grimar (George Rimar)
User

Projects

User does not belong to any projects.

User Details

User Since
Sep 21 2015, 12:36 AM (83 w, 1 d)

Recent Activity

Today

grimar added a comment to D32484: [ELF] Improve error message for incompatible section flags.

It seems both old and new way will not show proper location if section is coming from achieve object ?
We have std::string lld::toString(const InputFile *F) that can return "foo.a(bar.o)" for such case.

Also original implementation and one from this patch does not show where Sec is coming from.
And "for .bar" part probably does not make much sense because its already stated "with others with the same name "

I would do message more explicit, I am thinking about something like next format:

 Section .foo has flags incompatible with others with the same name:
>>> defined in foo.a(bar.o) with flags 0x403
>>> defined in zed.o with flags 0x3

Rui, what do you think ?

I think that output format looks better but I don't see how we can determine whichinput file the OutputSection flags are derived from as they are a bitwise or of multiple files.
I guess one option is adding list of InputSections that caused the Flags field to be updated (which would be empty if Sec->Flags == Sec->Sections[0]->Flags)?

Also with very long paths it might make sense to have the flags printed first?

Tue, Apr 25, 6:18 AM · lld
grimar added a comment to D32484: [ELF] Improve error message for incompatible section flags.

Ah sorry, Sec is an output section here. My suggestion will not work :) Need to think more about it.

Tue, Apr 25, 6:01 AM · lld
grimar added a comment to D32484: [ELF] Improve error message for incompatible section flags.

It seems both old and new way will not show proper location if section is coming from achieve object ?
We have std::string lld::toString(const InputFile *F) that can return "foo.a(bar.o)" for such case.

Tue, Apr 25, 5:57 AM · lld
grimar added inline comments to D32171: [ELF] - Implemented --defsym option.
Tue, Apr 25, 4:06 AM
grimar updated the diff for D32171: [ELF] - Implemented --defsym option.
  • Addressed review comments.
Tue, Apr 25, 4:06 AM
grimar retitled D32171: [ELF] - Implemented --defsym option from [ELF] - Implemented --defsym option #2 to [ELF] - Implemented --defsym option.
Tue, Apr 25, 3:50 AM
grimar abandoned D32082: [ELF] - Implemented --defsym option.

D32171 is on review instead

Tue, Apr 25, 3:50 AM
grimar added a comment to D32173: [ELF] - Assign absolute values in linkerscript correctly #1..

Comment above should be: "Honestly I think I do not like idea to calculate something..."

Tue, Apr 25, 3:44 AM
grimar added a comment to D32173: [ELF] - Assign absolute values in linkerscript correctly #1..
In D32173#736295, @ruiu wrote:

Sorry for the delay. I was on vacation last week.

So, this patch allows you compute absolute symbols addresses lazily. With that, you can create cycles like this.

foo = ABSOLUTE(bar - 0xff);
bar = ABSOLUTE(foo - 0xff);

What do you think it should be handled?

Also, you can create not a cycle but a series of expressions that should be calculated backwards, like this.

foo = ABSOLUTE(bar);
bar = ABSOLUTE(baz);
baz = 0xff;

What do you think about this?

Tue, Apr 25, 3:43 AM
grimar added inline comments to D32289: [ELF] - Linkerscript: support combination of linkerscript and --compress-debug-sections..
Tue, Apr 25, 3:27 AM

Yesterday

grimar added inline comments to D32228: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges.
Mon, Apr 24, 3:33 AM
grimar committed rL301170: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges.
[DWARF] - Take relocations in account when extracting ranges from .debug_ranges
Mon, Apr 24, 3:33 AM
grimar closed D32228: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges by committing rL301170: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges.
Mon, Apr 24, 3:32 AM

Fri, Apr 21

grimar updated the summary of D32355: [ELF] - Remove dead TLS relocations relative code for MIPS and ARM.
Fri, Apr 21, 8:44 AM
grimar created D32355: [ELF] - Remove dead TLS relocations relative code for MIPS and ARM.
Fri, Apr 21, 8:40 AM
grimar created D32354: [ELF] - Set DF_STATIC_TLS flag for i386 target..
Fri, Apr 21, 8:38 AM
grimar added inline comments to D32228: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges.
Fri, Apr 21, 3:20 AM
grimar updated the diff for D32228: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges.
  • Rebased.
  • Addressed review comments.
Fri, Apr 21, 3:20 AM
grimar added a comment to D32284: [DWARF] - Refactoring: localize handling of relocations in a single place..

