sbc100 (Sam Clegg)
User

Projects

User does not belong to any projects.

User Details

User Since
Sep 16 2016, 10:22 AM (108 w, 4 d)

Recent Activity

Fri, Oct 12

sbc100 updated the summary of D53240: [WebAssembly] WebAssemblyLowerEmscriptenEHSjLj: use getter/setter for accessing tempRet0.
Fri, Oct 12, 7:47 PM
sbc100 created D53240: [WebAssembly] WebAssemblyLowerEmscriptenEHSjLj: use getter/setter for accessing tempRet0.
Fri, Oct 12, 7:31 PM

Tue, Oct 9

sbc100 added a comment to D53054: Adapt OptTable::PrintHelp change in D51009.

Since no tests needed updating I guess we not testing much about the output of --help? Should we add some tests?

Tue, Oct 9, 4:52 PM
sbc100 added a comment to D52748: [WebAssembly] LSDA info generation.

Is that information what's used/needed to allow the linker to create a separate wasm segment per global?

Tue, Oct 9, 3:10 PM
sbc100 added a comment to D52748: [WebAssembly] LSDA info generation.

Sorry for not replying earlier. Yes, we should generate and preserve size information for data symbols if possible.

Tue, Oct 9, 2:12 PM

Fri, Oct 5

sbc100 added a comment to D52899: [TTI] Check that lowered type is floating point before calling isFabsFree.

Thanks for the tip, setting a negative threshold helped, and I was able to reduce the test case too.

Fri, Oct 5, 2:43 PM
sbc100 updated the diff for D52899: [TTI] Check that lowered type is floating point before calling isFabsFree.
  • add filecheck and minimize test
Fri, Oct 5, 2:43 PM

Thu, Oct 4

sbc100 added a comment to D52899: [TTI] Check that lowered type is floating point before calling isFabsFree.

I'm not sure it actually does the vectorization in the end. The bit-code that some out is basically identical.

Thu, Oct 4, 5:51 PM
sbc100 added inline comments to D52899: [TTI] Check that lowered type is floating point before calling isFabsFree.
Thu, Oct 4, 5:36 PM
sbc100 added a comment to D52914: [WebAssembly] Added test for inline assembly roundtrip..

I kind of prefer the test to be as minimal as possible even if derived from c source, but perhaps thats overly pedantic.

Thu, Oct 4, 5:06 PM
sbc100 added a comment to D52899: [TTI] Check that lowered type is floating point before calling isFabsFree.

This is the smallest test I could create. I would love it to be smaller but it seems hard to tickle this particular path.

Thu, Oct 4, 2:38 PM
sbc100 updated the diff for D52899: [TTI] Check that lowered type is floating point before calling isFabsFree.
  • add test
Thu, Oct 4, 2:36 PM
sbc100 added a comment to D52899: [TTI] Check that lowered type is floating point before calling isFabsFree.

I can trigger the same issue by using __float128 types and --target=i686-linux (where fp128 is emulated in software).

Thu, Oct 4, 12:51 PM
sbc100 retitled D52899: [TTI] Check that lowered type is floating point before calling isFabsFree from [SLPVectorizer] Check that lowered type is floating point before calling isFabsFree to [TTI] Check that lowered type is floating point before calling isFabsFree.
Thu, Oct 4, 11:58 AM
sbc100 created D52899: [TTI] Check that lowered type is floating point before calling isFabsFree.
Thu, Oct 4, 11:56 AM
sbc100 added inline comments to D52894: Update CallSite docs and add a new function (NFC).
Thu, Oct 4, 10:30 AM
sbc100 added a comment to D52837: [WebAssembly] Ignore DBG_VALUE in WebAssemblyCFGStackify pass when looking for block start.

Would bugpoint help here? Its specifically designed to minimize test case I think although I don't have a lot of experience with it myself.

