Ping (2)  https://lore.kernel.org/qemu-devel/[email protected]/
 
----------------
Кому: [email protected] ([email protected]);
Копия: Riku Voipio ([email protected]), Ilya Leoshkevich ([email protected]), Laurent Vivier ([email protected]), Chinmay Rath ([email protected]), Richard Henderson ([email protected]), David Hildenbrand ([email protected]), Thomas Huth ([email protected]), [email protected], Paolo Bonzini ([email protected]), Nicholas Piggin ([email protected]), [email protected];
Тема: [PATCH v3 0/4] Add support for PowerPC e500 little-endian pages;
15.09.2025, 15:46, "Danila Zhebryakov" <[email protected]>:

These changes make powerpc booke206 mmu pass TLB_BSWAP flag for pages marked as LE, and also fixes all the issues this causes.
 - added TLB_BSWAP to non-MMIO flags
 - inserted additional check for LE page when bswap-ing instructions
 - removed assert for cross-page TLB_BSWAP accesses
 - added gdbstub fix to account for TLB_BSWAP of the currently running code

BE- and LE- atomic operations (used by load-and-reserve and store-conditional instructions) were unified to support TLB_BSWAP

Some of the fixes are based on what is done in https://github.com/YetAnotherMod/qemu

V2:
 - Removed usage of PAGE_USER bit for LE (comment on previous version)
 - Refactored MMU code so directly setting TLB_BSWAP is possible
 - moved translation LE page detect to occur 1 time per translation (comment on previous version)
 - Removed redundant assert in atomic_mmu_lookup (comment on previous version)
 - Added assert on cross-endian accesses (ideally, there should be an exception)
 - Changed atomic need_bswap handling (comment on previous version)
 - Rebased and fixed atomics

V3
 - unfix assumed typo (if vs iff). Everything else unchanged

According to RM, unaligned accesses between LE and BE pages should cause an exception on e500.
However, (as I understand it) supporting this would require adding a new callback to TCGCPUOps, likely just for E500 or maybe a couple other CPUs doing this.
Such accesses are clearly not normal and usually don't occur in real applications, so I think just placing assert is OK here.

Danila Zhebryakov (4):
  accel/tcg: Unify big- and little- endian atomic ops
  target/ppc: refactor MMU helpers
  target/ppc: Add support for LE pages on PowerPC booke206 mmu
  target/ppc: fix GDB stub to work correctly with LE pages

 accel/tcg/atomic_common.c.inc | 36 +--
 accel/tcg/atomic_template.h | 326 ++++++++++------------------
 accel/tcg/cputlb.c | 29 ++-
 accel/tcg/tcg-runtime.h | 48 +---
 accel/tcg/user-exec.c | 2 +-
 include/accel/tcg/cpu-ldst-common.h | 51 ++---
 target/m68k/op_helper.c | 4 +-
 target/ppc/gdbstub.c | 17 +-
 target/ppc/internal.h | 3 +-
 target/ppc/mmu-booke.c | 67 +++---
 target/ppc/mmu-booke.h | 2 +-
 target/ppc/mmu-hash32.c | 30 +--
 target/ppc/mmu-hash32.h | 3 +-
 target/ppc/mmu-hash64.c | 14 +-
 target/ppc/mmu-hash64.h | 2 +-
 target/ppc/mmu-radix64.c | 20 +-
 target/ppc/mmu-radix64.h | 2 +-
 target/ppc/mmu_common.c | 75 ++++---
 target/ppc/mmu_helper.c | 15 +-
 target/ppc/translate.c | 42 +++-
 target/s390x/tcg/mem_helper.c | 6 +-
 tcg/tcg-op-ldst.c | 43 ++--
 22 files changed, 370 insertions(+), 467 deletions(-)
 

--
2.47.2
 

Reply via email to