Thanks, Adrian !

Fri, Apr 21, 2:26 AM
grimar committed rL300956: [DWARF] - Refactoring: localize handling of relocations in a single place..
[DWARF] - Refactoring: localize handling of relocations in a single place.
Fri, Apr 21, 2:25 AM
grimar closed D32284: [DWARF] - Refactoring: localize handling of relocations in a single place. by committing rL300956: [DWARF] - Refactoring: localize handling of relocations in a single place..
Fri, Apr 21, 2:25 AM
grimar accepted D32336: [ELF] Account for R_386_TLS_LDO_32 addend.

This looks good to me. Please wait for approve from Rui and/or Rafael.

Fri, Apr 21, 1:33 AM
grimar edited reviewers for D32336: [ELF] Account for R_386_TLS_LDO_32 addend, added: rafael; removed: rafael.espindola.
Fri, Apr 21, 1:30 AM

Thu, Apr 20

grimar created D32289: [ELF] - Linkerscript: support combination of linkerscript and --compress-debug-sections..
Thu, Apr 20, 5:23 AM
grimar added inline comments to D32171: [ELF] - Implemented --defsym option.
Thu, Apr 20, 3:56 AM
grimar updated the diff for D32171: [ELF] - Implemented --defsym option.
  • Addressed review comments.
Thu, Apr 20, 3:55 AM
grimar abandoned D31464: [ELF] - Stop producing broken entries in .debug_ranges section.

Its not yet clear how much it is really needed, abandoning for now.

Thu, Apr 20, 3:21 AM
grimar created D32284: [DWARF] - Refactoring: localize handling of relocations in a single place..
Thu, Apr 20, 3:08 AM
grimar added a comment to D32228: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges.

I suggest to split refactoring change that introduced getRelocatedValue() to different patch, because it touches many
places and landing it separatelly will allow to reduce amount of changes in this patch (after rebasing) and simplify the review probably.
I'll post it very soon.

Thu, Apr 20, 2:59 AM
grimar added inline comments to D32228: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges.
Thu, Apr 20, 2:36 AM
grimar updated the diff for D32228: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges.

Thanks for review, David ! All comments were addressed.

Thu, Apr 20, 2:36 AM

Wed, Apr 19

grimar retitled D32228: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges from [DWARF] - Take relocations in account when exctracting ranges from .debug_ranges to [DWARF] - Take relocations in account when extracting ranges from .debug_ranges.
Wed, Apr 19, 8:32 AM
grimar created D32228: [DWARF] - Take relocations in account when extracting ranges from .debug_ranges.
Wed, Apr 19, 8:27 AM
grimar added a comment to D32173: [ELF] - Assign absolute values in linkerscript correctly #1..
In D32173#730327, @ruiu wrote:

Is your example linker script valid for GNU linkers? I'm curious why you noticed this in the first place.

