some sort of assert into the `relocateNonAlloc` routine to catch this
case.
[1] ftp://www.linux-mips.org/pub/linux/mips/doc/ABI/MIPS-N32-ABI-Handbook.pdf
Differential revision: https://reviews.llvm.org/D26298
llvm-svn: 286052
Previously, we do this piece of code to iterate over all input sections.
for (elf::ObjectFile<ELFT> *F : Symtab.getObjectFiles())
for (InputSectionBase<ELFT> *S : F->getSections())
It turned out that this mechanisms doesn't work well with synthetic
input sections because synthetic input sections don't belong to any
input file.
This patch defines a vector that contains all input sections including
synthetic ones.
llvm-svn: 286051
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
As the state of lld gets more complicated, shutting down gets more
expensive.
In a normal lld run we can just call _exit immediately after renaming
the temporary output file. We still want the ability to run a full
shutdown since that is useful for detecting memory leaks.
This patch adds a --full-shutdown flag and changes lit to use it.
llvm-svn: 285224
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
In this patch partial gdb_index section is created.
For costructing the .gdb_index section 6 steps should be performed (details are in
SplitDebugInfo.cpp file header), this patch do first 3:
Creates proper section header.
Fills list of compilation units.
Types CU list area is not supposed to be supported, so it is ignored and therefore
can be treated as implemented either.
Differential revision: https://reviews.llvm.org/D24706
llvm-svn: 284708
-format=<foo>, -format <foo> and -b <foo> are all the same.
Previous code was intended to produce an error message with the
same spelling as given from the command line, but it actually
always printed out this string: "unknown -format= value:".
This is probably more confusing than "unknown -format value:".
So I changed the message.
llvm-svn: 284693
Previously, we were checking the existence of an entry symbol
too early. It was done before the linker script processor creates
symbols defined in scripts. Fixes bug 30743.
llvm-svn: 284676
The R_ARM_TARGET2 relocation is used in ARM exception tables to encode
a data dependency that will only be dereferenced by code in the
run-time support library. In a similar way to R_ARM_TARGET1 the
handling of the relocation is target specific, it maps to one of
R_ARM_ABS32, R_ARM_REL32 or R_ARM_GOT_PREL. The choice depends on the
run-time library. R_ARM_GOT_PREL is used for linux and BSD,
R_ARM_ABS32 and R_ARM_REL32 are used for bare-metal.
The command line option --target2=<target> can be used to select the
relocation used for R_ARM_TARGET2. The default is R_ARM_GOT_PREL.
Differential revision: https://reviews.llvm.org/D25684
llvm-svn: 284404
