probinson (Paul Robinson)
User

Projects

User Details

User Since
May 9 2013, 11:10 AM (236 w, 2 d)

Recent Activity

Yesterday

probinson added a comment to D40199: llvm-dwarfdump --verify is incorrectly saying all DW_AT_location attributes with locations lists are invalid..

The FORM that's appropriate for DW_AT_location to use as a reference varies with the DWARF version. In v2 and v3, it's FORM_data4 or FORM_data8, in v4 it's FORM_sec_offset. I haven't looked at the verifier much but it seems like using the correct FORM is something that ought to be verifiable.

Fri, Nov 17, 3:36 PM · debug-info
probinson added a project to D40199: llvm-dwarfdump --verify is incorrectly saying all DW_AT_location attributes with locations lists are invalid.: debug-info.
Fri, Nov 17, 3:16 PM · debug-info
probinson added inline comments to D40199: llvm-dwarfdump --verify is incorrectly saying all DW_AT_location attributes with locations lists are invalid..
Fri, Nov 17, 3:15 PM · debug-info
probinson created D40200: [DWARF] Fix handling of extended line-number opcodes.
Fri, Nov 17, 2:44 PM · debug-info

Thu, Nov 16

probinson added a comment to D40114: [DebugInfo] Dump a .debug_line section, with line-number program, without a compile unit..

Thanks for accepting, Jonas! FWIW accepting by itself does not cause Phab to send email, there still needs to be a comment.
And committing this has to wait on D39854 which still needs review.

Thu, Nov 16, 11:07 AM · debug-info

Wed, Nov 15

probinson created D40114: [DebugInfo] Dump a .debug_line section, with line-number program, without a compile unit..
Wed, Nov 15, 5:08 PM · debug-info
probinson added a comment to D38302: [X86][MC] Fixed crash when assembling a file that contains section with 64-bit alignment (PR34726).

@rafael this looks pretty benign to me, but MC is not my area. What do you think?

Wed, Nov 15, 3:44 PM
probinson added a member for debug-info: probinson.
Wed, Nov 15, 10:03 AM

Tue, Nov 14

probinson updated the diff for D39854: [DWARFv5] Support FORM_strp in .debug_line.dwo.

Use address-size zero for the DWARFDataExtractor, in the .dwo section.
This required relaxing an assertion in the line-table parser. If there's no line-number program, the address size isn't needed, and if there is one, attempting to retrieve an address will assert farther down.

Tue, Nov 14, 2:49 PM · debug-info

Mon, Nov 13

probinson added a comment to D39950: [DebugInfo] Stable sort symbols to remove non-deterministic ordering.

The results suggest a different problem. Emitting the aranges section should happen after the "kittens" and "rainbows" symbols are emitted. If the symbols were emitted, they should have nonzero and unique order. I don't see how shuffling the list can result in a different result for the aranges section.
Although MCStreamer::AssignFragment looks like it might be a victim of unspecified order-of-evaluation, still I would think the assigned orders remain nonzero and unique per symbol, and that's all that matters here.

Mon, Nov 13, 7:00 PM · debug-info
probinson added a comment to D39950: [DebugInfo] Stable sort symbols to remove non-deterministic ordering.

So there can be multiple symbols with the same offset, in this list? This change implies that the insertion order in the list is important and worth preserving. Why do we know that's true?

Mon, Nov 13, 10:35 AM · debug-info

Fri, Nov 10

probinson added a comment to D39854: [DWARFv5] Support FORM_strp in .debug_line.dwo.

Oh, head-smack! You don't need a unit to know the address size; the line table has always had that info, although indirectly.
DW_LNE_set_address (the only opcode that specifies an actual address) is an extended opcode, which has a ULEB telling you how long the operands are. Which is, one byte for the opcode, and.... the size of the address. Doh!

Fri, Nov 10, 5:09 PM · debug-info
probinson added a comment to D39896: Remove last Host usage from ArchSpec.

Drive by comment:

Fri, Nov 10, 2:19 PM
probinson added inline comments to D39854: [DWARFv5] Support FORM_strp in .debug_line.dwo.
Fri, Nov 10, 10:30 AM · debug-info
probinson added inline comments to D39854: [DWARFv5] Support FORM_strp in .debug_line.dwo.
Fri, Nov 10, 10:04 AM · debug-info

Thu, Nov 9

probinson added a comment to D39854: [DWARFv5] Support FORM_strp in .debug_line.dwo.

