[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #36 from Richard Biener --- Ok, so now we have this workaround in which means in theory this bug should be fixed... we seem to get a gcc-testresults posting for darwin about once a week (possibly for each snapshot?) so I'll try to

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #35 from Richard Biener --- Author: rguenth Date: Fri Mar 2 07:42:58 2018 New Revision: 258123 URL: https://gcc.gnu.org/viewcvs?rev=258123=gcc=rev Log: 2018-03-02 Richard Biener PR target/82005

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #34 from Richard Biener --- (In reply to Richard Biener from comment #31) > (In reply to Dominique d'Humieres from comment #30) > > > However I have restore my testing of the gfortran test suite with -g -flto > > > and I see the

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 Bug 82005 depends on bug 84645, which changed state. Bug 84645 Summary: -flto -g0 at compile-time vs. -flto -g at link time ICEs in add_dwarf_attr, at dwarf2out.c:4353 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84645 What

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 Eric Gallager changed: What|Removed |Added See Also|https://gcc.gnu.org/bugzill |

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 Eric Gallager changed: What|Removed |Added Keywords||patch URL|

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #31 from Richard Biener --- (In reply to Dominique d'Humieres from comment #30) > > However I have restore my testing of the gfortran test suite with -g -flto > > and I see the following new failures: > > > > FAIL:

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #30 from Dominique d'Humieres --- > However I have restore my testing of the gfortran test suite with -g -flto > and I see the following new failures: > > FAIL: gfortran.dg/namelist_14.f90 -g -flto (internal compiler error) >

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #29 from Dominique d'Humieres --- > How do the remaining failures look like? Part of them are darwin specific and several have already been reported (see

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #28 from rguenther at suse dot de --- On Wed, 14 Feb 2018, dominiq at lps dot ens.fr wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 > > --- Comment #27 from Dominique d'Humieres --- > > For completeness here is a

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #27 from Dominique d'Humieres --- > For completeness here is a fixed diff: With the patch the failures (-m32/-m64) went down from 1059 to 467.

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #26 from Richard Biener --- (In reply to rguent...@suse.de from comment #25) > On February 13, 2018 5:35:22 PM GMT+01:00, simon at pushface dot org > wrote: >

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #25 from rguenther at suse dot de --- On February 13, 2018 5:35:22 PM GMT+01:00, simon at pushface dot org wrote: >https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 > >--- Comment #24 from simon at

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #24 from simon at pushface dot org --- (In reply to Richard Biener from comment #22) >These are already implemented > by darwin in darwin_asm_lto_{start,end} where it could simply temporarily >

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #23 from Dominique d'Humieres --- > Can you check whether that solves all immediate issues with compiling > and linking? ... AFAICT no (preliminary results): FAIL: c-c++-common/asan/alloca_big_alignment.c -Os output pattern test

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #22 from Richard Biener --- So we don't seem to be able to do anything reasonable for GCC 8. This means triggering the fallback solution of disabling debug for Darwin. There's multiple possibilities of places to choose from where

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #21 from Dominique d'Humieres --- For the record, with the patch in comment 6 the test g++.dg/asan/use-after-scope-1.C compiled with /opt/gcc/gcc8w/bin/g++ -fsanitize=address -flto

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 Richard Biener changed: What|Removed |Added CC||simon at pushface dot org --- Comment

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 Richard Biener changed: What|Removed |Added Priority|P3 |P1

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #19 from Richard Biener --- (In reply to Iain Sandoe from comment #18) > So .. just some notes on thoughts so far, not much conclusion yet. > > (In reply to rguent...@suse.de from comment #17) > > On September 16, 2017 2:37:02 PM

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #18 from Iain Sandoe --- So .. just some notes on thoughts so far, not much conclusion yet. (In reply to rguent...@suse.de from comment #17) > On September 16, 2017 2:37:02 PM GMT+02:00, "iains at gcc dot gnu.org" >

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #17 from rguenther at suse dot de --- On September 16, 2017 2:37:02 PM GMT+02:00, "iains at gcc dot gnu.org" wrote: >https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 > >--- Comment #16 from Iain Sandoe

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #16 from Iain Sandoe --- So.. trying to understand what we need to do to get this working on Darwin. in the original .s we have ... .section __GNU_DWARF_LTO,__debug_info,regular,debug Lsection__debug_info_lto: Ldebug_info0:

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #15 from Dominique d'Humieres --- Results posted at https://gcc.gnu.org/ml/gcc-testresults/2017-09/msg01338.html. Hundreds of failures coming from this PR.

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 Richard Biener changed: What|Removed |Added CC||egallager at gcc dot gnu.org ---

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #13 from rguenther at suse dot de --- On Thu, 31 Aug 2017, dominiq at lps dot ens.fr wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 > > --- Comment #12 from Dominique d'Humieres --- > > On LTO tescases or on others?

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #12 from Dominique d'Humieres --- > On LTO tescases or on others? LTO is expected to fail in weird ways > until simple object support for mach-o is implemented. My gfortran tests run the additional options -g -flto. The dsymutil

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #11 from rguenther at suse dot de --- On August 30, 2017 8:18:07 PM GMT+02:00, "dominiq at lps dot ens.fr" wrote: >https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 > >--- Comment #10 from Dominique

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #10 from Dominique d'Humieres --- I can prune the warnings with regsub -all "(^|\n)while processing\[^\n\]*" $text "" text regsub -all "(^|\n)warning: could not find referenced DIE" $text "" text however I see failures of

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #9 from Dominique d'Humieres --- > If you manually do the extraction / renaming and link with the early > debug part it should work(?) (I didn't try myself, doing this > manually might be a bit tricky -- look at the compile

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #8 from rguenther at suse dot de --- On Wed, 30 Aug 2017, dominiq at lps dot ens.fr wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 > > --- Comment #7 from Dominique d'Humieres --- > With the patch in comment 6,

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #7 from Dominique d'Humieres --- With the patch in comment 6, compiling the test in comment 0 with -g -flto gives while processing /var/folders/8q/sh_swgz96r7f5vnn08f7fxr0gn/T//ccqniUp8.ltrans0.ltrans.o: warning: could not find

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #6 from Richard Biener --- To recap the situation we have sth like earlyDebug.o section .debug_info: sym1: CU1 ... CU2 at coff1 lateDebug.o section .debug_info: ... DW_TAG_imported_unit DW_AT_import $sym1 - .debug_info +

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #5 from Richard Biener --- If port maintainers cannot resolve this somehow the workaround for darwin is to force -g0 at LTO link time. Unfortunately darwin doesn't support -gstabs (anymore).

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 --- Comment #4 from Richard Biener --- That is, looking for documentation on Mach-O I find in mach-o/reloc.h * Another type of generic relocation, GENERIC_RELOC_SECTDIFF, is to support * the difference of two symbols defined in different

[Bug target/82005] [8 regression] early lto debug creates invalid assembly on Darwin

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82005 Richard Biener changed: What|Removed |Added Target Milestone|--- |8.0