Pretty small still, but there are two patches that ought

to get backported to stable, so no point in delaying.

r~

The following changes since commit a5ba0a7e4e150d1350a041f0d0ef9ca6c8d7c307:

Merge tag 'pull-aspeed-20241211' of https://github.com/legoater/qemu into staging (2024-12-11 15:16:47 +0000)

https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20241212

for you to fetch changes up to 7ac87b14a92234b6a89b701b4043ad6cf8bdcccf:

target/sparc: Use memcpy() and remove memcpy32() (2024-12-12 14:28:38 -0600)

tcg: Reset free_temps before tcg_optimize

tcg/riscv: Fix StoreStore barrier generation

include/exec: Introduce fpst alias in helper-head.h.inc

target/sparc: Use memcpy() and remove memcpy32()

Philippe Mathieu-Daudé (1):

target/sparc: Use memcpy() and remove memcpy32()

Richard Henderson (2):

tcg: Reset free_temps before tcg_optimize

include/exec: Introduce fpst alias in helper-head.h.inc

Roman Artemev (1):

tcg/riscv: Fix StoreStore barrier generation

include/tcg/tcg-temp-internal.h | 6 ++++++

accel/tcg/plugin-gen.c | 2 +-

target/sparc/win_helper.c | 26 ++++++++------------------

tcg/tcg.c | 5 ++++-

include/exec/helper-head.h.inc | 3 +++

tcg/riscv/tcg-target.c.inc | 2 +-

6 files changed, 23 insertions(+), 21 deletions(-)

When allocating new temps during tcg_optmize, do not re-use

any EBB temps that were used within the TB. We do not have

any idea what span of the TB in which the temp was live.

Introduce tcg_temp_ebb_reset_freed and use before tcg_optimize,

as well as replacing the equivalent in plugin_gen_inject and

tcg_func_start.

Cc: qemu-stable@nongnu.org

Fixes: fb04ab7ddd8 ("tcg/optimize: Lower TCG_COND_TST{EQ,NE} if unsupported")

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2711

Reported-by: wannacu <wannacu2049@gmail.com>

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>

---

include/tcg/tcg-temp-internal.h | 6 ++++++

accel/tcg/plugin-gen.c | 2 +-

tcg/tcg.c | 5 ++++-

3 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/include/tcg/tcg-temp-internal.h b/include/tcg/tcg-temp-internal.h

--- a/include/tcg/tcg-temp-internal.h

+++ b/include/tcg/tcg-temp-internal.h

@@ -XXX,XX +XXX,XX @@ TCGv_i64 tcg_temp_ebb_new_i64(void);

TCGv_ptr tcg_temp_ebb_new_ptr(void);

TCGv_i128 tcg_temp_ebb_new_i128(void);

+/* Forget all freed EBB temps, so that new allocations produce new temps. */

+static inline void tcg_temp_ebb_reset_freed(TCGContext *s)

+{

+ memset(s->free_temps, 0, sizeof(s->free_temps));

+}

#endif /* TCG_TEMP_FREE_H */

diff --git a/accel/tcg/plugin-gen.c b/accel/tcg/plugin-gen.c

index XXXXXXX..XXXXXXX 100644

--- a/accel/tcg/plugin-gen.c

+++ b/accel/tcg/plugin-gen.c

@@ -XXX,XX +XXX,XX @@ static void plugin_gen_inject(struct qemu_plugin_tb *plugin_tb)

* that might be live within the existing opcode stream.

* The simplest solution is to release them all and create new.

*/

- memset(tcg_ctx->free_temps, 0, sizeof(tcg_ctx->free_temps));

+ tcg_temp_ebb_reset_freed(tcg_ctx);

QTAILQ_FOREACH_SAFE(op, &tcg_ctx->ops, link, next) {

switch (op->opc) {

diff --git a/tcg/tcg.c b/tcg/tcg.c

index XXXXXXX..XXXXXXX 100644

--- a/tcg/tcg.c

+++ b/tcg/tcg.c

@@ -XXX,XX +XXX,XX @@ void tcg_func_start(TCGContext *s)

s->nb_temps = s->nb_globals;

/* No temps have been previously allocated for size or locality. */

- memset(s->free_temps, 0, sizeof(s->free_temps));

+ tcg_temp_ebb_reset_freed(s);

/* No constant temps have been previously allocated. */

for (int i = 0; i < TCG_TYPE_COUNT; ++i) {

@@ -XXX,XX +XXX,XX @@ int tcg_gen_code(TCGContext *s, TranslationBlock *tb, uint64_t pc_start)

}

#endif

+ /* Do not reuse any EBB that may be allocated within the TB. */

+ tcg_temp_ebb_reset_freed(s);

tcg_optimize(s);

reachable_code_pass(s);

2.43.0

From: Roman Artemev <roman.artemev@syntacore.com>

On RISC-V to StoreStore barrier corresponds

`fence w, w` not `fence r, r`

Cc: qemu-stable@nongnu.org

Fixes: efbea94c76b ("tcg/riscv: Add slowpath load and store instructions")