(Moving the why-do-we-need-the-unit discussion out of the inline comments.)

Thu, Nov 9, 2:57 PM · debug-info
probinson added inline comments to D39854: [DWARFv5] Support FORM_strp in .debug_line.dwo.
Thu, Nov 9, 1:34 PM · debug-info
probinson updated the diff for D39854: [DWARFv5] Support FORM_strp in .debug_line.dwo.

Address @dblaikie comments.

Thu, Nov 9, 12:53 PM · debug-info
probinson added inline comments to D39854: [DWARFv5] Support FORM_strp in .debug_line.dwo.
Thu, Nov 9, 12:02 PM · debug-info
probinson committed rL317827: Fix out-of-order stepping behavior in programs with hoisted constants..
Fix out-of-order stepping behavior in programs with hoisted constants.
Thu, Nov 9, 12:01 PM
probinson closed D38088: Fix out-of-order stepping behavior in programs with hoisted constants. by committing rL317827: Fix out-of-order stepping behavior in programs with hoisted constants..
Thu, Nov 9, 12:01 PM
probinson created D39854: [DWARFv5] Support FORM_strp in .debug_line.dwo.
Thu, Nov 9, 11:17 AM · debug-info

Tue, Nov 7

probinson committed rL317634: Reapply r317609 with a simpler sed script, thanks to Justin Bogner!.
Reapply r317609 with a simpler sed script, thanks to Justin Bogner!
Tue, Nov 7, 3:18 PM
probinson committed rL317628: Revert r317609, test fails on one bot.
Revert r317609, test fails on one bot
Tue, Nov 7, 2:39 PM
probinson committed rL317612: Convert a dwarfdump test from checked-in binary to assembler source..
Convert a dwarfdump test from checked-in binary to assembler source.
Tue, Nov 7, 12:36 PM
probinson committed rL317609: [DWARFv5] Add new test for previous commit..
[DWARFv5] Add new test for previous commit.
Tue, Nov 7, 12:13 PM
probinson committed rL317607: [DWARFv5] Support DW_FORM_strp in the .debug_line header..
[DWARFv5] Support DW_FORM_strp in the .debug_line header.
Tue, Nov 7, 11:58 AM
probinson closed D33155: [DWARFv5] Support FORM_strp in the line table header by committing rL317607: [DWARFv5] Support DW_FORM_strp in the .debug_line header..
Tue, Nov 7, 11:58 AM
probinson added a comment to D33155: [DWARFv5] Support FORM_strp in the line table header.

Why can't the other test be ELF-only assembler though?

Tue, Nov 7, 10:50 AM
probinson added a comment to D33155: [DWARFv5] Support FORM_strp in the line table header.

I left the existing test (dwarfdump-header.s) as a checked-in binary, because it tests both normal and split headers, and I don't know what MachO section names to use for the .dwo sections. My intuition is that MachO will never bother supporting split DWARF so maybe it's fine to leave it as is. If somebody does want me to convert the test, I'll need a list of section names to use, and I would do it as a follow-up. Just thought I'd put that out there.

Tue, Nov 7, 10:23 AM

Mon, Nov 6

probinson accepted D39629: Add a -D flag to FileCheck to define variables.

It does seem like this could be very useful in a sequence of RUN commands, where previously we'd have to add a bunch of run-specific checks.
LGTM.

Mon, Nov 6, 12:57 PM
probinson added inline comments to D39629: Add a -D flag to FileCheck to define variables.
Mon, Nov 6, 12:47 PM

Fri, Nov 3

probinson added a comment to D39605: [debuginfo-tests] Decouple debuginfo-tests from the clang source tree..

At Sony we build our own unified tree by pasting together several subtrees, including debuginfo-tests, and moving one will cause a bit of disruption. I remember the idea of moving debuginfo-tests has come up previously, and in principle I have no problem with it, but a more explicit announcement on llvm-dev/cfe-dev would be appropriate.

FWIW, debuginfo-tests is not actually checked out by default. You have to explicitly clone it at a location of your choosing. Before this patch, that location has to be clang/test, and after this patch, it has to be llvm/projects. How are your trees subtrees pasted together, is it something other than a mono-repo layout?

Fri, Nov 3, 2:50 PM
probinson added a comment to D39605: [debuginfo-tests] Decouple debuginfo-tests from the clang source tree..

