I wondered if the discrepancy I was seeing in fill patterns was due to some sort of non-deterministic resize being triggered. In theory we could resize away at any point which might account for the difference.
Signed-off-by: Alex Bennée <[email protected]> --- accel/tcg/cputlb.c | 2 ++ accel/tcg/trace-events | 1 + 2 files changed, 3 insertions(+) diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index c35df27caf..63f2a23709 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -201,6 +201,8 @@ static void tlb_mmu_resize_locked(CPUTLBDesc *desc, CPUTLBDescFast *fast, return; } + trace_tlb_resize(old_size, new_size); + g_free(fast->table); g_free(desc->fulltlb); diff --git a/accel/tcg/trace-events b/accel/tcg/trace-events index 31dda01c12..11b49a63f1 100644 --- a/accel/tcg/trace-events +++ b/accel/tcg/trace-events @@ -10,6 +10,7 @@ exec_tb_exit(void *last_tb, unsigned int flags) "tb:%p flags=0x%x" memory_notdirty_write_access(uint64_t vaddr, uint64_t ram_addr, unsigned size) "0x%" PRIx64 " ram_addr 0x%" PRIx64 " size %u" memory_notdirty_set_dirty(uint64_t vaddr) "0x%" PRIx64 tlb_fill(uint64_t vaddr, int size, int access_type, int mmu_idx) "0x%" PRIx64 "/%d %d %d" +tlb_resize(size_t old, size_t new) "%zu -> %zu" # translate-all.c translate_block(void *tb, uintptr_t pc, const void *tb_code) "tb:%p, pc:0x%"PRIxPTR", tb_code:%p" -- 2.39.2
