Re: r278655 - [CodeGen] Correctly implement the AVX512 psABI rules

LGTM. But it would be great if someone with more exposure to CFE could double check. -Gerolf > On Aug 14, 2016, at 11:39 PM, David Majnemer via cfe-commits > wrote: > > Author: majnemer > Date: Mon Aug 15 01:39:18 2016 > New Revision: 278655 > > URL:

Re: [PATCH] D22900: Revert r244207 - Mark calls in thunk functions as tail-call optimization

This is the IR I see in today’s trunk: *** IR Dump Before Module Verifier *** ; Function Attrs: noinline optsize ssp define i32 @_ZThn4_N1C4SeekE6_LARGE(%class.C* %this, %union._LARGE* byval align 4) unnamed_addr #2 align 2 { entry: %L = alloca %union._LARGE, align 8 %this.addr = alloca

Re: [PATCH] D22900: Revert r244207 - Mark calls in thunk functions as tail-call optimization

> On Jul 31, 2016, at 12:11 AM, Amjad Aboud wrote: > > aaboud added a comment. > > Reverting https://reviews.llvm.org/rL244207, would be fine if we assure that > PR24235, is fixed in another way. I think it is inappropriate for anyone who introduced a bug to condition

Re: [PATCH] D22900: Revert r244207 - Mark calls in thunk functions as tail-call optimization

> On Jul 31, 2016, at 1:46 AM, Amjad Aboud wrote: > > aaboud added a comment. > >> ISTM that the DWARF spec intended such thunks to be encoded as >> `DW_AT_trampoline`. That seems more appropriate than relying on codegen >> emitting a tailcall. This way the debugger

Re: [PATCH] D22900: Revert r244207 - Mark calls in thunk functions as tail-call optimization

Sounds like you care highly about your commit :-) > On Jul 29, 2016, at 10:10 AM, Michael Kuperstein wrote: > > mkuper resigned from this revision. > mkuper removed a reviewer: mkuper. > mkuper added a comment. > > I really don't understand anything about this. :-) > >

Re: [PATCH] D22900: Revert r244207 - Mark calls in thunk functions as tail-call optimization

Gerolf updated this revision to Diff 66075. Gerolf added a comment. Reduced test case. https://reviews.llvm.org/D22900 Files: lib/CodeGen/CGVTables.cpp test/CodeGenCXX/microsoft-abi-structors.cpp test/CodeGenCXX/tail-byval.cpp Index: test/CodeGenCXX/tail-byval.cpp

Re: [PATCH] D22900: Revert r244207 - Mark calls in thunk functions as tail-call optimization

Gerolf added a comment. Nope, I don't see the tail call. Anyway, I'll simplify my test case. Don't worry about it. clang++ -cc1 -x c++ -emit-llvm -triple i386-apple-darwin9 t.cpp cat t.ll: ; ModuleID = 't.cpp' source_filename = "t.cpp" target datalayout =

Re: [PATCH] D22900: Revert r244207 - Mark calls in thunk functions as tail-call optimization

Gerolf added a comment. Please add the options you used to compile? I can certainly shrink the test case a bit before I commit. https://reviews.llvm.org/D22900 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

[PATCH] D22900: Revert r244207 - Mark calls in thunk functions as tail-call optimization

Gerolf created this revision. Gerolf added reviewers: eli.friedman, mkuper. Gerolf added a subscriber: cfe-commits. This is just closing the loop for https://www.mail-archive.com/cfe-commits@lists.llvm.org/msg28837.html with a test case and fixes PR28748 which had been introduced by r244207.

Re: revert clang r244207 - Mark calls in thunk functions as tail-call optimization candidates? { Re: [PATCH] D21948: [DSE] fix - missing store to runtime stack in thunk with tail call bvval arg }

> On Jul 7, 2016, at 8:36 PM, Gerolf Hoflehner wrote: > > Ahmed pointed out the clang commit causing the issue is r244207 - Mark calls > in thunk functions as tail-call optimization candidates > > The purpose of the fix was to prevent thunks from showing up on the call