grimar (George Rimar)
User

Projects

User does not belong to any projects.

User Details

User Since
Sep 21 2015, 12:36 AM (108 w, 17 h)

Recent Activity

Today

grimar added inline comments to D38846: [ELF] - Linkerscript: Fix issues with SUBALIGN..
Mon, Oct 16, 9:35 AM
grimar added inline comments to D38724: [ELF] - Do not collect SHT_REL[A] sections unconditionally when --gc-sections and --emit-relocs used together..
Mon, Oct 16, 6:15 AM
grimar updated the diff for D38724: [ELF] - Do not collect SHT_REL[A] sections unconditionally when --gc-sections and --emit-relocs used together..
  • Addressed comments, added testcase.
Mon, Oct 16, 6:14 AM
grimar added inline comments to D38239: [ELF] - Define linkerscript symbols early..
Mon, Oct 16, 5:10 AM
grimar updated the diff for D38239: [ELF] - Define linkerscript symbols early..
  • Addressed review comments.
Mon, Oct 16, 5:10 AM
grimar committed rL315897: [llvm-dwarfdump] - Teach tool to parse DW_CFA_GNU_args_size..
[llvm-dwarfdump] - Teach tool to parse DW_CFA_GNU_args_size.
Mon, Oct 16, 3:26 AM
grimar closed D38879: [llvm-dwarfdump] - Teach tool to parse DW_CFA_GNU_args_size. by committing rL315897: [llvm-dwarfdump] - Teach tool to parse DW_CFA_GNU_args_size..
Mon, Oct 16, 3:26 AM
grimar added a comment to D38879: [llvm-dwarfdump] - Teach tool to parse DW_CFA_GNU_args_size..

Do you need to update the Verifier, too? Does llvm-dwarfdump -verify work on this input?

Mon, Oct 16, 3:23 AM
grimar added a comment to D38907: Give .note.gnu.build-id section alignment 4.

Probably needs testcase ?

Mon, Oct 16, 1:58 AM · lld
grimar added a comment to D38846: [ELF] - Linkerscript: Fix issues with SUBALIGN..
In D38846#898203, @ruiu wrote:

I'd think you are overthinking. I don't think it is a good idea to add that many parameters to various functions just to check for a misuse of some rarely used feature.

Mon, Oct 16, 1:49 AM

Fri, Oct 13

grimar updated the diff for D38239: [ELF] - Define linkerscript symbols early..
  • Addressed review comments.
Fri, Oct 13, 6:52 AM
grimar updated the diff for D38582: [ELF] - Get rid of LinkerScript::adjustSectionsBeforeSorting()..
  • Cosmetic changes.
Fri, Oct 13, 5:40 AM
grimar updated the diff for D38582: [ELF] - Get rid of LinkerScript::adjustSectionsBeforeSorting()..
  • Addressed review comments.
Fri, Oct 13, 5:26 AM
grimar created D38879: [llvm-dwarfdump] - Teach tool to parse DW_CFA_GNU_args_size..
Fri, Oct 13, 3:14 AM

Thu, Oct 12

grimar created D38846: [ELF] - Linkerscript: Fix issues with SUBALIGN..
Thu, Oct 12, 8:17 AM
grimar added inline comments to D38724: [ELF] - Do not collect SHT_REL[A] sections unconditionally when --gc-sections and --emit-relocs used together..
Thu, Oct 12, 3:42 AM
grimar updated the diff for D38724: [ELF] - Do not collect SHT_REL[A] sections unconditionally when --gc-sections and --emit-relocs used together..
  • Addressed review comments.
Thu, Oct 12, 3:42 AM
grimar added inline comments to D38239: [ELF] - Define linkerscript symbols early..
Thu, Oct 12, 2:11 AM
grimar updated the diff for D38239: [ELF] - Define linkerscript symbols early..
  • Addressed review comments.
Thu, Oct 12, 2:11 AM
grimar committed rL315569: [ELF] - Linkerscript: Add `~` as separate math token..
[ELF] - Linkerscript: Add `~` as separate math token.
Thu, Oct 12, 1:40 AM
grimar closed D36508: [ELF] - Linkerscript: Add `~` as separate math token. by committing rL315569: [ELF] - Linkerscript: Add `~` as separate math token..
Thu, Oct 12, 1:40 AM
grimar added a comment to D38721: [ELF] - Teach LLD to report line numbers for data symbols..

Sorry, I wanted to say "We want to support both ways, and this patch supports only second one, because ...", so it supports DW_FORM_string.