At Sony we build our own unified tree by pasting together several subtrees, including debuginfo-tests, and moving one will cause a bit of disruption. I remember the idea of moving debuginfo-tests has come up previously, and in principle I have no problem with it, but a more explicit announcement on llvm-dev/cfe-dev would be appropriate.

Fri, Nov 3, 2:29 PM

Wed, Nov 1

probinson added a comment to D39239: [AST] Incorrectly qualified unscoped enumeration as template actual parameter..

Have you tried the GDB suite yet? If it has no problems, and we can make LLDB happy, I'm okay with it.

Wed, Nov 1, 10:07 AM

Tue, Oct 31

probinson updated the diff for D33155: [DWARFv5] Support FORM_strp in the line table header.

Make the new test work for both ELF and Mach-O.
Simplify LLD change to never pass a Unit, as there's no test for it yet.

Tue, Oct 31, 4:54 PM

Mon, Oct 30

probinson added inline comments to D33155: [DWARFv5] Support FORM_strp in the line table header.
Mon, Oct 30, 2:33 PM

Fri, Oct 27

probinson added inline comments to D33155: [DWARFv5] Support FORM_strp in the line table header.
Fri, Oct 27, 11:37 AM
probinson updated the diff for D33155: [DWARFv5] Support FORM_strp in the line table header.

By passing a DWARFUnit into the line-table parser (if we have one), we allow it to handle indirect string forms, which require finding other sections. Previously the line-table parser couldn't do that (and didn't need to, prior to v5).
By passing DWARFFormParams to DWARFFormValue::extractValue() separately from the DWARFUnit, we allow interpreting forms according to the line-table's version/format, which could be different from the Unit's version/format.

Fri, Oct 27, 10:40 AM

Thu, Oct 26

probinson added a comment to D39215: Default to using in-tree clang for building test executables.

Has this gone in? I'm wondering because I starting playing with the monorepo, ran cmake with -DLLDB_TEST_COMPILER=$PWD/bin/clang, and today's test failure seems to be trying to build the test program with the system compiler (gcc) rather than my copy of clang. But it looks like you're deprecating LLDB_TEST_COMPILER?

Thu, Oct 26, 4:40 PM

Tue, Oct 24

probinson added a comment to D39239: [AST] Incorrectly qualified unscoped enumeration as template actual parameter..

Have you tried this change against the GDB and LLDB test suites? If they have failures then we should think about whether those tests are over-specific or whether we should put this under a tuning option.

Tue, Oct 24, 1:27 PM

Mon, Oct 23

probinson added a comment to D39069: CodeGen: Fix missing debug loc due to alloca.

Anytime the code between saveIP() and restoreIP() could set the current debug location, it needs to be saved/restored along with the insertion point. I have to say the problem is not obvious to me here, so maybe saveIP/restoreIP should be changed (or eliminated in favor of always using InsertPointGuard). I'm not seeing a lot of places where saveIP/restoreIP are used.

Mon, Oct 23, 4:40 PM

Oct 12 2017

probinson added inline comments to D38654: Parse DWARF information to reduce false positives..
Oct 12 2017, 5:08 PM

Oct 6 2017

probinson committed rL315078: For Windows, allow .exe extension in a test..
For Windows, allow .exe extension in a test.
Oct 6 2017, 10:14 AM

Sep 29 2017

probinson abandoned D35715: Preserve typedef names in debug info for template type parameters.

Abandoning. This change is irrelevant to the SCE debugger, and while I believe it could be made more complete and better reflect the original source (which is what the DWARF spec says names should be), I do not have time to pursue it.

Sep 29 2017, 3:58 PM
probinson committed rL314558: [PS4] Tidy up some debug-tuning v. triple decision-making..
[PS4] Tidy up some debug-tuning v. triple decision-making.
Sep 29 2017, 2:26 PM

Sep 28 2017

probinson committed rL314444: [DWARF] Allow forward declarations of a class template instantiation.
[DWARF] Allow forward declarations of a class template instantiation
Sep 28 2017, 11:39 AM
probinson closed D14358: DWARF's forward decl of a template should have template parameters. by committing rL314444: [DWARF] Allow forward declarations of a class template instantiation.
Sep 28 2017, 11:38 AM
probinson added a comment to D14358: DWARF's forward decl of a template should have template parameters..

I think I will go ahead and commit this; it doesn't change the status quo for CodeView and if something is warranted we can do that in a follow-up.

