Am 16. November 2026 01:34:29 UTC schrieb Bin Meng <[email protected]>:
>On Sun, Nov 8, 2026 at 8:04 PM Bin Meng <[email protected]> wrote:
>>
>> This series corrects SPR access, cache and MMU configuration, and
>> instruction availability in the Freescale Book E CPU models. It also
>> fixes host PVR matching and adds system TCG tests for e500mc, e5500
>> and e6500.
>>
>> The patches are grouped as follows:
>>
>> - patch #1-#3 and #5, bug fixes to PIR write access, the MAS5 SPR number
>> and cache block/line sizes (qemu-stable cc'ed)
>> - patch #4 and #6, complete e500mc/e5500 L1 cache configuration and
>> remove redundant e6500 cache geometry bits
>> - patch #7-#11, register the missing cache, MMU-assist and several
>> control SPRs
>> - patch #12-#14, split the PowerISA 2.05 instructions into integer and
>> floating-point groups, due to differences seen on e500 series cores
>> - patch #15-#17: Add TCG tests for illegal-instruction exceptions, byte
>> comparison and parity results, and FPSCR field-mask handling
>> - patch #18-#21, complete the e500 family MMU configuration, correct
>> PID register counts and express MMUCFG/TLBnCFG with named fields
>> - patch #22-#23, fix EPCR write width and host PVR revision matching
>> for Book E KVM
>>
>> Changes in v3:
>> - Cover the new system test directories in MAINTAINERS
>> - Add the missing SPDX license identifier to the README
>> - target/ppc: Build e6500 MMUCFG values from field encodings
>> - target/ppc: Complete MMU configuration for e500 family cores
>> - target/ppc: Limit e500mc and e5500 to one PID register
>> - target/ppc: Express e6500 TLB configuration with named fields
>> - target/ppc: Keep embedded processor control within 32 bits
>> - target/ppc: Match e500mc/e5500/e6500 host PVR revisions
>>
>> Changes in v2:
>> - Group ISA 2.05 features into integer and floating-point flags
>> - Use the first free bit for the floating-point group
>> - Align flag definitions and comment endings
>> - Enable the shared ISA 2.05 integer flag on e5500 and e6500
>> - Align the FPSCR W/L behavior flag comment ending
>> - Rebase the PPC64 test entry onto the Meson variable cleanup
>>
>> Bin Meng (23):
>> target/ppc: Correct PIR write access on e500 cores
>> target/ppc: Correct the MAS5 SPR number
>> target/ppc: Correct the e500mc and e5500 cache block size fields
>> target/ppc: Complete e500mc and e5500 L1 cache configuration
>> target/ppc: Correct cache line size for e6500
>> target/ppc: Drop redundant e6500 L1 cache geometry bits
>> target/ppc: Register L1CSR2 on e500mc/e5500/e6500 cores
>> target/ppc: Register L2CFG0 and L2CSR1 on e500mc and e5500
>> target/ppc: Register MAS5 and MAS8 on e500mc/e5500/e6500
>> target/ppc: Register PWRMGTCR0 on e6500
>> target/ppc: Register IVOR42 on e6500
>> target/ppc: Split the ISA 2.05 instruction feature group
>> target/ppc: Enable byte comparison and parity on e5500 and e6500
>> target/ppc: Ignore FPSCR W/L fields on e500mc, e5500 and e6500
>> tests/tcg/ppc: Add Book E illegal instruction tests
>> tests/tcg/ppc: Test byte comparison and parity on e5500/e6500
>> tests/tcg/ppc: Test FPSCR W/L handling on e500mc/e5500/e6500
>> target/ppc: Build e6500 MMUCFG values from field encodings
>> target/ppc: Complete MMU configuration for e500 family cores
>> target/ppc: Limit e500mc and e5500 to one PID register
>> target/ppc: Express e6500 TLB configuration with named fields
>> target/ppc: Keep embedded processor control within 32 bits
>> target/ppc: Match e500mc/e5500/e6500 host PVR revisions
>>
>> MAINTAINERS | 2 +-
>> target/ppc/cpu.h | 45 ++++++----
>> target/ppc/cpu_init.h | 3 +-
>> target/ppc/cpu_init.c | 99 +++++++++++++++++-----
>> target/ppc/translate.c | 10 ++-
>> target/ppc/translate/fixedpoint-impl.c.inc | 6 +-
>> target/ppc/translate/fp-impl.c.inc | 14 ++-
>> target/ppc/translate/fp-ops.c.inc | 9 +-
>> meson_options.txt | 4 +
>> scripts/meson-buildoptions.sh | 2 +
>> tests/tcg/meson.build | 1 +
>> tests/tcg/ppc/meson.build | 5 ++
>> tests/tcg/ppc/system/README | 24 ++++++
>> tests/tcg/ppc/system/boot.S | 92 ++++++++++++++++++++
>> tests/tcg/ppc/system/byte-parity.S | 33 ++++++++
>> tests/tcg/ppc/system/fpscr.S | 75 ++++++++++++++++
>> tests/tcg/ppc/system/isa.out | 1 +
>> tests/tcg/ppc/system/kernel.ld | 18 ++++
>> tests/tcg/ppc/system/macros.inc | 56 ++++++++++++
>> tests/tcg/ppc/system/meson.build | 39 +++++++++
>> tests/tcg/ppc/system/unsupported.S | 31 +++++++
>> tests/tcg/ppc64/meson.build | 2 +
>> tests/tcg/ppc64/system/meson.build | 33 ++++++++
>> 23 files changed, 554 insertions(+), 50 deletions(-)
>> create mode 100644 tests/tcg/ppc/meson.build
>> create mode 100644 tests/tcg/ppc/system/README
>> create mode 100644 tests/tcg/ppc/system/boot.S
>> create mode 100644 tests/tcg/ppc/system/byte-parity.S
>> create mode 100644 tests/tcg/ppc/system/fpscr.S
>> create mode 100644 tests/tcg/ppc/system/isa.out
>> create mode 100644 tests/tcg/ppc/system/kernel.ld
>> create mode 100644 tests/tcg/ppc/system/macros.inc
>> create mode 100644 tests/tcg/ppc/system/meson.build
>> create mode 100644 tests/tcg/ppc/system/unsupported.S
>> create mode 100644 tests/tcg/ppc64/system/meson.build
>>
>> ---
>
>Ping for review?
Hi Bin,
thanks for taking care of the e500 CPU models! Back in the days when I became
maintainer of the e500 machines, I deliberately didn't volunteer for the CPUs
since I don't have the bandwidth to do it. So go ahead, I'm fine with the
changes.
Best regards,
Bernhard