Thu, Oct 12, 1:00 AM

Wed, Oct 11

grimar updated the diff for D38319: [ELF] - Teach ICF to take FDEs into account when doing code folding..
  • Rebased.
Wed, Oct 11, 7:32 AM
grimar updated the diff for D38239: [ELF] - Define linkerscript symbols early..
  • Rebased/reimplemented after recent changes, added comments.
Wed, Oct 11, 6:49 AM
grimar updated the diff for D38724: [ELF] - Do not collect SHT_REL[A] sections unconditionally when --gc-sections and --emit-relocs used together..
  • Rebased, addressed comments.
Wed, Oct 11, 4:45 AM
grimar added inline comments to D38721: [ELF] - Teach LLD to report line numbers for data symbols..
Wed, Oct 11, 3:27 AM
grimar updated the diff for D38721: [ELF] - Teach LLD to report line numbers for data symbols..
  • Addressed review comments.
Wed, Oct 11, 3:27 AM
grimar added a comment to D36508: [ELF] - Linkerscript: Add `~` as separate math token..
  • Rebased.
Wed, Oct 11, 1:25 AM
grimar updated the diff for D36508: [ELF] - Linkerscript: Add `~` as separate math token..
Wed, Oct 11, 1:24 AM
grimar committed rL315442: [ELF] - Fix out of sync comment. NFC..
[ELF] - Fix out of sync comment. NFC.
Wed, Oct 11, 1:19 AM
grimar committed rL315441: [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}..
[ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}.
Wed, Oct 11, 1:14 AM
grimar closed D37736: [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}. by committing rL315441: [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}..
Wed, Oct 11, 1:14 AM
grimar retitled D37736: [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}. from [ELF] - Do not spread specific flags when emiting output sections. to [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}..
Wed, Oct 11, 1:12 AM

Tue, Oct 10

grimar updated the diff for D37736: [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}..
  • Updated comment.
Tue, Oct 10, 9:41 AM
grimar updated the diff for D38721: [ELF] - Teach LLD to report line numbers for data symbols..
  • Removed dead method, simplified testcase.
Tue, Oct 10, 8:26 AM
grimar added inline comments to D38721: [ELF] - Teach LLD to report line numbers for data symbols..
Tue, Oct 10, 8:05 AM
grimar updated the diff for D38721: [ELF] - Teach LLD to report line numbers for data symbols..
  • Fixed issue found by David.
Tue, Oct 10, 8:05 AM
grimar created D38724: [ELF] - Do not collect SHT_REL[A] sections unconditionally when --gc-sections and --emit-relocs used together..
Tue, Oct 10, 6:16 AM
grimar created D38721: [ELF] - Teach LLD to report line numbers for data symbols..
Tue, Oct 10, 5:07 AM
grimar updated the diff for D37736: [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}..
  • Included summary comment from discussion thread.
Tue, Oct 10, 1:35 AM
grimar added a comment to D37520: [ELF] - Fix removing of unused synthetic sections..

Ping.

Tue, Oct 10, 1:12 AM
grimar abandoned D37733: [ELF] - Implement --push-state/--pop-state..

Abandoning until will be decided LLD wants to support it.

Tue, Oct 10, 1:12 AM
grimar abandoned D38170: [ELF] - Drop SHF_LINK_ORDER flag from output..

In favor of D37736.

Tue, Oct 10, 1:09 AM
grimar abandoned D38613: [ELF] - Change how we deal with unused synthetic sections..
Tue, Oct 10, 1:07 AM
grimar abandoned D36579: [ELF] - Do not fail when set versions for linkerscript's symbol aliases.

In favor of D38239.

Tue, Oct 10, 1:06 AM
grimar abandoned D37731: [ELF] - Another fix for "Do not fail when set versions for linkerscript's symbol aliases".

In favor of D38239.

Tue, Oct 10, 1:06 AM
grimar committed rL315280: [ELF] - Improve "has non-ABS reloc" error..
[ELF] - Improve "has non-ABS reloc" error.
Tue, Oct 10, 12:55 AM
grimar closed D38623: [ELF] - Improve "has non-ABS reloc" error. by committing rL315280: [ELF] - Improve "has non-ABS reloc" error..
Tue, Oct 10, 12:55 AM
grimar added inline comments to D38712: [ELF] Try to not emit weird diagnostics on undefined symbols.
Tue, Oct 10, 12:47 AM

Mon, Oct 9

grimar added inline comments to D38570: Support: Rewrite Windows implementation of sys::fs::rename to be more POSIXy..
Mon, Oct 9, 10:13 AM
grimar added a comment to D36508: [ELF] - Linkerscript: Add `~` as separate math token..

ld.bfd behavior was fixed and going to be changed in upcoming 2.30 version:
https://sourceware.org/bugzilla/show_bug.cgi?id=22267

Mon, Oct 9, 6:06 AM
grimar reclaimed D36508: [ELF] - Linkerscript: Add `~` as separate math token..
Mon, Oct 9, 6:03 AM

Sat, Oct 7

grimar added a comment to D38613: [ELF] - Change how we deal with unused synthetic sections..

