[Qemu-devel] Proposal for host-utils

["J. Mayer" <l_indien@magic.fr>]

[-- Attachment #1: Type: text/plain, Size: 804 bytes --]

Following the previous discussions about host-utils implementations,
here's a patch with the following changes:
- move mulu64 and muls64 definitions from exec.h to host-utils.h, for
consistency
- include host-utils.h in more files to reflect this change
- make the optimized version of mulu64 / muls64 for amd64 hosts static
inline
- change clz64 to avoid 64 bits logical operations to optimize the 32
bits host case
- add ctz32, ctz64, cto32 and cto64, using the same method than ctlzxx /
cloxx implementations
- add ctpop8, ctpop16, ctpop32 and ctpop64, using the Sparc target
implementation method
  ctpop8 is used by the PowerPC target, I added ctpop16 for consistency
- change the Alpha and the PowerPC targets to use those helpers

Please comment.

-- 
J. Mayer <l_indien@magic.fr>
Never organized

[-- Attachment #2: host-utils.diff --]
[-- Type: text/x-patch, Size: 14689 bytes --]

Index: exec-all.h
===================================================================
RCS file: /sources/qemu/qemu/exec-all.h,v
retrieving revision 1.69
diff -u -d -d -p -r1.69 exec-all.h
--- exec-all.h	20 Oct 2007 19:45:43 -0000	1.69
+++ exec-all.h	27 Oct 2007 22:45:08 -0000
@@ -91,9 +91,6 @@ void optimize_flags_init(void);
 extern FILE *logfile;
 extern int loglevel;
 
-void muls64(int64_t *phigh, int64_t *plow, int64_t a, int64_t b);
-void mulu64(uint64_t *phigh, uint64_t *plow, uint64_t a, uint64_t b);
-
 int gen_intermediate_code(CPUState *env, struct TranslationBlock *tb);
 int gen_intermediate_code_pc(CPUState *env, struct TranslationBlock *tb);
 void dump_ops(const uint16_t *opc_buf, const uint32_t *opparam_buf);
Index: host-utils.c
===================================================================
RCS file: /sources/qemu/qemu/host-utils.c,v
retrieving revision 1.4
diff -u -d -d -p -r1.4 host-utils.c
--- host-utils.c	26 Oct 2007 22:35:01 -0000	1.4
+++ host-utils.c	27 Oct 2007 22:45:08 -0000
@@ -28,6 +28,7 @@
 //#define DEBUG_MULDIV
 
 /* Long integer helpers */
+#if !defined(__x86_64__)
 static void add128 (uint64_t *plow, uint64_t *phigh, uint64_t a, uint64_t b)
 {
     *plow += a;
@@ -69,17 +70,10 @@ static void mul64 (uint64_t *plow, uint6
     *phigh += v;
 }
 
-
 /* Unsigned 64x64 -> 128 multiplication */
 void mulu64 (uint64_t *plow, uint64_t *phigh, uint64_t a, uint64_t b)
 {
-#if defined(__x86_64__)
-    __asm__ ("mul %0\n\t"
-             : "=d" (*phigh), "=a" (*plow)
-             : "a" (a), "0" (b));
-#else
     mul64(plow, phigh, a, b);
-#endif
 #if defined(DEBUG_MULDIV)
     printf("mulu64: 0x%016llx * 0x%016llx = 0x%016llx%016llx\n",
            a, b, *phigh, *plow);
@@ -89,11 +83,6 @@ void mulu64 (uint64_t *plow, uint64_t *p
 /* Signed 64x64 -> 128 multiplication */
 void muls64 (uint64_t *plow, uint64_t *phigh, int64_t a, int64_t b)
 {
-#if defined(__x86_64__)
-    __asm__ ("imul %0\n\t"
-             : "=d" (*phigh), "=a" (*plow)
-             : "a" (a), "0" (b));
-#else
     int sa, sb;
 
     sa = (a < 0);
@@ -106,9 +95,9 @@ void muls64 (uint64_t *plow, uint64_t *p
     if (sa ^ sb) {
         neg128(plow, phigh);
     }
-#endif
 #if defined(DEBUG_MULDIV)
     printf("muls64: 0x%016llx * 0x%016llx = 0x%016llx%016llx\n",
            a, b, *phigh, *plow);
 #endif
 }
+#endif /* !defined(__x86_64__) */
Index: host-utils.h
===================================================================
RCS file: /sources/qemu/qemu/host-utils.h,v
retrieving revision 1.1
diff -u -d -d -p -r1.1 host-utils.h
--- host-utils.h	27 Oct 2007 13:05:54 -0000	1.1
+++ host-utils.h	27 Oct 2007 22:45:08 -0000
@@ -23,6 +23,26 @@
  * THE SOFTWARE.
  */
 
