[PULL 0/4] tcg patch queue

[PULL 0/4] tcg patch queue

[Richard Henderson]

The following changes since commit 67e41fe0cfb62e6cdfa659f0155417d17e5274ea:

  Merge tag 'pull-ppc-20220104' of https://github.com/legoater/qemu into staging (2022-01-04 07:23:27 -0800)

are available in the Git repository at:

  https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20220104

for you to fetch changes up to d7478d4229f0a2b2817a55487e6b17081099fae4:

  common-user: Fix tail calls to safe_syscall_set_errno_tail (2022-01-04 15:41:03 -0800)

----------------------------------------------------------------
Fix for safe_syscall_base.
Fix for folding of vector add/sub.
Fix build on loongarch64 with gcc 8.
Remove decl for qemu_run_machine_init_done_notifiers.

----------------------------------------------------------------
Philippe Mathieu-Daudé (1):
      linux-user: Fix trivial build error on loongarch64 hosts

Richard Henderson (2):
      tcg/optimize: Fix folding of vector ops
      common-user: Fix tail calls to safe_syscall_set_errno_tail

Xiaoyao Li (1):
      sysemu: Cleanup qemu_run_machine_init_done_notifiers()

 include/sysemu/sysemu.h                    |  1 -
 linux-user/host/loongarch64/host-signal.h  |  4 +--
 tcg/optimize.c                             | 49 +++++++++++++++++++++++-------
 common-user/host/i386/safe-syscall.inc.S   |  1 +
 common-user/host/mips/safe-syscall.inc.S   |  1 +
 common-user/host/x86_64/safe-syscall.inc.S |  1 +
 6 files changed, 42 insertions(+), 15 deletions(-)

[PULL 1/4] tcg/optimize: Fix folding of vector ops

[Richard Henderson]

Bitwise operations are easy to fold, because the operation is
identical regardless of element size.  But add and sub need
extra element size info that is not currently propagated.

Fixes: 2f9f08ba43d
Cc: qemu-stable@nongnu.org
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/799
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 tcg/optimize.c | 49 ++++++++++++++++++++++++++++++++++++++-----------
 1 file changed, 38 insertions(+), 11 deletions(-)

diff --git a/tcg/optimize.c b/tcg/optimize.c
index 2397f2cf93..e573000951 100644
--- a/tcg/optimize.c
+++ b/tcg/optimize.c
@@ -308,13 +308,13 @@ static uint64_t do_constant_folding_2(TCGOpcode op, uint64_t x, uint64_t y)
     CASE_OP_32_64(mul):
         return x * y;
 
-    CASE_OP_32_64(and):
+    CASE_OP_32_64_VEC(and):
         return x & y;
 
-    CASE_OP_32_64(or):
+    CASE_OP_32_64_VEC(or):
         return x | y;
 
-    CASE_OP_32_64(xor):
+    CASE_OP_32_64_VEC(xor):
         return x ^ y;
 
     case INDEX_op_shl_i32:
@@ -347,16 +347,16 @@ static uint64_t do_constant_folding_2(TCGOpcode op, uint64_t x, uint64_t y)
     case INDEX_op_rotl_i64:
         return rol64(x, y & 63);
 
-    CASE_OP_32_64(not):
+    CASE_OP_32_64_VEC(not):
         return ~x;
 
     CASE_OP_32_64(neg):
         return -x;
 
-    CASE_OP_32_64(andc):
+    CASE_OP_32_64_VEC(andc):
         return x & ~y;
 
-    CASE_OP_32_64(orc):
+    CASE_OP_32_64_VEC(orc):
         return x | ~y;
 
     CASE_OP_32_64(eqv):
@@ -751,6 +751,12 @@ static bool fold_const2(OptContext *ctx, TCGOp *op)
     return false;
 }
 
+static bool fold_commutative(OptContext *ctx, TCGOp *op)
+{
+    swap_commutative(op->args[0], &op->args[1], &op->args[2]);
+    return false;
+}
+
 static bool fold_const2_commutative(OptContext *ctx, TCGOp *op)
 {
     swap_commutative(op->args[0], &op->args[1], &op->args[2]);
@@ -905,6 +911,16 @@ static bool fold_add(OptContext *ctx, TCGOp *op)
     return false;
 }
 
+/* We cannot as yet do_constant_folding with vectors. */
+static bool fold_add_vec(OptContext *ctx, TCGOp *op)
+{
+    if (fold_commutative(ctx, op) ||
+        fold_xi_to_x(ctx, op, 0)) {
+        return true;
+    }
+    return false;
+}
+
 static bool fold_addsub2(OptContext *ctx, TCGOp *op, bool add)
 {
     if (arg_is_const(op->args[2]) && arg_is_const(op->args[3]) &&
@@ -1938,10 +1954,10 @@ static bool fold_sub_to_neg(OptContext *ctx, TCGOp *op)
     return false;
 }
 