Thu, Oct 4, 9:37 AM

Wed, Oct 3

sbc100 accepted D52850: [WebAssembly] Add WebAssembly to LLVM_ALL_TARGETS.
Wed, Oct 3, 4:10 PM
sbc100 added a comment to D52837: [WebAssembly] Ignore DBG_VALUE in WebAssemblyCFGStackify pass when looking for block start.

Is there any way to we can make this test case more focused?

Wed, Oct 3, 3:35 PM

Tue, Oct 2

sbc100 updated the diff for D49208: [WebAssembly] Update WebAssemblyLowerEmscriptenEHSjLj to support separate compilation..

Rebase

Tue, Oct 2, 2:28 PM
sbc100 added a comment to D52741: [AMDGPU] Add an AMDGPU pass to promote bitcast calls.

Your assessment is about right yes. We assembly supports indirect calls but the function signatures have to match in the webassembly type system. In fact the executable won't even validate (won't load) if it contains such invalid calls. So it sounds like the goals are not the same.

Tue, Oct 2, 10:15 AM

Mon, Oct 1

sbc100 added inline comments to D52748: [WebAssembly] LSDA info generation.
Mon, Oct 1, 3:04 PM

Fri, Sep 28

sbc100 added a comment to D49208: [WebAssembly] Update WebAssemblyLowerEmscriptenEHSjLj to support separate compilation..

ptal

Fri, Sep 28, 3:39 PM
sbc100 added inline comments to D49208: [WebAssembly] Update WebAssemblyLowerEmscriptenEHSjLj to support separate compilation..
Fri, Sep 28, 3:15 PM
sbc100 updated the diff for D49208: [WebAssembly] Update WebAssemblyLowerEmscriptenEHSjLj to support separate compilation..
  • feedback
Fri, Sep 28, 3:15 PM
sbc100 updated the diff for D49208: [WebAssembly] Update WebAssemblyLowerEmscriptenEHSjLj to support separate compilation..
  • feedback
Fri, Sep 28, 3:13 PM
sbc100 updated the diff for D49208: [WebAssembly] Update WebAssemblyLowerEmscriptenEHSjLj to support separate compilation..

Rebase and fix tests

Fri, Sep 28, 3:12 PM
sbc100 reclaimed D49208: [WebAssembly] Update WebAssemblyLowerEmscriptenEHSjLj to support separate compilation..

Re-opening since we need these symbols to be available pre-LTO.. i.e. before codegen.

Fri, Sep 28, 2:40 PM

Thu, Sep 27

sbc100 added a comment to D50721: [WebAssembly] Preserve function signatures during LTO.

ping..

Thu, Sep 27, 6:53 PM
sbc100 accepted D52622: [WebAssembly] Refactor use of signatures.

LGTM with nits

Thu, Sep 27, 12:54 PM
sbc100 added a comment to D52580: Refactor WasmSignature and use it for MCSymbolWasm.

Generally looks good. Yay for unblocking us from leaving experimental

Thu, Sep 27, 12:47 PM

Wed, Sep 26

sbc100 updated the summary of D52587: [WebAssembly] Add --[no]-export-dynamic to replace --export-default.
Wed, Sep 26, 6:50 PM
sbc100 added reviewers for D52587: [WebAssembly] Add --[no]-export-dynamic to replace --export-default: dschuff, ruiu.
Wed, Sep 26, 6:48 PM
sbc100 created D52587: [WebAssembly] Add --[no]-export-dynamic to replace --export-default.
Wed, Sep 26, 6:47 PM
sbc100 added a comment to D52048: [WebAssembly] Update docs.

Looks like I broke the doc builder: http://lab.llvm.org:8011/builders/lld-sphinx-docs/builds/26336

Wed, Sep 26, 5:34 PM
sbc100 added a comment to D49034: [WebAssembly] Move/clone DBG_VALUE during WebAssemblyRegStackify pass.

We we revert this or get it fixed? Its currently breaking the wasm waterfall: https://wasm-stat.us/console

Wed, Sep 26, 12:31 PM · debug-info

Tue, Sep 25

sbc100 added a comment to D52003: [WebAssembly] Add --export-default/--no-export-default options.

