[Bug other/78063] libbacktrace fails to handle cross CU DW_AT_abstract_origin
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78063 Tom de Vries changed: What|Removed |Added CC||vries at gcc dot gnu.org --- Comment #4 from Tom de Vries --- Created attachment 45622 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=45622=edit Tentative patch "[libbacktrace] Handle DW_FORM_ref_addr" Patch passes regular bootstrap and reg-test.
[Bug other/78063] libbacktrace fails to handle cross CU DW_AT_abstract_origin
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78063 Richard Biener changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed||2018-05-08 Ever confirmed|0 |1 --- Comment #3 from Richard Biener --- (In reply to Romain Geissler from comment #2) > Hi, > > As written in previous comments, this now breaks both libbacktrace tests + > all sanitizer tests using backtrace when using gcc >= 8 and an LTO > bootstrapped compiler. The sanitizer tests should behave OK (but I didn't recently double-check). I've adjusted their expected patterns to cope with the new DWARF. > Shall we XFAIL temporarily these tests in case of LTO bootstrap (if that is > even possible) ? > > Here is a list of tests which are failing for me: > > test1: [0]: missing file name or function name > FAIL: backtrace_full alloc stress > FAIL: edtest > test1: [0]: missing file name or function name > test1: [0]: missing file name or function name > test1: [0]: missing file name or function name > test1: [0]: missing file name or function name > test1: [0]: missing file name or function name > test1: [0]: missing file name or function name > test1: [0]: missing file name or function name > test1: [0]: missing file name or function name > test1: [0]: missing file name or function name > test1: [0]: missing file name or function name > FAIL: threaded backtrace_full noinline > FAIL: ttest I think the easiest fix would be to add -fno-lto to the test CFLAGS. We already build some of the host libraries with -fno-lto for various reasons. Ah, only libdecnumber it seems. Of course the best thing would be to fix this bug rather than working around it... > Cheers, > Romain
[Bug other/78063] libbacktrace fails to handle cross CU DW_AT_abstract_origin
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78063 Romain Geissler changed: What|Removed |Added CC||romain.geissler at amadeus dot com --- Comment #2 from Romain Geissler --- Hi, As written in previous comments, this now breaks both libbacktrace tests + all sanitizer tests using backtrace when using gcc >= 8 and an LTO bootstrapped compiler. Shall we XFAIL temporarily these tests in case of LTO bootstrap (if that is even possible) ? Here is a list of tests which are failing for me: test1: [0]: missing file name or function name FAIL: backtrace_full alloc stress FAIL: edtest test1: [0]: missing file name or function name test1: [0]: missing file name or function name test1: [0]: missing file name or function name test1: [0]: missing file name or function name test1: [0]: missing file name or function name test1: [0]: missing file name or function name test1: [0]: missing file name or function name test1: [0]: missing file name or function name test1: [0]: missing file name or function name test1: [0]: missing file name or function name FAIL: threaded backtrace_full noinline FAIL: ttest Cheers, Romain
[Bug other/78063] libbacktrace fails to handle cross CU DW_AT_abstract_origin
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78063 --- Comment #1 from Richard Biener --- Now manifests itself for LTO bootstrapped compilers.