Re: [Qemu-devel] [RFC PATCH v1 3/4] translate-all.c: Compute L1 page table properties at runtime

[Paolo Bonzini]

On 13/06/2016 11:08, address@hidden wrote:
> From: Vijaya Kumar K <address@hidden>
> 
> Remove L1 page mapping table properties computing
> statically using macros which is dependent on
> TARGET_PAGE_BITS. Drop macros V_L1_SIZE, V_L1_SHIFT,
> V_L1_BITS macros and replace with variables which are
> computed at early stage of VM boot.
> 
> Removing dependency can help to make TARGET_PAGE_BITS
> dynamic.
> 
> Signed-off-by: Vijaya Kumar K <address@hidden>
> ---
>  include/qemu-common.h |    1 +
>  translate-all.c       |   57 
> ++++++++++++++++++++++++++++++-------------------
>  vl.c                  |    3 +++
>  3 files changed, 39 insertions(+), 22 deletions(-)
> 
> diff --git a/include/qemu-common.h b/include/qemu-common.h
> index 1f2cb94..d5f0450 100644
> --- a/include/qemu-common.h
> +++ b/include/qemu-common.h
> @@ -129,6 +129,7 @@ int parse_debug_env(const char *name, int max, int 
> initial);
>  
>  const char *qemu_ether_ntoa(const MACAddr *mac);
>  void page_size_init(void);
> +void init_l1_page_table_param(void);
>  
>  /* returns non-zero if dump is in progress, otherwise zero is
>   * returned. */
> diff --git a/translate-all.c b/translate-all.c
> index 118e7d3..a580ca9 100644
> --- a/translate-all.c
> +++ b/translate-all.c
> @@ -57,6 +57,7 @@
>  #include "qemu/bitmap.h"
>  #include "qemu/timer.h"
>  #include "exec/log.h"
> +#include "qemu/error-report.h"
>  
>  //#define DEBUG_TB_INVALIDATE
>  //#define DEBUG_FLUSH
> @@ -99,25 +100,18 @@ typedef struct PageDesc {
>  #define V_L2_BITS 10
>  #define V_L2_SIZE (1 << V_L2_BITS)
>  
> -/* The bits remaining after N lower levels of page tables.  */
> -#define V_L1_BITS_REM \
> -    ((L1_MAP_ADDR_SPACE_BITS - TARGET_PAGE_BITS) % V_L2_BITS)
> -
> -#if V_L1_BITS_REM < 4
> -#define V_L1_BITS  (V_L1_BITS_REM + V_L2_BITS)
> -#else
> -#define V_L1_BITS  V_L1_BITS_REM
> -#endif
> -
> -#define V_L1_SIZE  ((target_ulong)1 << V_L1_BITS)
> -
> -#define V_L1_SHIFT (L1_MAP_ADDR_SPACE_BITS - TARGET_PAGE_BITS - V_L1_BITS)
> -
>  uintptr_t qemu_host_page_size;
>  intptr_t qemu_host_page_mask;
>  
> +/*
> + * L1 Mapping properties
> + */
> +static unsigned long v_l1_bits;
> +static unsigned long v_l1_size;
> +static unsigned long v_l1_shift;

Please make these uint8_t.

>  /* The bottom level has pointers to PageDesc */
> -static void *l1_map[V_L1_SIZE];
> +static void *l1_map;

You can make this array have a static V_L2_SIZE * 16 size too.  Peter,
what do you think?

>  /* code generation context */
>  TCGContext tcg_ctx;
> @@ -127,6 +121,25 @@ TCGContext tcg_ctx;
>  __thread int have_tb_lock;
>  #endif
>  
> +void init_l1_page_table_param(void)
> +{
> +    uint32_t v_l1_bits_rem;
> +
> +    assert(TARGET_PAGE_BITS);
> +    /* The bits remaining after N lower levels of page tables.  */
> +    v_l1_bits_rem = ((L1_MAP_ADDR_SPACE_BITS - TARGET_PAGE_BITS) % 
> V_L2_BITS);
> +    if (v_l1_bits_rem < 4)
> +        v_l1_bits = (v_l1_bits_rem + V_L2_BITS);
> +    else
> +        v_l1_bits = v_l1_bits_rem;
> +
> +    v_l1_size = ((target_ulong)1 << v_l1_bits);
> +    v_l1_shift = (L1_MAP_ADDR_SPACE_BITS - TARGET_PAGE_BITS - v_l1_bits);

Please assert that v_l1_shift % V_L2_BITS == 0.

> +    l1_map = g_malloc0(v_l1_size * sizeof(void *));
> +    if (!l1_map)
> +        error_report("Allocation faile for L1 MAP table\n");
> +}
> +
>  void tb_lock(void)
>  {
>  #ifdef CONFIG_USER_ONLY
> @@ -408,10 +421,10 @@ static PageDesc *page_find_alloc(tb_page_addr_t index, 
> int alloc)
>      int i;
>  
>      /* Level 1.  Always allocated.  */
> -    lp = l1_map + ((index >> V_L1_SHIFT) & (V_L1_SIZE - 1));
> +    lp = l1_map + ((index >> v_l1_shift) & (v_l1_size - 1));
>  
>      /* Level 2..N-1.  */
> -    for (i = V_L1_SHIFT / V_L2_BITS - 1; i > 0; i--) {
> +    for (i = v_l1_shift / V_L2_BITS - 1; i > 0; i--) {

Please cache v_l1_shift / V_L2_BITS - 1 into a new variable v_l2_levels.

>          void **p = atomic_rcu_read(lp);
>  
>          if (p == NULL) {
> @@ -819,8 +832,8 @@ static void page_flush_tb(void)
>  {
>      int i;
>  
> -    for (i = 0; i < V_L1_SIZE; i++) {
> -        page_flush_tb_1(V_L1_SHIFT / V_L2_BITS - 1, l1_map + i);
> +    for (i = 0; i < v_l1_size; i++) {
> +        page_flush_tb_1(v_l1_shift / V_L2_BITS - 1, l1_map + i);
>      }
>  }
>  
> @@ -1825,9 +1838,9 @@ int walk_memory_regions(void *priv, 
> walk_memory_regions_fn fn)
>      data.start = -1u;
>      data.prot = 0;
>  
> -    for (i = 0; i < V_L1_SIZE; i++) {
> -        int rc = walk_memory_regions_1(&data, (target_ulong)i << (V_L1_SHIFT 
> + TARGET_PAGE_BITS),
> -                                       V_L1_SHIFT / V_L2_BITS - 1, l1_map + 
> i);
> +    for (i = 0; i < v_l1_size; i++) {
> +        int rc = walk_memory_regions_1(&data, (target_ulong)i << (v_l1_shift 
> + TARGET_PAGE_BITS),
> +                                       v_l1_shift / V_L2_BITS - 1, l1_map + 
> i);
>          if (rc != 0) {
>              return rc;
>          }
> diff --git a/vl.c b/vl.c
> index b0bcc25..b6da265 100644
> --- a/vl.c
> +++ b/vl.c
> @@ -4044,6 +4044,9 @@ int main(int argc, char **argv, char **envp)
>      }
>      object_property_add_child(object_get_root(), "machine",
>                                OBJECT(current_machine), &error_abort);
> +
> +    init_l1_page_table_param();

Please call this from cpu_exec_init_all instead.

Thanks,

Paolo

>      cpu_exec_init_all();
>  
>      if (machine_class->hw_version) {
>