Commit Graph

38 Commits

Author SHA1 Message Date
Edd Dawson 0a6860521e [LLD][ELF][test] Fix CHECKs in map-file test
A repeated typo in lld/test/ELF/map-file.s prevented a number of checks from being executed.

CHECk-NEXT -> CHECK-NEXT
    ^             ^

After correcting the typo, a small adjustment was needed to match the size of the synthetic .comment section (which always contains "LLD 1.0" in the test environment).

Differential revision: https://reviews.llvm.org/D88023
2020-09-21 18:38:19 +03:00
Sam Clegg cc2da5554b [lld][WebAssembly] Add initial support for -Map/--print-map
Differential Revision: https://reviews.llvm.org/D77187
2020-09-12 16:10:51 -07:00
Fangrui Song cece7af586 [ELF][test] Add --match-full-lines to map-file.s to check leading and trailing spaces
Since we are going to touch so many lines, do some other cleanups.
Delete Inputs/map-file3.s and Inputs/map-file4.s which are tiny.
2020-04-15 11:49:52 -07:00
Fangrui Song 81cebfd008 [ELF][test] Change -o %t to -o /dev/null if the output is not needed 2020-02-12 21:54:50 -08:00
Fangrui Song d4306e90cb [ELF][X86] Allow PT_LOAD to have overlapping p_offset ranges on EM_X86_64
Port the D64906 technique to EM_X86_64.

Differential Revision: https://reviews.llvm.org/D67482

llvm-svn: 371958
2019-09-16 07:05:34 +00:00
Fangrui Song e8710ef1fb [ELF] Split RW PT_LOAD on the PT_GNU_RELRO boundary
Summary:
Based on Peter Collingbourne's suggestion in D56828.

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

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

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

Reviewers: espindola, ruiu, pcc

Reviewed By: ruiu

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

Tags: #llvm

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

llvm-svn: 356226
2019-03-15 01:29:57 +00:00
Fangrui Song 07f8daf05e [ELF] Simplify RelRo, TLS, NOBITS section ranks and make RW PT_LOAD start with RelRo
Old: PT_LOAD(.data | PT_GNU_RELRO(.data.rel.ro .bss.rel.ro) | .bss)
New: PT_LOAD(PT_GNU_RELRO(.data.rel.ro .bss.rel.ro) | .data .bss)

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

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

Other advantages:

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

Reviewers: ruiu, espindola, pcc

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

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

llvm-svn: 356117
2019-03-14 03:47:45 +00:00
Fangrui Song 3d87323a7e [ELF] Make non-writable non-executable PROGBITS sections closer to .text
This generalizes the old heuristic placing SHT_DYNSYM SHT_DYNSTR first in the readonly SHF_ALLOC segment.

Reviewers: espindola

Subscribers: emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D48406

llvm-svn: 335674
2018-06-26 22:13:32 +00:00
Han Shen b56030ee9e [lld] Mitigate relocation overflow [part 1 of 2].
This CL places .dynsym and .dynstr at the beginning of SHF_ALLOC
sections. We do this to mitigate the possibility that huge .dynsym and
.dynstr sections placed between ro-data and text sections cause
relocation overflow.

Differential Revision: https://reviews.llvm.org/D45788

llvm-svn: 332374
2018-05-15 17:02:35 +00:00
George Rimar 4c9ae67bb7 [ELF] - Revert of: r332038, r332054, r332060, r332061, r332062, r332063
This reverts "Mitigate relocation overflow [part 1 of 2]." and the following commits which
were trying to fix the bots.

At the moment of r332082, bots are still failing and we need to find the reason of test case breakages first of all.
http://lab.llvm.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-windows10pro-fast/builds/17042/steps/test/logs/stdio
http://lab.llvm.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/builds/29845/steps/test/logs/stdio

llvm-svn: 332085
2018-05-11 08:11:25 +00:00
Han Shen 6c0881c3cd Mitigate relocation overflow [part 1 of 2].
This CL is to mitigate R_X86_64_PC32 relocation overflow problems for huge binaries that has near 4G allocated sections.