Signed-off-by: Denis Tomashev <denis.tomashev@syntacore.com>

Signed-off-by: Roman Artemev <roman.artemev@syntacore.com>

Message-ID: <e2f2131e294a49e79959d4fa9ec02cf4@syntacore.com>

tcg/riscv/tcg-target.c.inc | 2 +-

1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc

--- a/tcg/riscv/tcg-target.c.inc

+++ b/tcg/riscv/tcg-target.c.inc

@@ -XXX,XX +XXX,XX @@ static void tcg_out_mb(TCGContext *s, TCGArg a0)

insn |= 0x02100000;

if (a0 & TCG_MO_ST_ST) {

- insn |= 0x02200000;

+ insn |= 0x01100000;

tcg_out32(s, insn);

2.43.0

From: Philippe Mathieu-Daudé <philmd@linaro.org>

Rather than manually copying each register, use

the libc memcpy(), which is well optimized nowadays.

Suggested-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

Message-ID: <20241205205418.67613-1-philmd@linaro.org>

target/sparc/win_helper.c | 26 ++++++++------------------

1 file changed, 8 insertions(+), 18 deletions(-)

diff --git a/target/sparc/win_helper.c b/target/sparc/win_helper.c

--- a/target/sparc/win_helper.c

+++ b/target/sparc/win_helper.c

@@ -XXX,XX +XXX,XX @@

#include "exec/helper-proto.h"

#include "trace.h"

-static inline void memcpy32(target_ulong *dst, const target_ulong *src)

-{

- dst[0] = src[0];

- dst[1] = src[1];

- dst[2] = src[2];

- dst[3] = src[3];

- dst[4] = src[4];

- dst[5] = src[5];

- dst[6] = src[6];

- dst[7] = src[7];

-}

-

void cpu_set_cwp(CPUSPARCState *env, int new_cwp)

/* put the modified wrap registers at their proper location */

if (env->cwp == env->nwindows - 1) {

- memcpy32(env->regbase, env->regbase + env->nwindows * 16);

+ memcpy(env->regbase, env->regbase + env->nwindows * 16,

+ sizeof(env->gregs));

}

env->cwp = new_cwp;

/* put the wrap registers at their temporary location */

if (new_cwp == env->nwindows - 1) {

- memcpy32(env->regbase + env->nwindows * 16, env->regbase);

+ memcpy(env->regbase + env->nwindows * 16, env->regbase,

+ sizeof(env->gregs));

}

env->regwptr = env->regbase + (new_cwp * 16);

@@ -XXX,XX +XXX,XX @@ void cpu_gl_switch_gregs(CPUSPARCState *env, uint32_t new_gl)

dst = get_gl_gregset(env, env->gl);

if (src != dst) {

- memcpy32(dst, env->gregs);

- memcpy32(env->gregs, src);

+ memcpy(dst, env->gregs, sizeof(env->gregs));

+ memcpy(env->gregs, src, sizeof(env->gregs));

}

@@ -XXX,XX +XXX,XX @@ void cpu_change_pstate(CPUSPARCState *env, uint32_t new_pstate)

/* Switch global register bank */

src = get_gregset(env, new_pstate_regs);

dst = get_gregset(env, pstate_regs);

- memcpy32(dst, env->gregs);

- memcpy32(env->gregs, src);

+ memcpy(dst, env->gregs, sizeof(env->gregs));

+ memcpy(env->gregs, src, sizeof(env->gregs));

} else {

trace_win_helper_no_switch_pstate(new_pstate_regs);

}

2.43.0

This allows targets to declare that the helper requires a

float_status pointer and instead of a generic void pointer.

include/exec/helper-head.h.inc | 3 +++

1 file changed, 3 insertions(+)

diff --git a/include/exec/helper-head.h.inc b/include/exec/helper-head.h.inc

--- a/include/exec/helper-head.h.inc

+++ b/include/exec/helper-head.h.inc

@@ -XXX,XX +XXX,XX @@

#define dh_alias_ptr ptr

#define dh_alias_cptr ptr

#define dh_alias_env ptr

+#define dh_alias_fpst ptr

#define dh_alias_void void

#define dh_alias_noreturn noreturn

#define dh_alias(t) glue(dh_alias_, t)

@@ -XXX,XX +XXX,XX @@

#define dh_ctype_ptr void *

#define dh_ctype_cptr const void *

#define dh_ctype_env CPUArchState *

+#define dh_ctype_fpst float_status *

#define dh_ctype_void void

#define dh_ctype_noreturn G_NORETURN void

#define dh_ctype(t) dh_ctype_##t

@@ -XXX,XX +XXX,XX @@

#define dh_typecode_f64 dh_typecode_i64

#define dh_typecode_cptr dh_typecode_ptr

#define dh_typecode_env dh_typecode_ptr

+#define dh_typecode_fpst dh_typecode_ptr

#define dh_typecode(t) dh_typecode_##t

#define dh_callflag_i32 0

2.43.0