[PATCH] x86/bitops: Replace __scanbit() with a simpler tzcnt()

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

There are 4 remaining users of __scanbit(), and they're capable of dealing
with max being BITS_PER_LONG.  Switch to a simpler tzcnt() helper, which also
removed some alternatives from behind the scenes.

No functional change.

Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
---
CC: Jan Beulich <jbeulich@xxxxxxxx>
CC: Roger Pau Monné <roger@xxxxxxxxxxxxxx>
CC: Teddy Astie <teddy.astie@xxxxxxxxxx>

Supercedes "x86/bitops: Improve code generation for __scanbit()"

Bloat-o-meter shows rather less churn:

  add/remove: 1/0 grow/shrink: 9/18 up/down: 244/-279 (-35)
  Function                                     old     new   delta
  is_cpu_primary                                 -      88     +88
  vmap_size                                     38      91     +53
  get_outstanding_claims                       198     214     +16
  csched2_unit_insert                          413     429     +16
  csched2_schedule                            5371    5387     +16
  __cpu_disable                                583     599     +16
  memory_add                                  4036    4049     +13
  csched2_alloc_domdata                        476     487     +11
  numa_init_array                              153     161      +8
  dump_execstate                               221     228      +7
  dump_numa                                    471     469      -2
  csched2_res_pick                            1351    1349      -2
  csched_schedule                             3790    3786      -4
  fixup_irqs                                  1655    1650      -5
  dump_numa.cold                               977     970      -7
  cpupool_assign_cpu_locked                    410     403      -7
  _assign_irq_vector                          1571    1564      -7
  vmx_sync_exit_bitmap                          73      65      -8
  numa_emulation                               487     479      -8
  __find_next_zero_bit                         145     137      -8
  get_free_buddy                              1014    1005      -9
  runq_tickle                                 1753    1737     -16
  node_to_scrub                                533     517     -16
  dom0_compute_nr_pages                       1424    1408     -16
  sched_init_vcpu                             1811    1791     -20
  microcode_nmi_callback                       300     260     -40
  heap_init_late                              1659    1611     -48
  do_microcode_update                         1014     958     -56

  Total: Before=2013616, After=2013581, chg -0.00%
  add/remove: 0/0 grow/shrink: 3/0 up/down: 36/0 (36)
  Data                                         old     new   delta
  symbols_names                             134797  134817     +20
  symbols_sorted_offsets                     69368   69376      +8
  symbols_offsets                            53456   53464      +8

For things like vmap_size() what has happened is that vm_size() has now been 
inlined.
---
 xen/arch/x86/bitops.c             |  4 ++--
 xen/arch/x86/include/asm/bitops.h | 31 ++++++++++++++++++-------------
 2 files changed, 20 insertions(+), 15 deletions(-)

diff --git a/xen/arch/x86/bitops.c b/xen/arch/x86/bitops.c
index 22cfb6d8ec3a..348e51202465 100644
--- a/xen/arch/x86/bitops.c
+++ b/xen/arch/x86/bitops.c
@@ -35,7 +35,7 @@ unsigned int __find_next_bit(
     if ( bit != 0 )
     {
         /* Look for a bit in the first word. */
-        set = __scanbit(*p >> bit, BITS_PER_LONG);
+        set = tzcnt(*p >> bit);
         if ( set < BITS_PER_LONG )
             return (offset + set);
         offset += BITS_PER_LONG - bit;
@@ -85,7 +85,7 @@ unsigned int __find_next_zero_bit(
     if ( bit != 0 )
     {
         /* Look for zero in the first word. */
-        set = __scanbit(~*p >> bit, BITS_PER_LONG);
+        set = tzcnt(~*p >> bit);
         if ( set < BITS_PER_LONG )
             return (offset + set);
         offset += BITS_PER_LONG - bit;
diff --git a/xen/arch/x86/include/asm/bitops.h 
b/xen/arch/x86/include/asm/bitops.h
index 557cee03baeb..f880d16a982f 100644
--- a/xen/arch/x86/include/asm/bitops.h
+++ b/xen/arch/x86/include/asm/bitops.h
@@ -320,18 +320,23 @@ extern unsigned int __find_first_zero_bit(
 extern unsigned int __find_next_zero_bit(
     const unsigned long *addr, unsigned int size, unsigned int offset);
 
-static always_inline unsigned int __scanbit(unsigned long val, unsigned int 
max)
+/*
+ * Search for the first set bit in @x, returning BITS_PER_LONG if @x was 0.
+ * See arch_ffs() for safety discussions.
+ */
+static always_inline unsigned int tzcnt(unsigned long x)
 {
-    if ( __builtin_constant_p(max) && max == BITS_PER_LONG )
-        alternative_io("bsf %[in],%[out]; cmovz %[max],%k[out]",
-                       "rep; bsf %[in],%[out]",
-                       X86_FEATURE_BMI1,
-                       [out] "=&r" (val),
-                       [in] "r" (val), [max] "r" (max));
-    else
-        asm ( "bsf %1,%0 ; cmovz %2,%k0"
-              : "=&r" (val) : "r" (val), "r" (max) );
-    return (unsigned int)val;
+    unsigned int r;
+
+    /*
+     * The destination register still needs pre-loading for older CPUs which
+     * will execute TZCNT as BSF.
+     */
+    asm ( "tzcnt %[x], %q[r]"
+          : [r] "=r" (r)
+          : [x] "rm" (x), "[r]" (BITS_PER_LONG) );
+
+    return r;
 }
 
 /**
@@ -358,7 +363,7 @@ static always_inline unsigned int __scanbit(unsigned long 
val, unsigned int max)
     if ( o__ >= s__ )                                                       \
         r__ = s__;                                                          \
     else if ( __builtin_constant_p(size) && s__ <= BITS_PER_LONG )          \
-        r__ = o__ + __scanbit(*a__ >> o__, s__);                            \
+        r__ = o__ + tzcnt(*a__ >> o__);                                     \
     else if ( __builtin_constant_p(off) && !o__ )                           \
         r__ = __find_first_bit(a__, s__);                                   \
     else                                                                    \
@@ -390,7 +395,7 @@ static always_inline unsigned int __scanbit(unsigned long 
val, unsigned int max)
     if ( o__ >= s__ )                                                       \
         r__ = s__;                                                          \
     else if ( __builtin_constant_p(size) && s__ <= BITS_PER_LONG )          \
-        r__ = o__ + __scanbit(~*a__ >> o__, s__);                           \
+        r__ = o__ + tzcnt(~*a__ >> o__);                                    \
     else if ( __builtin_constant_p(off) && !o__ )                           \
         r__ = __find_first_zero_bit(a__, s__);                              \
     else                                                                    \

base-commit: 204e45847d5a743b611de989835f651fcfd6ddc0
-- 
2.39.5