cputlb: Move cpu->pending_tlb_flush to env->tlb_c.pending_flush

Protect it with the tlb_lock instead of using atomics.
The move puts it in or near the same cacheline as the lock;
using the lock means we don't need a second atomic operation
in order to perform the update.  Which makes it cheap to also
update pending_flush in tlb_flush_by_mmuidx_async_work.

Tested-by: Emilio G. Cota <cota@braap.org>
Reviewed-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

1 files changed, 7 insertions, 1 deletions

diff --git a/include/exec/cpu-defs.h b/include/exec/cpu-defs.h
index 9005923b4d..659c73d2a1 100644
--- a/include/exec/cpu-defs.h
+++ b/include/exec/cpu-defs.h
@@ -145,8 +145,14 @@ typedef struct CPUIOTLBEntry {
  * Data elements that are shared between all MMU modes.
  */
 typedef struct CPUTLBCommon {
-    /* lock serializes updates to tlb_table and tlb_v_table */
+    /* Serialize updates to tlb_table and tlb_v_table, and others as noted. */
     QemuSpin lock;
+    /*
+     * Within pending_flush, for each bit N, there exists an outstanding
+     * cross-cpu flush for mmu_idx N.  Further cross-cpu flushes to that
+     * mmu_idx may be discarded.  Protected by tlb_c.lock.
+     */
+    uint16_t pending_flush;
 } CPUTLBCommon;
 
 /*