On Tue, 08 Dec 2020 14:56:08 PST (-0800), Alistair Francis wrote:
> The RISC-V QEMU port currently has lot of preprocessor directives that
> check if we are targetting a 32-bit or 64-bit CPU. This means that the
> 64-bit RISC-V target can not run 32-bit CPUs. This is different to most
> other QEMU architectures and doesn't allow us to mix xlens (such as when
> running Hypervisors with different xlen guests).
> This series is a step toward removing some of those to allow us to use
> 32-bit CPUs on 64-bit RISC-V targets.
> v2:
> - Rebase on the latest RISC-V tree
>
> Alistair Francis (15):
> hw/riscv: Expand the is 32-bit check to support more CPUs
> target/riscv: Add a TYPE_RISCV_CPU_BASE CPU
> riscv: spike: Remove target macro conditionals
> riscv: virt: Remove target macro conditionals
> hw/riscv: boot: Remove compile time XLEN checks
> hw/riscv: virt: Remove compile time XLEN checks
> hw/riscv: spike: Remove compile time XLEN checks
> hw/riscv: sifive_u: Remove compile time XLEN checks
> target/riscv: fpu_helper: Match function defs in HELPER macros
> target/riscv: Add a riscv_cpu_is_32bit() helper function
> target/riscv: Specify the XLEN for CPUs
> target/riscv: cpu: Remove compile time XLEN checks
> target/riscv: cpu_helper: Remove compile time XLEN checks
> target/riscv: csr: Remove compile time XLEN checks
> target/riscv: cpu: Set XLEN independently from target
>
> include/hw/riscv/boot.h | 8 +-
> include/hw/riscv/spike.h | 6 --
> include/hw/riscv/virt.h | 6 --
> target/riscv/cpu.h | 8 ++
> target/riscv/cpu_bits.h | 4 +-
> hw/riscv/boot.c | 69 +++++++++------
> hw/riscv/sifive_u.c | 57 ++++++------
> hw/riscv/spike.c | 50 ++++++-----
> hw/riscv/virt.c | 36 ++++----
> target/riscv/cpu.c | 84 ++++++++++++------
> target/riscv/cpu_helper.c | 12 +--
> target/riscv/csr.c | 182 ++++++++++++++++++++------------------
> target/riscv/fpu_helper.c | 16 ++--
> 13 files changed, 302 insertions(+), 236 deletions(-)
Thanks, this is awesome! That's always been one of the headaches floating
around the port that I was hoping to have some time to sort out :). I just
gave it a quick look, but
Reviewed-by: Palmer Dabbelt <palmerdabbelt@google.com>
Acked-by: Palmer Dabbelt <palmerdabbelt@google.com>
as it certainly seems better than before.