cpu: Move icount_decr to CPUNegativeOffsetState

Amusingly, we had already ignored the comment to keep this value
at the end of CPUState.  This restores the minimum negative offset
from TCG_AREG0 for code generation.

For the couple of uses within qom/cpu.c, without NEED_CPU_H, add
a pointer from the CPUState object to the IcountDecr object within
CPUNegativeOffsetState.

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

4 files changed, 31 insertions, 29 deletions

diff --git a/include/exec/cpu-all.h b/include/exec/cpu-all.h
index 5ae83405c8..253dd1d9a5 100644
--- a/include/exec/cpu-all.h
+++ b/include/exec/cpu-all.h
@@ -380,6 +380,7 @@ int cpu_exec(CPUState *cpu);
 static inline void cpu_set_cpustate_pointers(ArchCPU *cpu)
 {
     cpu->parent_obj.env_ptr = &cpu->env;
+    cpu->parent_obj.icount_decr_ptr = &cpu->neg.icount_decr;
 }
 
 /**
diff --git a/include/exec/cpu-defs.h b/include/exec/cpu-defs.h
index 921fbb4c36..c067994e5c 100644
--- a/include/exec/cpu-defs.h
+++ b/include/exec/cpu-defs.h
@@ -33,6 +33,7 @@
 #include "exec/hwaddr.h"
 #endif
 #include "exec/memattrs.h"
+#include "qom/cpu.h"
 
 #include "cpu-param.h"
 
@@ -232,7 +233,7 @@ typedef struct CPUTLB {
  * before CPUArchState, as a field named "neg".
  */
 typedef struct CPUNegativeOffsetState {
-    /* Empty */
+    IcountDecr icount_decr;
 } CPUNegativeOffsetState;
 
 #endif
diff --git a/include/exec/gen-icount.h b/include/exec/gen-icount.h
index 9cfa6ccce5..f7669b6841 100644
--- a/include/exec/gen-icount.h
+++ b/include/exec/gen-icount.h
@@ -5,8 +5,6 @@
 
 /* Helpers for instruction counting code generation.  */
 
-#define ENV_OFFSET   offsetof(ArchCPU, env)
-
 static TCGOp *icount_start_insn;
 
 static inline void gen_tb_start(TranslationBlock *tb)
@@ -21,7 +19,8 @@ static inline void gen_tb_start(TranslationBlock *tb)
     }
 
     tcg_gen_ld_i32(count, cpu_env,
-                   -ENV_OFFSET + offsetof(CPUState, icount_decr.u32));
+                   offsetof(ArchCPU, neg.icount_decr.u32) -
+                   offsetof(ArchCPU, env));
 
     if (tb_cflags(tb) & CF_USE_ICOUNT) {
         imm = tcg_temp_new_i32();
@@ -39,7 +38,8 @@ static inline void gen_tb_start(TranslationBlock *tb)
 
     if (tb_cflags(tb) & CF_USE_ICOUNT) {
         tcg_gen_st16_i32(count, cpu_env,
-                         -ENV_OFFSET + offsetof(CPUState, icount_decr.u16.low));
+                         offsetof(ArchCPU, neg.icount_decr.u16.low) -
+                         offsetof(ArchCPU, env));
     }
 
     tcg_temp_free_i32(count);
@@ -60,14 +60,18 @@ static inline void gen_tb_end(TranslationBlock *tb, int num_insns)
 static inline void gen_io_start(void)
 {
     TCGv_i32 tmp = tcg_const_i32(1);
-    tcg_gen_st_i32(tmp, cpu_env, -ENV_OFFSET + offsetof(CPUState, can_do_io));
+    tcg_gen_st_i32(tmp, cpu_env,
+                   offsetof(ArchCPU, parent_obj.can_do_io) -
+                   offsetof(ArchCPU, env));
     tcg_temp_free_i32(tmp);
 }
 
