[Bug target/82092] gcc fails to link genmodes on darwin (cfiStartsArray[i] != cfiStartsArray[i-1])
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82092 --- Comment #10 from Ryan Schmidt --- Is gcc8 ever going to be buildable on macOS again? It's been unbuildable for 11 months now.
[Bug target/82092] gcc fails to link genmodes on darwin (cfiStartsArray[i] != cfiStartsArray[i-1])
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82092 --- Comment #9 from Ryan Schmidt --- Is it safe to use the patch? Will a gcc built with the patch produce correct code? If so, I would like to include it in MacPorts so that I can update our gcc8 port to a newer version. Because of this problem, we have not been able to do so since June 2017.
[Bug target/82092] gcc fails to link genmodes on darwin (cfiStartsArray[i] != cfiStartsArray[i-1])
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82092 --- Comment #8 from Jürgen Reuter --- Another thing I observed: one difference between the laptop where the bootstrap worked and the one where it didn't work (besides the different chipsets i5 vs. i7) is that on the one where the bootstrap doesn't work, the MAC OS X Security Integrity Protocol (SIP) is activated, while on the one where the bootstrap worked it is deactivated. Could it be that during the bootstrap procedure the SIP causes some trouble in that some of the settings of the gcc to be bootstrapped are being blocked?
[Bug target/82092] gcc fails to link genmodes on darwin (cfiStartsArray[i] != cfiStartsArray[i-1])
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82092 --- Comment #7 from Jürgen Reuter --- I can confirm that the patch above indeed does work. Interestingly, I don't see the bootstrap on all combinations of Apple Hardware and MAC OS X versions. Apparently, the problem (for me) only shows up on an 2011 Apple i5 with both Sierra and High Sierra and XCode 9.X. AFAI can say, the problem doesn't arise for a 2015 i7 with the same combinations of OSX and XCode versions. Back to the fix by Dominique: yes, it works for the problematic combination I checked.
[Bug target/82092] gcc fails to link genmodes on darwin (cfiStartsArray[i] != cfiStartsArray[i-1])
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82092 Ryan Schmidt changed: What|Removed |Added CC||gcc at ryandesign dot com --- Comment #6 from Ryan Schmidt --- (In reply to Dominique d'Humieres from comment #3) > --- ../_clean/gcc/config/darwin.c 2017-09-18 15:49:48.0 +0200 > +++ gcc/config/darwin.c 2017-09-23 21:00:41.0 +0200 > @@ -3201,6 +3201,10 @@ darwin_override_options (void) >flag_reorder_blocks = 1; > } > > +// FIXME ; Jam this off until we figure out what codegen issues are > caused > +flag_reorder_blocks_and_partition = 0; > +flag_reorder_blocks = 1; > + > /* FIXME: flag_objc_sjlj_exceptions is no longer needed since there is > only > one valid choice of exception scheme for each runtime. */ > if (!global_options_set.x_flag_objc_sjlj_exceptions) > > Could you please test this patch? Yes, this patch does allow gcc 8-20171217 to build for me on macOS Sierra with Xcode 9.2. Is this patch safe for us to commit to MacPorts to be able to update our gcc8 port?
[Bug target/82092] gcc fails to link genmodes on darwin (cfiStartsArray[i] != cfiStartsArray[i-1])
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82092 --- Comment #5 from Jürgen Reuter --- Interestingly, I get the problem only with Xcode 9.0 and MAC OS X Sierra Darwin 16.6, but not with Xcode 9.0 and Mac OS X High Sierra Darwin 16.7.
[Bug target/82092] gcc fails to link genmodes on darwin (cfiStartsArray[i] != cfiStartsArray[i-1])
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82092 --- Comment #4 from Iain Sandoe --- (I don't see the reported problem on Darwin15 with any assembler/linker pair I have) However,we're certain there's an underlying issue. We're seeing a variety of problems which are all pointing some difficulty introduced by reorder and partition. My current suspicion is that some code fragment isn't a legal atom. With current trunk I'm seeing repeatable dwarf verify errors for eh sections in some of the libgcc objects - will be debugging these. Most likely this is also a dup of 81733. So - Dominique's comment #3 is worth testing to determine if the issue you see is another in this series. I don't think it's the same underlying cause as 57438 (that was caused by function bodies being elided when the function contained an inline __builtin_unreachable()). Arguably, 0-sized FDEs could be dropped by ld64 (I have a patch that does this), since an exception cannot originate from a 0-sized region. However, it would still be worth warning even so since it probably indicates a code-gen issue earlier in the toolchain (as we suspect here).
[Bug target/82092] gcc fails to link genmodes on darwin (cfiStartsArray[i] != cfiStartsArray[i-1])
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82092 Dominique d'Humieres changed: What|Removed |Added CC||iains at gcc dot gnu.org --- Comment #3 from Dominique d'Humieres --- On darwin10, starting at revision r249105, bootstrap fails with Undefined symbols: "CIE.eh", referenced from: CIE in _negvdi2_s.o unless I use --- ../_clean/gcc/config/darwin.c 2017-09-18 15:49:48.0 +0200 +++ gcc/config/darwin.c 2017-09-23 21:00:41.0 +0200 @@ -3201,6 +3201,10 @@ darwin_override_options (void) flag_reorder_blocks = 1; } +// FIXME ; Jam this off until we figure out what codegen issues are caused +flag_reorder_blocks_and_partition = 0; +flag_reorder_blocks = 1; + /* FIXME: flag_objc_sjlj_exceptions is no longer needed since there is only one valid choice of exception scheme for each runtime. */ if (!global_options_set.x_flag_objc_sjlj_exceptions) Could you please test this patch?
[Bug target/82092] gcc fails to link genmodes on darwin (cfiStartsArray[i] != cfiStartsArray[i-1])
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82092 Eric Gallager changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed||2017-09-24 CC||egallager at gcc dot gnu.org Ever confirmed|0 |1 --- Comment #2 from Eric Gallager --- I'll take the dup as confirmation.
[Bug target/82092] gcc fails to link genmodes on darwin (cfiStartsArray[i] != cfiStartsArray[i-1])
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82092 Eric Gallager changed: What|Removed |Added CC||juergen.reuter at desy dot de --- Comment #1 from Eric Gallager --- *** Bug 82315 has been marked as a duplicate of this bug. ***