Wed, Apr 19, 5:41 AM
grimar added inline comments to D32171: [ELF] - Implemented --defsym option.
Wed, Apr 19, 12:43 AM
grimar updated the diff for D32171: [ELF] - Implemented --defsym option.
  • Updated testcase to show we can use aliasing.
Wed, Apr 19, 12:43 AM

Tue, Apr 18

grimar added inline comments to D32171: [ELF] - Implemented --defsym option.
Tue, Apr 18, 9:31 AM
grimar updated the diff for D32171: [ELF] - Implemented --defsym option.
  • Addressed review comment.
Tue, Apr 18, 9:31 AM
grimar updated the summary of D32173: [ELF] - Assign absolute values in linkerscript correctly #1..
Tue, Apr 18, 9:24 AM
grimar created D32174: [ELF] - Assign absolute values in linkerscript correctly #2..
Tue, Apr 18, 9:24 AM
grimar created D32173: [ELF] - Assign absolute values in linkerscript correctly #1..
Tue, Apr 18, 9:21 AM
grimar created D32171: [ELF] - Implemented --defsym option.
Tue, Apr 18, 9:19 AM
grimar abandoned D31943: [Support] - Implemented zlib::StreamCompression class..

Was not used in LLD, abandoning for now.

Tue, Apr 18, 6:41 AM

Mon, Apr 17

grimar added inline comments to D31941: [ELF] - Implemented --compress-debug-sections option..
Mon, Apr 17, 2:22 AM
grimar committed rL300445: [ELF] - Add compress-debug-output-i386.s testcase..
[ELF] - Add compress-debug-output-i386.s testcase.
Mon, Apr 17, 2:16 AM
grimar committed rL300444: [ELF] - Implemented --compress-debug-sections option..
[ELF] - Implemented --compress-debug-sections option.
Mon, Apr 17, 2:11 AM
grimar added a comment to D31941: [ELF] - Implemented --compress-debug-sections option..

Thanks for review, Rui !

Mon, Apr 17, 2:11 AM
grimar closed D31941: [ELF] - Implemented --compress-debug-sections option. by committing rL300444: [ELF] - Implemented --compress-debug-sections option..
Mon, Apr 17, 2:11 AM
grimar added a comment to D32082: [ELF] - Implemented --defsym option.
In D32082#727317, @ruiu wrote:

This doesn't look good as you changed too many places for this little feature. Do you really have to support expressions that contain operators? I think you can start with supporting an assignments that consist of only two symbols in the form of -defsym=<sym1>=<sym2>.

Mon, Apr 17, 12:16 AM

Fri, Apr 14

grimar updated the diff for D32082: [ELF] - Implemented --defsym option.
  • Removed useless change.
Fri, Apr 14, 7:19 AM
grimar created D32082: [ELF] - Implemented --defsym option.
Fri, Apr 14, 7:11 AM
grimar added inline comments to D31941: [ELF] - Implemented --compress-debug-sections option..
Fri, Apr 14, 3:34 AM
grimar updated the diff for D31941: [ELF] - Implemented --compress-debug-sections option..
  • Addressed review comments.
Fri, Apr 14, 3:34 AM
grimar committed rL300317: [ELF] - Linkerscript: make section with no content to be SHT_PROGBITS by….
[ELF] - Linkerscript: make section with no content to be SHT_PROGBITS by…
Fri, Apr 14, 2:49 AM
grimar closed D32016: [ELF] - Linkerscript: make section with no content to be SHT_PROGBITS by default. by committing rL300317: [ELF] - Linkerscript: make section with no content to be SHT_PROGBITS by….
Fri, Apr 14, 2:49 AM
grimar committed rL300316: [ELF] - Cleanup of align.s testcase. NFC..
[ELF] - Cleanup of align.s testcase. NFC.
Fri, Apr 14, 2:43 AM
grimar committed rL300315: [ELF] LinkerScript: Don't assign zero to all regular symbols.
[ELF] LinkerScript: Don't assign zero to all regular symbols
Fri, Apr 14, 2:36 AM
grimar closed D31984: [ELF] LinkerScript: Don't assign zero to all regular symbols by committing rL300315: [ELF] LinkerScript: Don't assign zero to all regular symbols.
Fri, Apr 14, 2:36 AM

Thu, Apr 13

grimar created D32016: [ELF] - Linkerscript: make section with no content to be SHT_PROGBITS by default..
Thu, Apr 13, 8:04 AM
grimar updated the summary of D31941: [ELF] - Implemented --compress-debug-sections option..
Thu, Apr 13, 6:19 AM
grimar added a comment to D31941: [ELF] - Implemented --compress-debug-sections option..
In D31941#725216, @ruiu wrote:

What I was saying was that you can compress multiple sections concurrently, right? I didn't say you should try to compress single stream of data using multiple threads. I meant, since we have multiple .debug_ sections, you could compress them separately.