-static bool fold_sub(OptContext *ctx, TCGOp *op)
+/* We cannot as yet do_constant_folding with vectors. */
+static bool fold_sub_vec(OptContext *ctx, TCGOp *op)
 {
-    if (fold_const2(ctx, op) ||
-        fold_xx_to_i(ctx, op, 0) ||
+    if (fold_xx_to_i(ctx, op, 0) ||
         fold_xi_to_x(ctx, op, 0) ||
         fold_sub_to_neg(ctx, op)) {
         return true;
@@ -1949,6 +1965,11 @@ static bool fold_sub(OptContext *ctx, TCGOp *op)
     return false;
 }
 
+static bool fold_sub(OptContext *ctx, TCGOp *op)
+{
+    return fold_const2(ctx, op) || fold_sub_vec(ctx, op);
+}
+
 static bool fold_sub2(OptContext *ctx, TCGOp *op)
 {
     return fold_addsub2(ctx, op, false);
@@ -2052,9 +2073,12 @@ void tcg_optimize(TCGContext *s)
          * Sorted alphabetically by opcode as much as possible.
          */
         switch (opc) {
-        CASE_OP_32_64_VEC(add):
+        CASE_OP_32_64(add):
             done = fold_add(&ctx, op);
             break;
+        case INDEX_op_add_vec:
+            done = fold_add_vec(&ctx, op);
+            break;
         CASE_OP_32_64(add2):
             done = fold_add2(&ctx, op);
             break;
@@ -2193,9 +2217,12 @@ void tcg_optimize(TCGContext *s)
         CASE_OP_32_64(sextract):
             done = fold_sextract(&ctx, op);
             break;
-        CASE_OP_32_64_VEC(sub):
+        CASE_OP_32_64(sub):
             done = fold_sub(&ctx, op);
             break;
+        case INDEX_op_sub_vec:
+            done = fold_sub_vec(&ctx, op);
+            break;
         CASE_OP_32_64(sub2):
             done = fold_sub2(&ctx, op);
             break;
-- 
2.25.1

[PULL 2/4] linux-user: Fix trivial build error on loongarch64 hosts

[Richard Henderson]

From: Philippe Mathieu-Daudé <f4bug@amsat.org>

When building using GCC 8.3.0 on loongarch64 (Loongnix) we get:

  In file included from ../linux-user/signal.c:33:
  ../linux-user/host/loongarch64/host-signal.h: In function ‘host_signal_write’:
  ../linux-user/host/loongarch64/host-signal.h:57:9: error: a label can only be part of a statement and a declaration is not a statement
         uint32_t sel = (insn >> 15) & 0b11111111111;
         ^~~~~~~~

We don't use the 'sel' variable more than once, so drop it.

Meson output for the record:

  Host machine cpu family: loongarch64
  Host machine cpu: loongarch64
  C compiler for the host machine: cc (gcc 8.3.0 "cc (Loongnix 8.3.0-6.lnd.vec.27) 8.3.0")
  C linker for the host machine: cc ld.bfd 2.31.1-system

Fixes: ad812c3bd65 ("linux-user: Implement CPU-specific signal handler for loongarch64 hosts")
Reported-by: Song Gao <gaosong@loongson.cn>
Suggested-by: Song Gao <gaosong@loongson.cn>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: WANG Xuerui <git@xen0n.name>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220104215027.2180972-1-f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 linux-user/host/loongarch64/host-signal.h | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/linux-user/host/loongarch64/host-signal.h b/linux-user/host/loongarch64/host-signal.h
index 05e2c82371..7effa24251 100644
--- a/linux-user/host/loongarch64/host-signal.h
+++ b/linux-user/host/loongarch64/host-signal.h
@@ -54,9 +54,7 @@ static inline bool host_signal_write(siginfo_t *info, ucontext_t *uc)
         }
         break;
     case 0b001110: /* indexed, atomic, bounds-checking memory operations */
-        uint32_t sel = (insn >> 15) & 0b11111111111;
-
-        switch (sel) {
+        switch ((insn >> 15) & 0b11111111111) {
         case 0b00000100000: /* stx.b */
         case 0b00000101000: /* stx.h */
         case 0b00000110000: /* stx.w */
-- 
2.25.1

[PULL 3/4] sysemu: Cleanup qemu_run_machine_init_done_notifiers()

[Richard Henderson]

From: Xiaoyao Li <xiaoyao.li@intel.com>

Remove qemu_run_machine_init_done_notifiers() since no implementation
and user.