OK to land?

Tue, Sep 25, 2:05 PM

Mon, Sep 24

sbc100 accepted D49446: [WebAssembly] Move .debug_line section address of dead function outside section range.
Mon, Sep 24, 4:21 PM

Fri, Sep 21

sbc100 added a comment to rL342707: [WebAssembly] Revert r342701, "Add WebAssembly to LLVM_ALL_TARGTS.".

Do you have a link the broken builder? I'm happy to take a look at fixing this.

Fri, Sep 21, 8:43 AM

Thu, Sep 20

sbc100 added a comment to D43211: [WebAssembly] promote from experimental to normal target.

Woah! This a big step people! Congratulations.

Thu, Sep 20, 5:56 PM
sbc100 added a comment to D43675: [WebAssembly] Rename imported/exported memory symbol to __linear_memory.

I've been persuaded I think that we should rename this, and let emscripten coerce it as it sees fit. However I think maybe we should have common scheme for all these synthetic/internal exports. How about: __wasm_memory, __wasm_table, __wasm_data_start, etc?

Thu, Sep 20, 5:19 PM

Sep 13 2018

sbc100 added a comment to D52003: [WebAssembly] Add --export-default/--no-export-default options.

Hmm, I see what you mean. I thing is that --export-default is the default, so we would need to detect weather it was set from the command or not which would complicate the code a little. Also, all is a superset of default I'm not sure its necessary.

Sep 13 2018, 5:23 PM
sbc100 added a comment to D50976: [WebAssembly] Update Config member to match command line option. NFC..

ping

Sep 13 2018, 5:19 PM
sbc100 created D52054: Add missing MC dependency to llvm-exegesis/CMakeLists.txt.
Sep 13 2018, 1:03 PM
sbc100 added a reviewer for D52054: Add missing MC dependency to llvm-exegesis/CMakeLists.txt: DiamondLovesYou.
Sep 13 2018, 1:03 PM
sbc100 added reviewers for D52048: [WebAssembly] Update docs: ruiu, jgravelle-google.
Sep 13 2018, 11:02 AM
sbc100 updated the diff for D52048: [WebAssembly] Update docs.

Update README

Sep 13 2018, 11:02 AM
sbc100 created D52048: [WebAssembly] Update docs.
Sep 13 2018, 10:55 AM
sbc100 added a reviewer for D52003: [WebAssembly] Add --export-default/--no-export-default options: jgravelle-google.
Sep 13 2018, 10:50 AM
sbc100 updated the diff for D51562: [WebAssembly] Fix signature of `main` in FixFunctionBitcasts.

-rebase

Sep 13 2018, 10:09 AM

Sep 12 2018

sbc100 added a reviewer for D52003: [WebAssembly] Add --export-default/--no-export-default options: ruiu.
Sep 12 2018, 2:10 PM
sbc100 created D52003: [WebAssembly] Add --export-default/--no-export-default options.
Sep 12 2018, 2:09 PM
sbc100 added a comment to D51991: [WebAssembly] Make tied inline asm operands work again.

I don't know this code myself, but if I'm right in assuming this fixes the current waterfall failures then please commit and we can followup is others have comments.

Sep 12 2018, 9:29 AM

Sep 5 2018

sbc100 added a comment to D51562: [WebAssembly] Fix signature of `main` in FixFunctionBitcasts.

ping

Sep 5 2018, 3:10 PM

Aug 31 2018

sbc100 updated the diff for D51562: [WebAssembly] Fix signature of `main` in FixFunctionBitcasts.

fix

Aug 31 2018, 5:09 PM
sbc100 added a comment to D51562: [WebAssembly] Fix signature of `main` in FixFunctionBitcasts.

I'm not sure why the logic of this flag is set to true by default and then we check for "!TemporaryWorkarounds".