By examining those binaries, there're 2 issues contributes to the problem:
1). huge ".dynsym" and ".dynstr" stands in the way between .rodata and .text
2). _init_array_start/end are placed at 0 if no ".init_array" presents, this causes .text relocation against them become more prone to overflow.

This CL addresses 1st problem (the 2nd will be addressed in another CL.) by assigning a smaller sortrank to .dynsym and .dynstr thus they no longer stand in between.

llvm-svn: 332038
2018-05-10 20:44:42 +00:00
Fangrui Song 95851515d7 Add a CIE with length 0 unconditionally.
Summary: This is not technically required, but glibc unwind-dw2-fde.c classify_object_over_fdes expects there is a CIE record length 0 as a terminator.

Reviewers: ruiu, espindola

Subscribers: emaste, arichardson, llvm-commits

Differential Revision: https://reviews.llvm.org/D46566

llvm-svn: 331708
2018-05-08 01:19:16 +00:00
Rafael Espindola ab0cce5f1f Replace SharedSymbols with Defined when creating copy relocations.
This is slightly simpler to read IMHO. Now if a symbol has a position
in the file, it is Defined.

The main motivation is that with this a SharedSymbol doesn't need a
section, which reduces the size of SymbolUnion.

With this the peak allocation when linking chromium goes from 568.1 to
564.2 MB.

llvm-svn: 330966
2018-04-26 17:58:58 +00:00
Rui Ueyama 4e150c4c88 Make "Size" column in the map file one characters shorter.
Previously, "size" column is 9 characters long which is too long
at least for 32-bit (because at maximum it needs 8 columns). This
patch make it one column shorter than before. That's also a reasonable
default for 64-bit.

llvm-svn: 329317
2018-04-05 17:20:18 +00:00
Rui Ueyama 16a130bc78 Fix column size in the map file.
Size can be narrow, but LMA should be the same width as VMA.

llvm-svn: 329312
2018-04-05 16:45:37 +00:00
George Rimar ee01b1d390 [ELF] - Print LMA in a -Map file.
Currently, LLD prints VA, but not LMA in a map file.
It seems can be useful to print both to reveal layout
details and patch implements it.

Differential revision: https://reviews.llvm.org/D44899

llvm-svn: 329271
2018-04-05 10:51:06 +00:00
Rui Ueyama 5a9529f4a2 Do not show alignment 0 because that is equivalent to 1.
Differential Revision: https://reviews.llvm.org/D44991

llvm-svn: 329233
2018-04-04 21:25:37 +00:00
Rui Ueyama a5c1ed340f [ELF] Add .eh_frame pieces to map file
This patch is a simplified version of https://reviews.llvm.org/D42960
written by Andrew Ng.

Differential Revision: https://reviews.llvm.org/D44168

llvm-svn: 327574
2018-03-14 21:18:18 +00:00
Rafael Espindola f931c179ca Don't depend on "call foo" producing a X86_64_PC32.
Newer versions of the gnu assembler produce a X86_64_PLT32 for
calls. There is a change under review in llvm to do the same, so update
the tests to not depend on it.

We can still produce a R_X86_64_PC32 with ".long foo - .".

llvm-svn: 325379
2018-02-16 20:05:58 +00:00
George Rimar ba7afd7043 [ELF] - Show .plt symbols in -Map output.
As mentioned in PR35471, shared functions for which
.plt entry address is used shows up in bfd's map files. 
Patch teaches LLD to do the same.

Differential revision: https://reviews.llvm.org/D40839

llvm-svn: 319879
2017-12-06 08:36:21 +00:00
George Rimar 31cd5da582 Recommit r319747 "[ELF] - Show copy relocated symbols in a --Map file output."
With fix:
Specify -soname for input dso to fix up the .dynstr section
size in different environments.

Original commit message:

As mentioned in PR35471, copied symbols did not show
in --Map output. Patch fixes that.

