Re: [Qemu-devel] [PATCH v6 00/37] AArch64 system emulation

[Peter Crosthwaite <peter.crosthwaite@xilinx.com>]

On Fri, Apr 11, 2014 at 2:14 AM, Peter Maydell <peter.maydell@linaro.org> wrote:
> Here's v5 of the AArch64 system emulation patchset.
> Still missing/TODO:
>  * SMP support (needs PSCI emulation in QEMU; being prototyped)
>  * save/restore (I have a patch which adds this but I think it will
>    look better if we consolidate AArch32 cpsr and AArch64 pstate
>    handling)
> but both of these I think are best done once this main series
> is committed to master.
>
> The changes v5->v6 are pretty minor, and these patches have
> been kicking around onlist for a long time, so I plan to
> put these in a pull request pretty much as soon as we reopen
> trunk after 2.0 releases. Last chance for review!
>

Ill do another sweep before the end of the week and review those Ive
skipped over so far for completeness.

Regards,
Peter

>
> Changes v5->v6:
>  * add extract64() when putting together 32-bit CBAR value
>  * be consistent about int vs bool for 1 bit fields in the
>    syn_insn_abort/syn_data_abort functions
>  * added some FIXMEs about inaccurate syndrome info for A32/T32
>    Neon unallocated insns when FP is disabled
>  * decided that using a15mpcore_priv in the virt machine is actually
>    the best approach, and added a suitable comment
> Changes v4->v5:
>  * new patches:
>    + MVFR registers
>    + various extra system registers
>    + don't expose wildcards for ARMv8
>    + make A15's CBAR R/O
>    + support interprocessing in set_pc
>  * minor tweaks per review (I haven't always taken the review
>    suggestion; see mail threads on the previous version of the
>    patch series)
>  * DC ZVA: use helper_ret_stb_mmu
>            make tlb_vaddr_to_host take param for access type
>  * NB: I didn't make gen_exception and gen_exception_internal shared
>    in patch 5: I think keeping the A64 and A32 decoders independent
>    is preferable
>  * have syn_insn_abort and syn_data_abort set the syndrome bit
>    for 'exception to same level' rather than making caller do it
> Changes v3->v4:
>  * reviewed patches from bottom of stack got committed to master
>  * new patches at top of stack
>  * addressed review issues on v8 mmu translation patch and DAIF patch
>
> thanks
> -- PMM
>
>
> Peter Maydell (33):
>   target-arm: Split out private-to-target functions into internals.h
>   target-arm: Implement AArch64 DAIF system register
>   target-arm: Define exception record for AArch64 exceptions
>   target-arm: Provide correct syndrome information for cpreg access
>     traps
>   target-arm: Add support for generating exceptions with syndrome
>     information
>   target-arm: A64: Correctly fault FP/Neon if CPACR.FPEN set
>   target-arm: A64: Add assertion that FP access was checked
>   target-arm: Fix VFP enables for AArch32 EL0 under AArch64 EL1
>   target-arm: Don't mention PMU in debug feature register
>   target-arm: A64: Implement DC ZVA
>   target-arm: Use dedicated CPU state fields for ARM946 access bit
>     registers
>   target-arm: Add AArch64 ELR_EL1 register.
>   target-arm: Implement SP_EL0, SP_EL1
>   target-arm: Implement AArch64 SPSR_EL1
>   target-arm: Move arm_log_exception() into internals.h
>   target-arm: Implement ARMv8 MVFR registers
>   target-arm: Add Cortex-A57 processor
>   hw/arm/virt: Add support for Cortex-A57
>   target-arm: Implement AArch64 views of AArch32 ID registers
>   target-arm: Implement AArch64 view of CONTEXTIDR
>   target-arm: Implement AArch64 view of ACTLR
>   target-arm: Implement ISR_EL1 register
>   target-arm: Remove THUMB2EE feature from AArch64 'any' CPU
>   target-arm: Don't expose wildcard ID register definitions for ARMv8
>   target-arm: Replace wildcarded cpreg definitions with precise ones for
>     ARMv8
>   target-arm: Implement auxiliary fault status registers
>   target-arm: Implement AArch64 address translation operations
>   target-arm: Implement RVBAR register
>   target-arm: Implement Cortex-A57 implementation-defined system
>     registers
>   target-arm: Implement CBAR for Cortex-A57
>   target-arm: Make Cortex-A15 CBAR read-only
>   target-arm: Handle the CPU being in AArch32 mode in the AArch64 set_pc
>   target-arm: Dump 32-bit CPU state if 64 bit CPU is in AArch32
>
> Rob Herring (4):
>   target-arm: Provide syndrome information for MMU faults
>   target-arm: Add v8 mmu translation support
>   target-arm: Implement AArch64 views of fault status and data registers
>   target-arm: Implement AArch64 EL1 exception handling
>
>  hw/arm/virt.c               |   8 +
>  include/exec/softmmu_exec.h |  52 +++
>  linux-user/main.c           |  56 ++-
>  target-arm/cpu-qom.h        |  10 +-
>  target-arm/cpu.c            |  30 +-
>  target-arm/cpu.h            |  81 +++--
>  target-arm/cpu64.c          | 115 ++++++-
>  target-arm/helper-a64.c     |  75 ++++
>  target-arm/helper.c         | 809 +++++++++++++++++++++++++++++++++-----------
>  target-arm/helper.h         |   7 +-
>  target-arm/internals.h      | 267 +++++++++++++++
>  target-arm/kvm32.c          |  19 +-
>  target-arm/kvm64.c          |  71 +++-
>  target-arm/machine.c        |  13 +-
>  target-arm/op_helper.c      |  92 ++++-
>  target-arm/translate-a64.c  | 441 +++++++++++++++++++++---
>  target-arm/translate.c      | 195 ++++++++---
>  target-arm/translate.h      |  23 +-
>  18 files changed, 1987 insertions(+), 377 deletions(-)
>  create mode 100644 target-arm/internals.h
>
> --
> 1.9.1
>
>