+#if defined(__x86_64__)
+static always_inline void mulu64 (uint64_t *plow, uint64_t *phigh,
+                                  uint64_t a, uint64_t b)
+{
+    __asm__ ("mul %0\n\t"
+             : "=d" (*phigh), "=a" (*plow)
+             : "a" (a), "0" (b));
+}
+static always_inline void muls64 (uint64_t *plow, uint64_t *phigh,
+                                  int64_t a, int64_t b)
+{
+    __asm__ ("imul %0\n\t"
+             : "=d" (*phigh), "=a" (*plow)
+             : "a" (a), "0" (b));
+}
+#else
+void muls64(int64_t *phigh, int64_t *plow, int64_t a, int64_t b);
+void mulu64(uint64_t *phigh, uint64_t *plow, uint64_t a, uint64_t b);
+#endif
+
 /* Note that some of those functions may end up calling libgcc functions,
    depending on the host machine. It is up to the target emulation to
    cope with that. */
@@ -68,37 +88,113 @@ static always_inline int clz64(uint64_t 
 {
     int cnt = 0;
 
-    if (!(val & 0xFFFFFFFF00000000ULL)) {
+    if (!(val >> 32)) {
         cnt += 32;
-        val <<= 32;
+    } else {
+        val >>= 32;
     }
-    if (!(val & 0xFFFF000000000000ULL)) {
+
+    return cnt + clz32(val);
+}
+
+static always_inline int clo64(uint64_t val)
+{
+    return clz64(~val);
+}
+
+static always_inline int ctz32 (uint32_t val)
+{
+    int cnt;
+
+    cnt = 0;
+    if (!(val & 0x0000FFFFUL)) {
         cnt += 16;
-        val <<= 16;
+        val >>= 16;
     }
-    if (!(val & 0xFF00000000000000ULL)) {
+    if (!(val & 0x000000FFUL)) {
         cnt += 8;
-        val <<= 8;
+        val >>= 8;
     }
-    if (!(val & 0xF000000000000000ULL)) {
+    if (!(val & 0x0000000FUL)) {
         cnt += 4;
-        val <<= 4;
+        val >>= 4;
     }
-    if (!(val & 0xC000000000000000ULL)) {
+    if (!(val & 0x00000003UL)) {
         cnt += 2;
-        val <<= 2;
+        val >>= 2;
     }
-    if (!(val & 0x8000000000000000ULL)) {
+    if (!(val & 0x00000001UL)) {
         cnt++;
-        val <<= 1;
+        val >>= 1;
     }
-    if (!(val & 0x8000000000000000ULL)) {
+    if (!(val & 0x00000001UL)) {
         cnt++;
     }
+
     return cnt;
 }
 