Sep 28 2017, 10:56 AM

Sep 27 2017

probinson added inline comments to D38282: llvm-dwarfdump: implement --find=<name>.
Sep 27 2017, 3:28 PM
probinson added a reviewer for D14358: DWARF's forward decl of a template should have template parameters.: rnk.

+rnk for the CodeView question.

Sep 27 2017, 1:58 PM
probinson updated the diff for D14358: DWARF's forward decl of a template should have template parameters..

command-line option is cc1 not driver
internal flag moved from LangOpts to CodeGenOpts and renamed
simplified test

Sep 27 2017, 12:36 PM
probinson added a comment to D14358: DWARF's forward decl of a template should have template parameters..

I would prefer to eliminate the <params> from the instance name as well, because our debugger reconstructs a name more to its liking from the parameter children. However, IIUC the name with params is used for deduplication in LTO, so that is probably not such a good idea. :-)

Though you have this out of tree? How do you cope with LTO there?

Sep 27 2017, 11:50 AM
probinson added a comment to D14358: DWARF's forward decl of a template should have template parameters..

Does this have to be exposed through the driver or could this be a cc1 option only?

Sep 27 2017, 11:10 AM

Sep 26 2017

probinson updated the diff for D14358: DWARF's forward decl of a template should have template parameters..

Add a command-line flag to control emitting the template parameter children. Default to on for SCE debugger tuning.
I am not super excited about my choice of option name or the help text; alternate suggestions welcome.

Sep 26 2017, 3:30 PM

Sep 21 2017

probinson accepted D38088: Fix out-of-order stepping behavior in programs with hoisted constants..

Well then! LGTM.

Sep 21 2017, 4:56 PM
probinson added a comment to D38088: Fix out-of-order stepping behavior in programs with hoisted constants..

This came out of our sample-PGO work, wasn't sure it would be kosher for me to approve a patch from the guy in the next cube.
It's completely in line with the other similar patches that Andrea and Dehao have been doing. So if Craig is happy, I'm happy.

Sep 21 2017, 4:39 PM

Sep 20 2017

probinson added inline comments to D37971: [dwarfdump] Add verbose output for .debug-line section.
Sep 20 2017, 4:47 PM

Sep 15 2017

probinson added a comment to D37157: Fix Bug 30978 by emitting cv file checksums..

A few minor commentary points; nothing huge. I've been looking at this mainly with the thought of having to do something similar for DWARF 5 eventually, but it seems likely there won't be a lot to leverage as CodeView and DWARF appear to associate the hashes with files pretty differently. Oh well.

Sep 15 2017, 10:53 AM

Sep 14 2017

probinson added inline comments to D37157: Fix Bug 30978 by emitting cv file checksums..
Sep 14 2017, 4:58 PM
probinson added a comment to D37852: [dwarfdump] Make .eh_frame an alias for .debug_frame.

My impression is that .eh_frame and .debug_frame are interchangeable and in fact you get only one or the other in any given compilation. So, if you ask for .debug_frame and the object has .eh_frame, I think we should still dump it; and vice versa.

Sep 14 2017, 1:33 PM

Sep 13 2017

probinson added inline comments to D37771: llvm-dwarfdump: automatically dump both regular and .dwo variant of sections.
Sep 13 2017, 2:52 PM
probinson added inline comments to D37771: llvm-dwarfdump: automatically dump both regular and .dwo variant of sections.
Sep 13 2017, 2:36 PM
probinson added inline comments to D37771: llvm-dwarfdump: automatically dump both regular and .dwo variant of sections.
Sep 13 2017, 11:55 AM
probinson added inline comments to D37771: llvm-dwarfdump: automatically dump both regular and .dwo variant of sections.
Sep 13 2017, 11:44 AM
probinson added a comment to D37756: [lit] Force site configs to be run before source-tree configs.

If you use cmake to generate Visual Studio projects on Windows, do they use lit to run the tests? If so, do they run with this patch correctly? I have no opinion one way or the other but it's an environment nobody seems to have thought of before (including me).

I just tested running check-llvm from inside Visual Studio using a Visual Studio generated CMake project and it worked fine. I didn't run the rest of them, since if one works they all should.

Sep 13 2017, 8:18 AM

Sep 12 2017

probinson added a comment to D37771: llvm-dwarfdump: automatically dump both regular and .dwo variant of sections.

