Bug 13298 tracked 100s of ptest failures for arm64 in valgrind. That is no longer occuring, and we have them updated to run most of the tests. there are 12 remaining ptest failures that should be tracked by this bug. List of failing ptests: gdbserver_tests/hgtls memcheck/tests/dw4 memcheck/tests/leak_cpp_interior memcheck/tests/varinfo1 memcheck/tests/varinfo2 memcheck/tests/varinfo3 memcheck/tests/varinfo4 memcheck/tests/varinfo5 memcheck/tests/varinfo6 memcheck/tests/varinforestrict helgrind/tests/hg05_race2 helgrind/tests/tc20_verifywrap
Correction: the bug mentioned above should be Bug 13520.
After some investigation, it seems like all the tests are failing because there is this error exists on arm64, but never on x86-64: --16431-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0xa3 --16431-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0xa3 --16431-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0xa3 --16431-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0xa3 --16431-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0xa3 --16431-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0xa3 --16431-- warning: evaluate_Dwarf3_Expr: unhandled DW_OP_ 0xa3
FWIW: elfutils-native/usr/include/dwarf.h: DW_OP_entry_value = 0xa3,
A missing case in https://sourceware.org/git/?p=valgrind.git;a=blob;f=coregrind/m_debuginfo/d3basics.c;h=d48589a30ca2555761ad221fe400b44297d1ca00;hb=HEAD#l701. Of course I have no idea what the entry_value opcode is for...
Bulk move to 5.0-M4
Park in 5.99 until I'm sure I have time to work on an issue.