-static always_inline int clo64(uint64_t val)
+static always_inline int cto32 (uint32_t val)
 {
-    return clz64(~val);
+    return ctz32(~val);
+}
+
+static always_inline int ctz64 (uint64_t val)
+{
+    int cnt;
+
+    cnt = 0;
+    if (!((uint32_t)val)) {
+        cnt += 32;
+        val >>= 32;
+    }
+
+    return cnt + ctz32(val);
+}
+
+static always_inline int cto64 (uint64_t val)
+{
+    return ctz64(~val);
+}
+
+static always_inline int ctpop8 (uint8_t val)
+{
+    val = (val & 0x55) + ((val >> 1) & 0x55);
+    val = (val & 0x33) + ((val >> 2) & 0x33);
+    val = (val & 0x0f) + ((val >> 4) & 0x0f);
+
+    return val;
+}
+
+static always_inline int ctpop16 (uint16_t val)
+{
+    val = (val & 0x5555) + ((val >> 1) & 0x5555);
+    val = (val & 0x3333) + ((val >> 2) & 0x3333);
+    val = (val & 0x0f0f) + ((val >> 4) & 0x0f0f);
+    val = (val & 0x00ff) + ((val >> 8) & 0x00ff);
+
+    return val;
+}
+
+static always_inline int ctpop32 (uint32_t val)
+{
+    val = (val & 0x55555555) + ((val >>  1) & 0x55555555);
+    val = (val & 0x33333333) + ((val >>  2) & 0x33333333);
+    val = (val & 0x0f0f0f0f) + ((val >>  4) & 0x0f0f0f0f);
+    val = (val & 0x00ff00ff) + ((val >>  8) & 0x00ff00ff);
+    val = (val & 0x0000ffff) + ((val >> 16) & 0x0000ffff);
+
+    return val;
+}
+
+static always_inline int ctpop64 (uint64_t val)
+{
+    val = (val & 0x5555555555555555ULL) + ((val >>  1) & 0x5555555555555555ULL);
+    val = (val & 0x3333333333333333ULL) + ((val >>  2) & 0x3333333333333333ULL);
+    val = (val & 0x0f0f0f0f0f0f0f0fULL) + ((val >>  4) & 0x0f0f0f0f0f0f0f0fULL);
+    val = (val & 0x00ff00ff00ff00ffULL) + ((val >>  8) & 0x00ff00ff00ff00ffULL);
+    val = (val & 0x0000ffff0000ffffULL) + ((val >> 16) & 0x0000ffff0000ffffULL);
+    val = (val & 0x00000000ffffffffULL) + ((val >> 32) & 0x00000000ffffffffULL);
+
+    return val;
 }
Index: target-alpha/op.c
===================================================================
RCS file: /sources/qemu/qemu/target-alpha/op.c,v
retrieving revision 1.4
diff -u -d -d -p -r1.4 op.c
--- target-alpha/op.c	25 Oct 2007 23:34:44 -0000	1.4
+++ target-alpha/op.c	27 Oct 2007 22:45:08 -0000
@@ -22,6 +22,7 @@
 
 #include "config.h"
 #include "exec.h"
+#include "host-utils.h"
 
 #include "op_helper.h"
 
Index: target-alpha/op_helper.c
===================================================================
RCS file: /sources/qemu/qemu/target-alpha/op_helper.c,v
retrieving revision 1.5
diff -u -d -d -p -r1.5 op_helper.c
--- target-alpha/op_helper.c	25 Oct 2007 23:34:44 -0000	1.5
+++ target-alpha/op_helper.c	27 Oct 2007 22:45:08 -0000
@@ -19,6 +19,7 @@
  */
 
 #include "exec.h"
+#include "host-utils.h"
 #include "softfloat.h"
 
 #include "op_helper.h"
@@ -211,87 +212,17 @@ void helper_mulqv ()
 
 void helper_ctpop (void)
 {
-    int n;
-
-    for (n = 0; T0 != 0; n++)
-        T0 = T0 ^ (T0 - 1);
-    T0 = n;
+    T0 = ctpop64(T0);
 }
 
 void helper_ctlz (void)
 {
-    uint32_t op32;
-    int n;
-
-    n = 0;
-    if (!(T0 & 0xFFFFFFFF00000000ULL)) {
-        n += 32;
-        T0 <<= 32;
-    }
-    /* Make it easier for 32 bits hosts */
-    op32 = T0 >> 32;
-    if (!(op32 & 0xFFFF0000UL)) {
-        n += 16;
-        op32 <<= 16;
-    }
-    if (!(op32 & 0xFF000000UL)) {
-        n += 8;
-        op32 <<= 8;
-    }
-    if (!(op32 & 0xF0000000UL)) {
-        n += 4;
-        op32 <<= 4;
-    }
-    if (!(op32 & 0xC0000000UL)) {
-        n += 2;
-        op32 <<= 2;
-    }
-    if (!(op32 & 0x80000000UL)) {
-        n++;
-        op32 <<= 1;
-    }
-    if (!(op32 & 0x80000000UL)) {
-        n++;
-    }
-    T0 = n;
+    T0 = clz64(T0);
 }
 