Aug 31 2018, 5:08 PM
sbc100 added reviewers for D51562: [WebAssembly] Fix signature of `main` in FixFunctionBitcasts: jgravelle-google, sunfish.
Aug 31 2018, 5:07 PM
sbc100 updated the diff for D51562: [WebAssembly] Fix signature of `main` in FixFunctionBitcasts.

cleanup

Aug 31 2018, 5:07 PM
sbc100 created D51562: [WebAssembly] Fix signature of `main` in FixFunctionBitcasts.
Aug 31 2018, 5:02 PM

Aug 29 2018

sbc100 added a reviewer for D51460: [WebAssembly] Be a little more conservative in WebAssemblyFixFunctionBitcasts: jgravelle-google.
Aug 29 2018, 3:19 PM
sbc100 created D51460: [WebAssembly] Be a little more conservative in WebAssemblyFixFunctionBitcasts.
Aug 29 2018, 3:15 PM
sbc100 accepted D51449: [WebAssembly] clang-format (NFC).
Aug 29 2018, 1:46 PM

Aug 22 2018

sbc100 added inline comments to D51065: [WebAssembly] Ensure relocation entries are ordered by offset.
Aug 22 2018, 10:14 AM
sbc100 updated the diff for D51065: [WebAssembly] Ensure relocation entries are ordered by offset.
  • fixup
  • stable_sort
Aug 22 2018, 10:14 AM

Aug 21 2018

sbc100 updated the summary of D51065: [WebAssembly] Ensure relocation entries are ordered by offset.
Aug 21 2018, 4:25 PM
sbc100 retitled D51065: [WebAssembly] Ensure relocation entries are ordered by offset from [WebAssembly] Ensure relocation are entries are ordered by offset to [WebAssembly] Ensure relocation entries are ordered by offset.
Aug 21 2018, 2:24 PM
sbc100 updated subscribers of D51063: [WebAssembly] Optimise relocation iteration to remove n^2 loop. NFC..
Aug 21 2018, 2:23 PM
sbc100 added a comment to D51063: [WebAssembly] Optimise relocation iteration to remove n^2 loop. NFC..

This change depends on relocations being alwasy stored in offset order, and I'm not convinced that is always that case. I created this change to ensure that llvm always create reloction sections like this but I'm not under what circumstance is might currently fail to do so: https://reviews.llvm.org/D51065

Aug 21 2018, 2:23 PM
sbc100 created D51065: [WebAssembly] Ensure relocation entries are ordered by offset.
Aug 21 2018, 2:21 PM
sbc100 updated the summary of D51063: [WebAssembly] Optimise relocation iteration to remove n^2 loop. NFC..
Aug 21 2018, 2:11 PM
sbc100 updated the diff for D51063: [WebAssembly] Optimise relocation iteration to remove n^2 loop. NFC..
  • cleanup
Aug 21 2018, 2:11 PM
sbc100 created D51063: [WebAssembly] Optimise relocation iteration to remove n^2 loop. NFC..
Aug 21 2018, 2:06 PM
sbc100 added inline comments to D49446: [WebAssembly] Move .debug_line section address of dead function outside section range.
Aug 21 2018, 11:42 AM
sbc100 added inline comments to D49446: [WebAssembly] Move .debug_line section address of dead function outside section range.
Aug 21 2018, 11:35 AM

Aug 20 2018

sbc100 retitled D50976: [WebAssembly] Update Config member to match command line option. NFC. from [WebAssembly] Update Config member to match command line option to [WebAssembly] Update Config member to match command line option. NFC..
Aug 20 2018, 10:01 AM
sbc100 created D50976: [WebAssembly] Update Config member to match command line option. NFC..
Aug 20 2018, 10:01 AM

Aug 17 2018

sbc100 added a comment to D50721: [WebAssembly] Preserve function signatures during LTO.

Ping

Aug 17 2018, 12:43 PM

Aug 16 2018

sbc100 accepted D50859: [WebAssembly] Remove temporary workaround for function bitcasts.

Yay!

Aug 16 2018, 12:00 PM

Aug 15 2018

sbc100 added a comment to D50477: WIP: Ensure that the type of size_t is represended as one of the fixed width types.