 static inline void gen_io_end(void)
 {
     TCGv_i32 tmp = tcg_const_i32(0);
-    tcg_gen_st_i32(tmp, cpu_env, -ENV_OFFSET + offsetof(CPUState, can_do_io));
+    tcg_gen_st_i32(tmp, cpu_env,
+                   offsetof(ArchCPU, parent_obj.can_do_io) -
+                   offsetof(ArchCPU, env));
     tcg_temp_free_i32(tmp);
 }
 
diff --git a/include/qom/cpu.h b/include/qom/cpu.h
index 98e12d914c..5ee0046b62 100644
--- a/include/qom/cpu.h
+++ b/include/qom/cpu.h
@@ -232,17 +232,25 @@ typedef struct CPUClass {
     bool gdb_stop_before_watchpoint;
 } CPUClass;
 
+/*
+ * Low 16 bits: number of cycles left, used only in icount mode.
+ * High 16 bits: Set to -1 to force TCG to stop executing linked TBs
+ * for this CPU and return to its top level loop (even in non-icount mode).
+ * This allows a single read-compare-cbranch-write sequence to test
+ * for both decrementer underflow and exceptions.
+ */
+typedef union IcountDecr {
+    uint32_t u32;
+    struct {
 #ifdef HOST_WORDS_BIGENDIAN
-typedef struct icount_decr_u16 {
-    uint16_t high;
-    uint16_t low;
-} icount_decr_u16;
+        uint16_t high;
+        uint16_t low;
 #else
-typedef struct icount_decr_u16 {
-    uint16_t low;
-    uint16_t high;
-} icount_decr_u16;
+        uint16_t low;
+        uint16_t high;
 #endif
+    } u16;
+} IcountDecr;
 
 typedef struct CPUBreakpoint {
     vaddr pc;
@@ -314,11 +322,6 @@ struct qemu_work_item;
  * @crash_occurred: Indicates the OS reported a crash (panic) for this CPU
  * @singlestep_enabled: Flags for single-stepping.
  * @icount_extra: Instructions until next timer event.
- * @icount_decr: Low 16 bits: number of cycles left, only used in icount mode.
- * High 16 bits: Set to -1 to force TCG to stop executing linked TBs for this
- * CPU and return to its top level loop (even in non-icount mode).
- * This allows a single read-compare-cbranch-write sequence to test
- * for both decrementer underflow and exceptions.
  * @can_do_io: Nonzero if memory-mapped IO is safe. Deterministic execution
  * requires that IO only be performed on the last instruction of a TB
  * so that interrupts take effect immediately.
@@ -328,6 +331,7 @@ struct qemu_work_item;
  * @as: Pointer to the first AddressSpace, for the convenience of targets which
  *      only have a single AddressSpace
  * @env_ptr: Pointer to subclass-specific CPUArchState field.
+ * @icount_decr_ptr: Pointer to IcountDecr field within subclass.
  * @gdb_regs: Additional GDB registers.
  * @gdb_num_regs: Number of total registers accessible to GDB.
  * @gdb_num_g_regs: Number of registers in GDB 'g' packets.
@@ -387,6 +391,7 @@ struct CPUState {
     MemoryRegion *memory;
 
     void *env_ptr; /* CPUArchState */
+    IcountDecr *icount_decr_ptr;
 
     /* Accessed in parallel; all accesses must be atomic */
     struct TranslationBlock *tb_jmp_cache[TB_JMP_CACHE_SIZE];
@@ -441,15 +446,6 @@ struct CPUState {
 
     bool ignore_memory_transaction_failures;
 
-    /* Note that this is accessed at the start of every TB via a negative
-       offset from AREG0.  Leave this field at the end so as to make the
-       (absolute value) offset as small as possible.  This reduces code
-       size, especially for hosts without large memory offsets.  */
-    union {
-        uint32_t u32;
-        icount_decr_u16 u16;
-    } icount_decr;
-
     struct hax_vcpu_state *hax_vcpu;
 
     int hvf_fd;