 void helper_cttz (void)
 {
-    uint32_t op32;
-    int n;
-
-    n = 0;
-    if (!(T0 & 0x00000000FFFFFFFFULL)) {
-        n += 32;
-        T0 >>= 32;
-    }
-    /* Make it easier for 32 bits hosts */
-    op32 = T0;
-    if (!(op32 & 0x0000FFFFUL)) {
-        n += 16;
-        op32 >>= 16;
-    }
-    if (!(op32 & 0x000000FFUL)) {
-        n += 8;
-        op32 >>= 8;
-    }
-    if (!(op32 & 0x0000000FUL)) {
-        n += 4;
-        op32 >>= 4;
-    }
-    if (!(op32 & 0x00000003UL)) {
-        n += 2;
-        op32 >>= 2;
-    }
-    if (!(op32 & 0x00000001UL)) {
-        n++;
-        op32 >>= 1;
-    }
-    if (!(op32 & 0x00000001UL)) {
-        n++;
-    }
-    T0 = n;
+    T0 = clz32(T0);
 }
 
 static inline uint64_t byte_zap (uint64_t op, uint8_t mskb)
Index: target-i386/helper.c
===================================================================
RCS file: /sources/qemu/qemu/target-i386/helper.c,v
retrieving revision 1.91
diff -u -d -d -p -r1.91 helper.c
--- target-i386/helper.c	26 Oct 2007 22:35:02 -0000	1.91
+++ target-i386/helper.c	27 Oct 2007 22:45:09 -0000
@@ -18,6 +18,7 @@
  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
 #include "exec.h"
+#include "host-utils.h"
 
 //#define DEBUG_PCALL
 
Index: target-ppc/op.c
===================================================================
RCS file: /sources/qemu/qemu/target-ppc/op.c,v
retrieving revision 1.59
diff -u -d -d -p -r1.59 op.c
--- target-ppc/op.c	27 Oct 2007 17:54:30 -0000	1.59
+++ target-ppc/op.c	27 Oct 2007 22:45:09 -0000
@@ -22,6 +22,7 @@
 
 #include "config.h"
 #include "exec.h"
+#include "host-utils.h"
 #include "helper_regs.h"
 #include "op_helper.h"
 
@@ -1508,14 +1509,14 @@ void OPPROTO op_andi_T1_64 (void)
 /* count leading zero */
 void OPPROTO op_cntlzw (void)
 {
-    T0 = _do_cntlzw(T0);
+    do_cntlzw();
     RETURN();
 }
 
 #if defined(TARGET_PPC64)
 void OPPROTO op_cntlzd (void)
 {
-    T0 = _do_cntlzd(T0);
+    do_cntlzd();
     RETURN();
 }
 #endif
Index: target-ppc/op_helper.c
===================================================================
RCS file: /sources/qemu/qemu/target-ppc/op_helper.c,v
retrieving revision 1.56
diff -u -d -d -p -r1.56 op_helper.c
--- target-ppc/op_helper.c	27 Oct 2007 17:59:46 -0000	1.56
+++ target-ppc/op_helper.c	27 Oct 2007 22:45:09 -0000
@@ -18,6 +18,7 @@
  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
 #include "exec.h"
+#include "host-utils.h"
 
 #include "helper_regs.h"
 #include "op_helper.h"
@@ -381,6 +382,18 @@ void do_subfzeo_64 (void)
 }
 #endif
 
+void do_cntlzw (void)
+{
+    T0 = clz32(T0);
+}
+
+#if defined(TARGET_PPC64)
+void do_cntlzd (void)
+{
+    T0 = clz64(T0);
+}
+#endif
+
 /* shift right arithmetic helper */
 void do_sraw (void)
 {
@@ -438,16 +451,6 @@ void do_srad (void)
 }
 #endif
 
-static always_inline int popcnt (uint32_t val)
-{
-    int i;
-
-    for (i = 0; val != 0;)
-        val = val ^ (val - 1);
-
-    return i;
-}
-
 void do_popcntb (void)
 {
     uint32_t ret;
@@ -455,7 +458,7 @@ void do_popcntb (void)
 
     ret = 0;
     for (i = 0; i < 32; i += 8)
-        ret |= popcnt((T0 >> i) & 0xFF) << i;
+        ret |= ctpop8((T0 >> i) & 0xFF) << i;
     T0 = ret;
 }
 