Thu, Apr 13, 6:12 AM
grimar removed a dependency for D31941: [ELF] - Implemented --compress-debug-sections option.: D31943: [Support] - Implemented zlib::StreamCompression class..
Thu, Apr 13, 6:11 AM
grimar removed a dependent revision for D31943: [Support] - Implemented zlib::StreamCompression class.: D31941: [ELF] - Implemented --compress-debug-sections option..
Thu, Apr 13, 6:11 AM
grimar updated the diff for D31941: [ELF] - Implemented --compress-debug-sections option..
  • Added multithreaded compression of output sections.
  • Removed streaming zlib compression.
  • Addressed review comments.
Thu, Apr 13, 6:10 AM
grimar committed rL300188: [DWARF] - Simplify (use dyn_cast instead of isa + cast)..
[DWARF] - Simplify (use dyn_cast instead of isa + cast).
Thu, Apr 13, 3:05 AM
grimar updated the diff for D31943: [Support] - Implemented zlib::StreamCompression class..
  • Updated stubs implementation for case when zlib was not available(defined) (update of signatures and names, code was not compilable in that case).
Thu, Apr 13, 2:52 AM

Wed, Apr 12

grimar added inline comments to D31943: [Support] - Implemented zlib::StreamCompression class..
Wed, Apr 12, 9:03 AM
grimar updated the diff for D31943: [Support] - Implemented zlib::StreamCompression class..

Thanks for review, David ! All comments addressed.

Wed, Apr 12, 9:03 AM
grimar added inline comments to D31747: [DWARF] - Refactoring of DWARFContextInMemory implementation..
Wed, Apr 12, 8:36 AM
grimar added a comment to D31941: [ELF] - Implemented --compress-debug-sections option..
In D31941#723864, @ruiu wrote:

This patch doesn't use multiple cores to compress .debug section contents. In theory, if there are multiple .debug sections, we can compress them separately. Can you do that?

Wed, Apr 12, 5:40 AM
grimar added a comment to D31747: [DWARF] - Refactoring of DWARFContextInMemory implementation..

Thanks for review, Adrian !

Wed, Apr 12, 2:14 AM
grimar committed rL300039: [DWARF] - Refactoring of DWARFContextInMemory implementation..
[DWARF] - Refactoring of DWARFContextInMemory implementation.
Wed, Apr 12, 2:11 AM
grimar closed D31747: [DWARF] - Refactoring of DWARFContextInMemory implementation. by committing rL300039: [DWARF] - Refactoring of DWARFContextInMemory implementation..
Wed, Apr 12, 2:11 AM

Tue, Apr 11

grimar updated the diff for D31943: [Support] - Implemented zlib::StreamCompression class..
  • Last minux fix for llvm_unreachable messages text.
Tue, Apr 11, 9:08 AM
grimar added a dependent revision for D31943: [Support] - Implemented zlib::StreamCompression class.: D31941: [ELF] - Implemented --compress-debug-sections option..
Tue, Apr 11, 8:55 AM
grimar added a dependency for D31941: [ELF] - Implemented --compress-debug-sections option.: D31943: [Support] - Implemented zlib::StreamCompression class..
Tue, Apr 11, 8:55 AM
grimar created D31943: [Support] - Implemented zlib::StreamCompression class..
Tue, Apr 11, 8:54 AM
grimar edited reviewers for D31941: [ELF] - Implemented --compress-debug-sections option., added: ruiu; removed: rui314.
Tue, Apr 11, 8:51 AM
grimar created D31941: [ELF] - Implemented --compress-debug-sections option..
Tue, Apr 11, 8:50 AM