Differential revision: https://reviews.llvm.org/D40785

llvm-svn: 319769
2017-12-05 14:23:18 +00:00
George Rimar fe52e0064b [ELF] - Revert r319747 "[ELF] - Show copy relocated symbols in a --Map file output.".
It broke bots:
http://lab.llvm.org:8011/builders/lld-x86_64-darwin13/builds/15570/steps/test_lld/logs/stdio
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/10930/steps/check-lld%20asan/logs/stdio
http://lab.llvm.org:8011/builders/lld-x86_64-win7/builds/16336/steps/test%20lld/logs/stdio

According to outputs,
.dynstr section has different sizes for some reason.

llvm-svn: 319749
2017-12-05 09:38:05 +00:00
George Rimar d9ae97cdb6 [ELF] - Show copy relocated symbols in a --Map file output.
As mentioned in PR35471, copied symbols did not show
in --Map output. Patch fixes that.

Differential revision: https://reviews.llvm.org/D40785

llvm-svn: 319747
2017-12-05 09:13:45 +00:00
Dmitry Mikulin 1e30f07ce7 Currently lld creates a single section to collect all commons. There is no way
to separate commons based on file name patterns. The following linker script
construct does not work because commons are allocated before section placement
is done and the only synthesized BssSection that holds all commons has no file
associated with it:
SECTIONS { .common_0 : { *file0.o(COMMON) }}

This patch changes the allocation of commons to create a section per common
symbol and let the section logic do the layout.

Differential revision: https://reviews.llvm.org/D37489

llvm-svn: 312796
2017-09-08 16:22:43 +00:00
Rafael Espindola db1af69342 Avoid crash with local abs symbol.
llvm-svn: 310736
2017-08-11 17:47:12 +00:00
George Rimar f2fe963d83 [ELF] - Do not omit common symbols when -Map is given.
This is PR33886,

previously we did not output common symbols to map,
patch fixes that.

Differential revision: https://reviews.llvm.org/D36466

llvm-svn: 310703
2017-08-11 11:34:04 +00:00
Rui Ueyama 3012b371fd Change the format of the map file.
Previously, we printed out input sections and input files in
separate columns as shown below.

  Address          Size             Align Out     In      File    Symbol
  0000000000201000 0000000000000015     4 .text
  0000000000201000 000000000000000e     4         .text
  0000000000201000 000000000000000e     4                 foo.o
  0000000000201000 0000000000000000     0                         _start
  0000000000201005 0000000000000000     0                         f(int)
  000000000020100e 0000000000000000     0                         local
  0000000000201010 0000000000000002     4                 bar.o
  0000000000201010 0000000000000000     0                         foo
  0000000000201011 0000000000000000     0                         bar

This format doesn't make much sense because for each input section,
there's always exactly one input file. This patch changes the format
to this.

  Address          Size             Align Out     In      Symbol
  0000000000201000 0000000000000015     4 .text
  0000000000201000 000000000000000e     4         foo.o:(.text)
  0000000000201000 0000000000000000     0                 _start
  0000000000201005 0000000000000000     0                 f(int)
  000000000020100e 0000000000000000     0                 local
  0000000000201010 0000000000000002     4         bar.o:(.text)
  0000000000201010 0000000000000000     0                 foo
  0000000000201011 0000000000000000     0                 bar

Differential Revision: https://reviews.llvm.org/D32657

llvm-svn: 301683
2017-04-28 20:38:27 +00:00
Rui Ueyama b882e5910d Speed up the -Map option.
We found that some part of code for the -Map option takes O(m*n)
where m is the number of input sections in some file and n is
the number of symbols in the same file. If you do LTO, we usually
have only a few object files as inputs for the -Map option
feature, so this performance characteristic was worse than I
expected.

This patch rewrites the -Map option feature to speed it up.
I eliminated the O(m*n) bottleneck and also used multi-threading.

As a result, clang link time with the -Map option improved from
18.7 seconds to 11.2 seconds. Without -Map, it takes 7.7 seconds,
so the -Map option is now about 3x faster than before for this
test case (from 11.0 seconds to 3.5 seconds.) The generated output
file size was 223 MiB, and the file contains 1.2M lines.