UI is tricky. I think I lean in the direction that if the section doesn't exist, don't print anything, even the "contents:" line.
This is certainly how I want --all to behave, and obviously it's easiest and self-consistent if that's how the tool always behaves. So for example --debug-types when there is no .debug_types section should print nothing.
I could see an argument that if I explicitly asked for a particular section and it doesn't exist, I should get the "contents:" line with nothing after it. But that's really not how I want --all to behave.

Sep 12 2017, 3:38 PM
probinson added inline comments to D37768: [IR] Add llvm.dbg.addr, a control-dependent version of llvm.dbg.declare.
Sep 12 2017, 3:26 PM
probinson added a comment to D37756: [lit] Force site configs to be run before source-tree configs.

If you use cmake to generate Visual Studio projects on Windows, do they use lit to run the tests? If so, do they run with this patch correctly? I have no opinion one way or the other but it's an environment nobody seems to have thought of before (including me).

Sep 12 2017, 3:06 PM
probinson added inline comments to D37771: llvm-dwarfdump: automatically dump both regular and .dwo variant of sections.
Sep 12 2017, 2:41 PM

Sep 11 2017

probinson added a comment to D37714: llvm-dwarfdump: Replace -debug-dump=sect option with individual options..

What about supporting both options, but --debug-info also implies --debug-info-dwo?

Sep 11 2017, 4:16 PM
probinson added a comment to D37714: llvm-dwarfdump: Replace -debug-dump=sect option with individual options..

I would kind of like one option to dump whichever of .debug_info or .debug_info.dwo (etc) there is, so I don't have to know up front which section to be asking for. Would also be more user-friendly to people who kind of know DWARF but not at that level of detail.

Sep 11 2017, 4:07 PM
probinson accepted D37625: [DWARF] Incorrect prologue end line record..

I see that findPrologueEndLoc() is also checking the FrameSetup flag, so at least this makes the handling consistent, and relying on FrameSetup actually seems cleaner than relying on source locations. If you have time to investigate why the source locations seem wrong, that would be great.
LGTM after you fix the comment Adrian pointed out.

Sep 11 2017, 10:51 AM

Sep 8 2017

probinson added a comment to D37625: [DWARF] Incorrect prologue end line record..

I think the argument (and the test) would be stronger if it was based on the instruction sequence rather than offsets. For example if you generate asm and then show prologue_end is immediately before the call to get_arg.
Also we generally like to trim the IR source by removing unnecessary attribute sets.

Sep 8 2017, 11:27 AM
probinson added a comment to D37604: Disable debuginfo-tests for non-native configurations.

This seems reasonable to me, thanks!
When you commit this, could you please double-check that the tests are still running on the green dragon builders? I'll also keep an eye on them.

Sep 8 2017, 10:54 AM
probinson closed D37604: Disable debuginfo-tests for non-native configurations.

r312803. Forgot to put the tag in the commit message....

Sep 8 2017, 10:13 AM
probinson committed rL312803: Restrict debuginfo-tests to native configurations..
Restrict debuginfo-tests to native configurations.
Sep 8 2017, 10:12 AM

Sep 7 2017

probinson added inline comments to D37602: Properly hook debuginfo-tests up to lit and CMake.
Sep 7 2017, 5:08 PM
probinson added a comment to D37604: Disable debuginfo-tests for non-native configurations.

The new file goes in the debuginfo-tests directory. It's a separate project so that's probably not obvious from the diff.

Sep 7 2017, 4:49 PM
probinson created D37604: Disable debuginfo-tests for non-native configurations.
Sep 7 2017, 4:48 PM
probinson committed rL312751: [DWARF] Line 0 should not have a discriminator..
[DWARF] Line 0 should not have a discriminator.
Sep 7 2017, 3:19 PM
probinson closed D37364: [DWARF] Line 0 should not have a discriminator by committing rL312751: [DWARF] Line 0 should not have a discriminator..
Sep 7 2017, 3:19 PM
probinson added a comment to D37364: [DWARF] Line 0 should not have a discriminator.

Seems OK to me

For a minor perk - would it be possible to test the line table length? Since the benefit (smaller line table) of this change isn't readily apparent in the high level line table printing that dwarfdump does. I know it'd make the test a bit more brittle.

Sep 7 2017, 3:04 PM
probinson updated the diff for D37364: [DWARF] Line 0 should not have a discriminator.

Seriously simplified test (derived from DebugInfo/X86/discriminator2.ll, as suggested by dblaikie).