I'm not sure how I feel about this.
It's true that reduces the number of lines to implement part of this feature.
OTOH, it scatters details that were in a single place into multiple (three, to be precise).
In some sense,it makes following this code harder to read.

Sat, Oct 7, 6:47 AM
grimar added a comment to D38625: [ELF] - Ignore non-absolute R_386_GOTPC relocation in debug sections..
In D38625#890908, @ruiu wrote:

Where is the bug for gcc, and what was a response from gcc develoeprs?

Sat, Oct 7, 6:16 AM

Fri, Oct 6

grimar created D38625: [ELF] - Ignore non-absolute R_386_GOTPC relocation in debug sections..
Fri, Oct 6, 7:01 AM
grimar created D38623: [ELF] - Improve "has non-ABS reloc" error..
Fri, Oct 6, 6:32 AM
grimar updated the diff for D37736: [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}..
  • Addressed review comment: added testcase for testing SHF_TLS flag spreading.
Fri, Oct 6, 3:46 AM
grimar reclaimed D37736: [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}..
In D37736#890129, @ruiu wrote:

I still don't think I understand the point of this patch. This patch in the original form says that the patch is to fix a crash bug, but it does not crash anymore (is this correct?)

No, we did not land anything to fix this crash yet.

Fri, Oct 6, 3:20 AM
grimar committed rL315054: [ELF] - Do not produce broken .dynamic section with --no-rosegment.
[ELF] - Do not produce broken .dynamic section with --no-rosegment
Fri, Oct 6, 3:08 AM
grimar closed D38580: [ELF] - Do not produce broken .dynamic section with --no-rosegment by committing rL315054: [ELF] - Do not produce broken .dynamic section with --no-rosegment.
Fri, Oct 6, 3:07 AM
grimar committed rL315053: [ELF] - Do not produce broken .dynsym with --no-rosegment..
[ELF] - Do not produce broken .dynsym with --no-rosegment.
Fri, Oct 6, 2:58 AM
grimar closed D38579: [ELF] - Do not produce broken .dynsym with --no-rosegment. by committing rL315053: [ELF] - Do not produce broken .dynsym with --no-rosegment..
Fri, Oct 6, 2:58 AM
grimar committed rL315052: [ELF] - Fix BB after r315051..
[ELF] - Fix BB after r315051.
Fri, Oct 6, 2:47 AM
grimar committed rL315051: [ELF] - Do --hash-style=both by default..
[ELF] - Do --hash-style=both by default.
Fri, Oct 6, 2:40 AM
grimar closed D38407: [ELF] - Do --hash-style=both by default. by committing rL315051: [ELF] - Do --hash-style=both by default..
Fri, Oct 6, 2:39 AM
grimar added a comment to D38393: [ELF] - Stop removing sections in removeUnusedSyntheticSections()..
In D38393#886125, @ruiu wrote:

On second thought, I think removeUnusedSyntehticSections is a source of confusion. This mechanism seems too hacky, and that's why you need to write many pieces of tricky code. I feels we could design a better mechanism. Can you visit this again to design something new?

Fri, Oct 6, 2:21 AM
grimar created D38613: [ELF] - Change how we deal with unused synthetic sections..
Fri, Oct 6, 2:18 AM

Thu, Oct 5

grimar added inline comments to D38360: [ELF] Set Dot initially to --image-base value when using linker scripts.
Thu, Oct 5, 6:51 AM
grimar updated the diff for D38582: [ELF] - Get rid of LinkerScript::adjustSectionsBeforeSorting()..
  • Better version.
Thu, Oct 5, 6:40 AM
grimar created D38582: [ELF] - Get rid of LinkerScript::adjustSectionsBeforeSorting()..
Thu, Oct 5, 6:19 AM
grimar updated the diff for D38579: [ELF] - Do not produce broken .dynsym with --no-rosegment..
  • Simplified testcase.
Thu, Oct 5, 4:39 AM
grimar updated the diff for D38580: [ELF] - Do not produce broken .dynamic section with --no-rosegment.
  • Simplified testcase.
Thu, Oct 5, 4:28 AM
grimar created D38580: [ELF] - Do not produce broken .dynamic section with --no-rosegment.
Thu, Oct 5, 4:14 AM
grimar updated the summary of D38579: [ELF] - Do not produce broken .dynsym with --no-rosegment..
Thu, Oct 5, 4:07 AM
grimar created D38579: [ELF] - Do not produce broken .dynsym with --no-rosegment..
Thu, Oct 5, 3:54 AM
grimar updated the diff for D38407: [ELF] - Do --hash-style=both by default..
  • Applied code from D38573 as requested.
Thu, Oct 5, 1:40 AM
grimar committed rL314973: [MC] - llvm-mc hangs on non-english characters..
[MC] - llvm-mc hangs on non-english characters.
Thu, Oct 5, 1:17 AM
grimar closed D38545: [MC] - llvm-mc hangs on non-english characters. by committing rL314973: [MC] - llvm-mc hangs on non-english characters..
Thu, Oct 5, 1:17 AM