Differential Revision: https://reviews.llvm.org/D32631

llvm-svn: 301659
2017-04-28 17:19:13 +00:00
James Henderson b7a90ef48e [ELF] Fail the link early if the map file path is invalid
As with the changes made in r297645, we do not want a potentially long link to
be run, if it will ultimately fail because the map file is not writable. This
change reuses the same functionality as the output file path check. See
https://reviews.llvm.org/D30449 for further justification and explanations.

Reviewers: ruiu

Differential Revision: https://reviews.llvm.org/D31603

llvm-svn: 299420
2017-04-04 09:42:24 +00:00
Rafael Espindola b691ccf0a5 Revert "Add terminator to .eh_frame sections"
This reverts commit r296378.

I am pretty sure this is incorrect. In particular, for just

        .cfi_startproc
        nop
        .cfi_endproc

We now add an extra 4 zeros that neither bfd nor gold add.

llvm-svn: 296503
2017-02-28 18:55:08 +00:00
Rui Ueyama 1720ef1343 Add terminator to .eh_frame sections
Patch by Mark Kettenis.

Currenlty ld.lld does not add a terminator (a CIE with its length field
set to zero) to the .eh_frame sections it generates.  While the relevant
standards (the AMD64 SysV ABI and the Linux LSB) are not explicit about
this, such a terminator is expected by some unwinder implementations and
seems to be always emitted by ld.bfd. In addition to that, the Linux LSB

  https://refspecs.linuxfoundation.org/LSB_5.0.0/LSB-Core-generic/LSB-Core-generic/ehframechpt.html#EHFRAME

explicitly says that

  The .eh_frame section shall contain 1 or more Call Frame Information
  (CFI) records.

Currently, if the .eh_frame sections of the input files only contain
terminators, ld.lld emits a zero=sized .eh_frame section
which clearly doesn't meet that requirement.

The diff makes sure a terminator gets added to each .eh_frame section
and adjusts all the relevant tests to account for that.  An additional
test isn't needed as these adjustments mean that the existence of the
terminator is tested for by several tests already.

Differential Revision: https://reviews.llvm.org/D30335

llvm-svn: 296378
2017-02-27 20:44:59 +00:00
Rafael Espindola 66b4e21534 Convert EhOutputSection to be a synthetic section.
With this we complete the transition out of special output sections,
and with the previous patches it should be possible to merge
OutputSectionBase and OuputSection.

llvm-svn: 296023
2017-02-23 22:06:28 +00:00
Rafael Espindola 9e9754b520 Replace MergeOutputSection with a synthetic section.
With a synthetic merge section we can have, for example, a single
.rodata section with stings, fixed sized constants and non merge
constants.

I can be simplified further by not setting Entsize, but that is
probably better done is a followup patch.

This should allow some cleanup in the linker script code now that
every output section command maps to just one output section.

llvm-svn: 294005
2017-02-03 13:06:18 +00:00
Rui Ueyama eb1ad400f9 Print alignment in decimal instead of hexadecimal.
Previously, we were printing out something like this for
sections/symbols with alignment 16

  0000000000201000 0000000000000182    10 .data

which I think confusing. I think printing it in decimal is better.

Differential Revision: https://reviews.llvm.org/D29258

llvm-svn: 293685
2017-01-31 20:42:34 +00:00
Rui Ueyama 40eaa9968d Return early if writeMapFile failed.
This patch adds a test for an invalid output path for -Map option,
though that test is not for verifying that we are using error()
instead of fatal() in writeMapFile.

llvm-svn: 292336
2017-01-18 03:34:38 +00:00
Rui Ueyama 1705f99c77 Add -print-map and -M options.
llvm-svn: 292046
2017-01-15 02:52:34 +00:00
Rafael Espindola 7c783a49a2 Try to fix the windows bots.
llvm-svn: 291989
2017-01-14 02:45:53 +00:00
Rafael Espindola 1ebfc59c89 Implement -Map.
The format is not exactly the same as the one in bfd since bfd always
follows a linker script and prints it along.

llvm-svn: 291958
2017-01-13 21:05:46 +00:00