Sep 7 2017, 2:28 PM

Sep 5 2017

probinson updated subscribers of D35532: [CMake] Update GetSVN.cmake to handle Repo.

The patch seems trivial enough but I do not have a good idea of the subtle differences in the various environments.
I have subscribed a couple of people who probably can tell whether this change makes sense.

Sep 5 2017, 5:02 PM
probinson added a comment to D37364: [DWARF] Line 0 should not have a discriminator.

An MIR test that set up the situation more directly might be better, although I am not super confident in my ability to hand-reduce an MIR input.

Sep 5 2017, 10:35 AM
probinson added a comment to D37364: [DWARF] Line 0 should not have a discriminator.

The change looks good to me. But the test is quite large and hard to understand. I don't think it's necessary to have the test generated from a source file, just forge a test with discriminator attached to line 0, and then checks if discriminators is omitted in the generated code. I guess it should suffice with < 20 LOC

Sep 5 2017, 10:22 AM

Aug 31 2017

probinson created D37364: [DWARF] Line 0 should not have a discriminator.
Aug 31 2017, 5:46 PM
probinson added a comment to D37038: Replace temp MD nodes with unique/distinct before cloning.

I now think the tactic to pursue is making sure the containing DICompositeType(s) for the method have complete descriptions, and RAUW the temporary nodes, prior to calling CloneFunction. Actually wading around in the MDNodes, Types, and Decls to accomplish that result promises to be "interesting" and "educational" (advice/suggestions welcome).

Aug 31 2017, 11:34 AM

Aug 29 2017

probinson added a comment to D37038: Replace temp MD nodes with unique/distinct before cloning.

This may have gotten lost earlier: Would it be possible to instruct CloneFunction to not clone any temporary MDNodes via one of the flags that are passed to the ValueMapper?

Aug 29 2017, 5:34 PM
probinson added a comment to D37038: Replace temp MD nodes with unique/distinct before cloning.

This may have gotten lost earlier: Would it be possible to instruct CloneFunction to not clone any temporary MDNodes via one of the flags that are passed to the ValueMapper?

Aug 29 2017, 11:04 AM

Aug 28 2017

probinson added a comment to D37038: Replace temp MD nodes with unique/distinct before cloning.

Dumping the whole module, the temporary DICompositeType node for CBdVfsImpl is used both as the scope node for the DISubprogram, and also as the baseType of a DIDerivedType which is a pointer type in the type list for the subprogram.

Aug 28 2017, 5:43 PM
probinson added a comment to D37038: Replace temp MD nodes with unique/distinct before cloning.

finalizeSubprogram() retrieves the variables from the subprogram and handles them; what is it missing?

Aug 28 2017, 4:56 PM
probinson added a comment to D37038: Replace temp MD nodes with unique/distinct before cloning.

Trying to understand the broader context here, I looked back through the list of revisions mentioned in PR33930 to see if that helped.

Aug 28 2017, 3:50 PM

Aug 25 2017

probinson added a comment to D36501: Add flag to request Clang is ABI-compatible with older versions of itself.

Sorry it took so long to wrap my head around this, but it has been about 20 years since I've had to accommodate an intentional ABI breakage, and that's actually what you want to do. Remembering that case, I have a model for this one, and I understand what you are doing now.

Aug 25 2017, 5:44 PM
probinson added inline comments to D37038: Replace temp MD nodes with unique/distinct before cloning.
Aug 25 2017, 4:59 PM
probinson added inline comments to D36501: Add flag to request Clang is ABI-compatible with older versions of itself.
Aug 25 2017, 4:19 PM
probinson added a comment to D36993: [llvm-dwarfdump] Print type names in DW_AT_type DIEs.

I'm happy, but Adrian should have a chance to comment too.

Aug 25 2017, 2:35 PM
probinson added a comment to D6379: Only warn about DWARF2 supporting one section per compilation unit for code sections.
In D6379#852974, @dim wrote:

Actually I see that we also get this warning on our .section ".note.GNU-stack","",@progbits directives, which are in most of our hand-written .S files. Since the type of that section is progbits, this fix won't work for it.

However, interestingly, clang also outputs these same .note.GNU-stack directives in .s output for compiled files, but it does *not* warn on those, even when using -gdwarf-2. So how is it avoiding them in that case?

Aug 25 2017, 2:25 PM
probinson added inline comments to D36501: Add flag to request Clang is ABI-compatible with older versions of itself.
Aug 25 2017, 2:18 PM