@@ -467,7 +470,7 @@ void do_popcntb_64 (void)
 
     ret = 0;
     for (i = 0; i < 64; i += 8)
-        ret |= popcnt((T0 >> i) & 0xFF) << i;
+        ret |= ctpop8((T0 >> i) & 0xFF) << i;
     T0 = ret;
 }
 #endif
@@ -1925,14 +1928,14 @@ static always_inline uint32_t _do_eaddw 
 static always_inline int _do_ecntlsw (uint32_t val)
 {
     if (val & 0x80000000)
-        return _do_cntlzw(~val);
+        return clz32(~val);
     else
-        return _do_cntlzw(val);
+        return clz32(val);
 }
 
 static always_inline int _do_ecntlzw (uint32_t val)
 {
-    return _do_cntlzw(val);
+    return clz32(val);
 }
 
 static always_inline uint32_t _do_eneg (uint32_t val)
Index: target-ppc/op_helper.h
===================================================================
RCS file: /sources/qemu/qemu/target-ppc/op_helper.h,v
retrieving revision 1.24
diff -u -d -d -p -r1.24 op_helper.h
--- target-ppc/op_helper.h	27 Oct 2007 17:54:30 -0000	1.24
+++ target-ppc/op_helper.h	27 Oct 2007 22:45:09 -0000
@@ -75,6 +75,10 @@ void do_nego (void);
 void do_subfe (void);
 void do_subfmeo (void);
 void do_subfzeo (void);
+void do_cntlzw (void);
+#if defined(TARGET_PPC64)
+void do_cntlzd (void);
+#endif
 void do_sraw (void);
 #if defined(TARGET_PPC64)
 void do_adde_64 (void);
@@ -285,78 +289,6 @@ void do_evfsctsiz (void);
 void do_evfsctuiz (void);
 #endif /* defined(TARGET_PPCEMB) */
 
-/* Inlined helpers: used in micro-operation as well as helpers */
-/* Generic fixed-point helpers */
-static always_inline int _do_cntlzw (uint32_t val)
-{
-    int cnt = 0;
-    if (!(val & 0xFFFF0000UL)) {
-        cnt += 16;
-        val <<= 16;
-    }
-    if (!(val & 0xFF000000UL)) {
-        cnt += 8;
-        val <<= 8;
-    }
-    if (!(val & 0xF0000000UL)) {
-        cnt += 4;
-        val <<= 4;
-    }
-    if (!(val & 0xC0000000UL)) {
-        cnt += 2;
-        val <<= 2;
-    }
-    if (!(val & 0x80000000UL)) {
-        cnt++;
-        val <<= 1;
-    }
-    if (!(val & 0x80000000UL)) {
-        cnt++;
-    }
-    return cnt;
-}
-
-static always_inline int _do_cntlzd (uint64_t val)
-{
-    int cnt = 0;
-#if HOST_LONG_BITS == 64
-    if (!(val & 0xFFFFFFFF00000000ULL)) {
-        cnt += 32;
-        val <<= 32;
-    }
-    if (!(val & 0xFFFF000000000000ULL)) {
-        cnt += 16;
-        val <<= 16;
-    }
-    if (!(val & 0xFF00000000000000ULL)) {
-        cnt += 8;
-        val <<= 8;
-    }
-    if (!(val & 0xF000000000000000ULL)) {
-        cnt += 4;
-        val <<= 4;
-    }
-    if (!(val & 0xC000000000000000ULL)) {
-        cnt += 2;
-        val <<= 2;
-    }
-    if (!(val & 0x8000000000000000ULL)) {
-        cnt++;
-        val <<= 1;
-    }
-    if (!(val & 0x8000000000000000ULL)) {
-        cnt++;
-    }
-#else
-    /* Make it easier on 32 bits host machines */
-    if (!(val >> 32))
-        cnt = _do_cntlzw(val) + 32;
-    else
-        cnt = _do_cntlzw(val >> 32);
-#endif
-    return cnt;
-}
-
 #if defined(TARGET_PPCEMB)
 /* SPE extension */
 /* Single precision floating-point helpers */