Please ignore this change, I had thought we wanted size_t to match either int32_t or int64_t, but it turns out that isn't a requirement and that code that expects this to be true needs fixing.

Aug 15 2018, 3:26 PM

Aug 14 2018

sbc100 added a reviewer for D50729: [WebAssembly] Don't compress LEBs by default: yurydelendik.
Aug 14 2018, 4:52 PM
sbc100 accepted D50728: [WebAssembly] Delete a specific push number from test expectations.
Aug 14 2018, 12:19 PM
sbc100 updated the diff for D50729: [WebAssembly] Don't compress LEBs by default.
  • add check for -r
Aug 14 2018, 12:15 PM
sbc100 added a comment to D50729: [WebAssembly] Don't compress LEBs by default.

Perhaps I should make this -z compress although I'm not sure what the policy is in using -z keywords vs regular flags.

Aug 14 2018, 12:13 PM
sbc100 added a reviewer for D50729: [WebAssembly] Don't compress LEBs by default: ruiu.
Aug 14 2018, 12:12 PM
sbc100 created D50729: [WebAssembly] Don't compress LEBs by default.
Aug 14 2018, 12:12 PM
sbc100 added a reviewer for D50721: [WebAssembly] Preserve function signatures during LTO: ruiu.
Aug 14 2018, 11:43 AM
sbc100 updated the diff for D50721: [WebAssembly] Preserve function signatures during LTO.
  • fix type mismatch
  • add test
Aug 14 2018, 11:38 AM
sbc100 created D50721: [WebAssembly] Preserve function signatures during LTO.
Aug 14 2018, 9:44 AM

Aug 8 2018

sbc100 abandoned D50477: WIP: Ensure that the type of size_t is represended as one of the fixed width types.
Aug 8 2018, 5:51 PM
sbc100 updated the diff for D50477: WIP: Ensure that the type of size_t is represended as one of the fixed width types.

remove debugging

Aug 8 2018, 2:11 PM
sbc100 created D50477: WIP: Ensure that the type of size_t is represended as one of the fixed width types.
Aug 8 2018, 2:10 PM
sbc100 added a comment to D50472: Fix wasm backend compilation on gcc 5.4: variable name cannot match class.

Hmm. I don't know the convention. Not a huge and of "The" prefix. LGTM otherwise.

Aug 8 2018, 1:53 PM
sbc100 added a comment to D50424: [WebAssembly] Group rodata into a single output segment.

I think you are right about the optimized builds. One thing I thought we might want to preserve is that data objects from the same section should be grouped together contiguously.. but I think we can probably do that while at the same time putting it all in a single segment.

Aug 8 2018, 11:02 AM
sbc100 added a comment to D40526: [WebAssembly] Change size_t to `unsigned long`.

Is this related to the issue reported in the thread here?

Aug 8 2018, 10:55 AM
sbc100 added a comment to D40526: [WebAssembly] Change size_t to `unsigned long`.

Should we change int32_t as well? Or does the above code just need fixing? I'm a little concerned because this code is very portable which implies that WebAssembly might be doing something unusual here.

Aug 8 2018, 9:41 AM
sbc100 added a comment to D40526: [WebAssembly] Change size_t to `unsigned long`.

I'm running into an issue with an internal codebase that assumes that size_t is equivalent to either int32_t or int64_t which this change invalidates.

Aug 8 2018, 9:39 AM
sbc100 accepted D50387: [WASM] Fix overflow when reading custom section.

I agree this feels a bit awkward, but I prefer to custom bounds checking I think.

Aug 8 2018, 8:36 AM

Aug 7 2018

sbc100 added reviewers for D50424: [WebAssembly] Group rodata into a single output segment: fitzgen, ruiu.
Aug 7 2018, 9:03 PM
sbc100 created D50424: [WebAssembly] Group rodata into a single output segment.
Aug 7 2018, 9:02 PM
sbc100 added inline comments to D50387: [WASM] Fix overflow when reading custom section.
Aug 7 2018, 3:48 PM