Fixes: f66dc8737c9 ("vl: move all generic initialization out of vl.c")
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220104024136.1433545-1-xiaoyao.li@intel.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 include/sysemu/sysemu.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/include/sysemu/sysemu.h b/include/sysemu/sysemu.h
index 8fae667172..b9421e03ff 100644
--- a/include/sysemu/sysemu.h
+++ b/include/sysemu/sysemu.h
@@ -16,7 +16,6 @@ extern bool qemu_uuid_set;
 void qemu_add_exit_notifier(Notifier *notify);
 void qemu_remove_exit_notifier(Notifier *notify);
 
-void qemu_run_machine_init_done_notifiers(void);
 void qemu_add_machine_init_done_notifier(Notifier *notify);
 void qemu_remove_machine_init_done_notifier(Notifier *notify);
 
-- 
2.25.1

[PULL 4/4] common-user: Fix tail calls to safe_syscall_set_errno_tail

[Richard Henderson]

For the ABIs in which the syscall return register is not
also the first function argument register, move the errno
value into the correct place.

Fixes: a3310c0397e2 ("linux-user: Move syscall error detection into safe_syscall_base")
Reported-by: Laurent Vivier <laurent@vivier.eu>
Tested-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220104190454.542225-1-richard.henderson@linaro.org>
---
 common-user/host/i386/safe-syscall.inc.S   | 1 +
 common-user/host/mips/safe-syscall.inc.S   | 1 +
 common-user/host/x86_64/safe-syscall.inc.S | 1 +
 3 files changed, 3 insertions(+)

diff --git a/common-user/host/i386/safe-syscall.inc.S b/common-user/host/i386/safe-syscall.inc.S
index baf5400a29..9c45e56e48 100644
--- a/common-user/host/i386/safe-syscall.inc.S
+++ b/common-user/host/i386/safe-syscall.inc.S
@@ -120,6 +120,7 @@ safe_syscall_end:
         pop     %ebp
         .cfi_adjust_cfa_offset -4
         .cfi_restore ebp
+        mov     %eax, (%esp)
         jmp     safe_syscall_set_errno_tail
 
         .cfi_endproc
diff --git a/common-user/host/mips/safe-syscall.inc.S b/common-user/host/mips/safe-syscall.inc.S
index fc75a337d1..6a44614970 100644
--- a/common-user/host/mips/safe-syscall.inc.S
+++ b/common-user/host/mips/safe-syscall.inc.S
@@ -141,6 +141,7 @@ safe_syscall_end:
 1:      USE_ALT_CP(t0)
         SETUP_GPX(t1)
         SETUP_GPX64(t0, t1)
+        move    a0, v0
         PTR_LA  t9, safe_syscall_set_errno_tail
         jr      t9
 
diff --git a/common-user/host/x86_64/safe-syscall.inc.S b/common-user/host/x86_64/safe-syscall.inc.S
index a20927a783..d1a67a303a 100644
--- a/common-user/host/x86_64/safe-syscall.inc.S
+++ b/common-user/host/x86_64/safe-syscall.inc.S
@@ -99,6 +99,7 @@ safe_syscall_end:
 1:      pop     %rbp
         .cfi_def_cfa_offset 8
         .cfi_restore rbp
+        mov     %eax, %edi
         jmp     safe_syscall_set_errno_tail
         .cfi_endproc
 
-- 
2.25.1

Re: [PULL 0/4] tcg patch queue

[Richard Henderson]

On 1/4/22 4:40 PM, Richard Henderson wrote:
> The following changes since commit 67e41fe0cfb62e6cdfa659f0155417d17e5274ea:
> 
>    Merge tag 'pull-ppc-20220104' of https://github.com/legoater/qemu into staging (2022-01-04 07:23:27 -0800)
> 
> are available in the Git repository at:
> 
>    https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20220104
> 
> for you to fetch changes up to d7478d4229f0a2b2817a55487e6b17081099fae4:
> 
>    common-user: Fix tail calls to safe_syscall_set_errno_tail (2022-01-04 15:41:03 -0800)
> 
> ----------------------------------------------------------------
> Fix for safe_syscall_base.
> Fix for folding of vector add/sub.
> Fix build on loongarch64 with gcc 8.
> Remove decl for qemu_run_machine_init_done_notifiers.
> 
> ----------------------------------------------------------------
> Philippe Mathieu-Daudé (1):
>        linux-user: Fix trivial build error on loongarch64 hosts
> 
> Richard Henderson (2):
>        tcg/optimize: Fix folding of vector ops
>        common-user: Fix tail calls to safe_syscall_set_errno_tail
> 
> Xiaoyao Li (1):
>        sysemu: Cleanup qemu_run_machine_init_done_notifiers()
> 
>   include/sysemu/sysemu.h                    |  1 -
>   linux-user/host/loongarch64/host-signal.h  |  4 +--
>   tcg/optimize.c                             | 49 +++++++++++++++++++++++-------
>   common-user/host/i386/safe-syscall.inc.S   |  1 +
>   common-user/host/mips/safe-syscall.inc.S   |  1 +
>   common-user/host/x86_64/safe-syscall.inc.S |  1 +
>   6 files changed, 42 insertions(+), 15 deletions(-)

Applied.

r~