Sat, Apr 8

grimar added inline comments to D31747: [DWARF] - Refactoring of DWARFContextInMemory implementation..
Sat, Apr 8, 12:15 AM
grimar updated the diff for D31747: [DWARF] - Refactoring of DWARFContextInMemory implementation..
  • Addressed review comments.
Sat, Apr 8, 12:15 AM

Fri, Apr 7

grimar closed D30699: [ELF] - Stop producing broken output for R_386_GOT32X relocation..

r299812

Fri, Apr 7, 11:29 PM
grimar committed rL299812: [ELF] - Stop producing broken output for R_386_GOT32[X] relocations..
[ELF] - Stop producing broken output for R_386_GOT32[X] relocations.
Fri, Apr 7, 11:27 PM
grimar added inline comments to D30699: [ELF] - Stop producing broken output for R_386_GOT32X relocation..
Fri, Apr 7, 11:24 PM
grimar updated the summary of D30699: [ELF] - Stop producing broken output for R_386_GOT32X relocation..
Fri, Apr 7, 11:20 PM

Thu, Apr 6

grimar updated the diff for D31747: [DWARF] - Refactoring of DWARFContextInMemory implementation..
  • Removed unrelative clang-format changes (file is not clang-formated, I think we want to format it separatelly).
Thu, Apr 6, 2:31 AM
grimar created D31747: [DWARF] - Refactoring of DWARFContextInMemory implementation..
Thu, Apr 6, 2:26 AM
grimar added inline comments to D30699: [ELF] - Stop producing broken output for R_386_GOT32X relocation..
Thu, Apr 6, 12:58 AM
grimar updated the diff for D30699: [ELF] - Stop producing broken output for R_386_GOT32X relocation..
  • Addressed review comments.
Thu, Apr 6, 12:57 AM

Wed, Apr 5

grimar updated the diff for D30699: [ELF] - Stop producing broken output for R_386_GOT32X relocation..
  • Addressed review comments,
Wed, Apr 5, 4:28 AM
grimar committed rL299528: [ELF] - Define __bss_start symbol..
[ELF] - Define __bss_start symbol.
Wed, Apr 5, 3:16 AM
grimar closed D30419: [ELF] - Define __bss_start symbol. by committing rL299528: [ELF] - Define __bss_start symbol..
Wed, Apr 5, 3:16 AM
grimar committed rL299525: Revert r299506 "Simplify. NFC.".
Revert r299506 "Simplify. NFC."
Wed, Apr 5, 2:32 AM

Tue, Apr 4

grimar updated the diff for D30419: [ELF] - Define __bss_start symbol..
  • Addressed review comment properly (previously I changed naming ElfSym::BssStart to ElfSym::Bss,

but did not update comment as was requested, now it is fixed).

Tue, Apr 4, 9:35 PM
grimar added a comment to D31136: [DWARF] - Speedup handling of relocations in DWARFContextInMemory..

Ping. (not sure we will use this change in LLD, since D31424 looks better, but it still looks like a good cleanup and speedup for other users. If no I'll abandon this).

Tue, Apr 4, 5:46 AM
grimar added a comment to D30419: [ELF] - Define __bss_start symbol..
In D30419#714407, @ruiu wrote:

You don't really need three tests because in this feature you don't really need to (re-)verify that our section merging logic is working fine. Just reduce it to one test.

Tue, Apr 4, 5:19 AM
grimar updated the diff for D30419: [ELF] - Define __bss_start symbol..
  • Addressed review comments.
Tue, Apr 4, 5:17 AM
grimar added inline comments to D30699: [ELF] - Stop producing broken output for R_386_GOT32X relocation..
Tue, Apr 4, 3:47 AM
grimar updated the diff for D30699: [ELF] - Stop producing broken output for R_386_GOT32X relocation..
  • Addressed review comments.
Tue, Apr 4, 3:47 AM
grimar added inline comments to D31464: [ELF] - Stop producing broken entries in .debug_ranges section.
Tue, Apr 4, 3:30 AM