Hi Jeff, Thanks for all your support. Yes, this is running in an MDCP VM. The host is x86_64-pc-linux-gnu and the target is riscv64-unknown-elf.
I fixed the QEMU setup and reran the tests in an MDCP VM using the locally built QEMU `riscv-qemu` DejaGNU board for `riscv64-unknown-elf`. The PR126245-specific tests pass at all tested optimization levels, including the LTO configurations. The full testsuite results have also improved substantially. For GCC, unexpected failures decreased from 39,015 to 1,795 and unresolved testcase decreased from 69,521 to 148. The G++ run still reports 17,535 unexpected failures, so I am continuing to classify the remaining failures and analyse. Thanks, Naveen >