Wed, Oct 4

grimar added a comment to D38545: [MC] - llvm-mc hangs on non-english characters. .
In D38545#888290, @rnk wrote:

lgtm

Wed, Oct 4, 9:16 AM
grimar created D38545: [MC] - llvm-mc hangs on non-english characters. .
Wed, Oct 4, 8:47 AM
grimar added a comment to D38239: [ELF] - Define linkerscript symbols early..

Ping.

Wed, Oct 4, 7:39 AM
grimar added a comment to D38180: [ELF] - ICF: improve support of SHF_LINK_ORDER sections..

Ping.

Wed, Oct 4, 7:37 AM
grimar abandoned D37736: [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}..
Wed, Oct 4, 7:36 AM
grimar added a comment to D37736: [ELF] - Do not set output section flags except SHF_{ALLOC,WRITE,EXECINSTR}..

Abandoning in favor of D38170.

Wed, Oct 4, 7:35 AM
grimar committed rL314899: [gold-plugin] - Fix compilation after LLVM update (r314883). NFC..
[gold-plugin] - Fix compilation after LLVM update (r314883). NFC.
Wed, Oct 4, 4:02 AM
grimar committed rL314889: [ELF] - Get rid of precompiled input objects from testcases..
[ELF] - Get rid of precompiled input objects from testcases.
Wed, Oct 4, 2:48 AM
grimar closed D38505: [ELF] - Get rid of precompiled input objects from testcases. by committing rL314889: [ELF] - Get rid of precompiled input objects from testcases..
Wed, Oct 4, 2:48 AM
grimar added inline comments to D38407: [ELF] - Do --hash-style=both by default..
Wed, Oct 4, 2:29 AM
grimar updated the diff for D38407: [ELF] - Do --hash-style=both by default..
  • Addressed review comments.
Wed, Oct 4, 2:29 AM
grimar committed rL314884: [ELF] - Update after LLVM r314883 change. NFC..
[ELF] - Update after LLVM r314883 change. NFC.
Wed, Oct 4, 1:52 AM
grimar committed rL314883: [MC] - Don't assert when non-english characters are used..
[MC] - Don't assert when non-english characters are used.
Wed, Oct 4, 1:52 AM
grimar closed D38461: [MC] - Don't assert when non-english characters are used. by committing rL314883: [MC] - Don't assert when non-english characters are used..
Wed, Oct 4, 1:51 AM
grimar updated the summary of D38461: [MC] - Don't assert when non-english characters are used..
Wed, Oct 4, 1:50 AM

Tue, Oct 3

grimar added a comment to D38491: [ELF] Decompress debug info sections early.

I have no futher comments, this LGTM, please wait for other reviewers approval. Thanks !

Tue, Oct 3, 9:16 AM
grimar added inline comments to D38491: [ELF] Decompress debug info sections early.
Tue, Oct 3, 8:56 AM
grimar updated the diff for D38407: [ELF] - Do --hash-style=both by default..
  • Use --hash-style=sysv to simulate previous behavior in testcases instead of modifying values checked.
Tue, Oct 3, 8:31 AM
grimar created D38505: [ELF] - Get rid of precompiled input objects from testcases..
Tue, Oct 3, 7:50 AM
grimar added a comment to D38393: [ELF] - Stop removing sections in removeUnusedSyntheticSections()..
In D38393#886125, @ruiu wrote:

On second thought, I think removeUnusedSyntehticSections is a source of confusion. This mechanism seems too hacky, and that's why you need to write many pieces of tricky code. I feels we could design a better mechanism. Can you visit this again to design something new?

Tue, Oct 3, 5:33 AM
grimar updated the diff for D38461: [MC] - Don't assert when non-english characters are used..
  • Added helpers to StringExtras as was suggested.
Tue, Oct 3, 5:13 AM
grimar added inline comments to D38491: [ELF] Decompress debug info sections early.
Tue, Oct 3, 2:07 AM