From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Subject: [PATCH v2 19/30] accel/tcg: Add have_lse2 support in ldst_atomicity
Date: Wed, 15 Feb 2023 16:57:28 -1000 [thread overview]
Message-ID: <20230216025739.1211680-20-richard.henderson@linaro.org> (raw)
In-Reply-To: <20230216025739.1211680-1-richard.henderson@linaro.org>
Add fast paths for FEAT_LSE2, using the detection in tcg.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
accel/tcg/ldst_atomicity.c.inc | 37 ++++++++++++++++++++++++++++++----
1 file changed, 33 insertions(+), 4 deletions(-)
diff --git a/accel/tcg/ldst_atomicity.c.inc b/accel/tcg/ldst_atomicity.c.inc
index 9a95ac327d..277629f241 100644
--- a/accel/tcg/ldst_atomicity.c.inc
+++ b/accel/tcg/ldst_atomicity.c.inc
@@ -41,6 +41,8 @@
* but we're using tcg/tci/ instead.
*/
# define HAVE_al16_fast false
+#elif defined(__aarch64__)
+# define HAVE_al16_fast likely(have_lse2)
#elif defined(__x86_64__)
# define HAVE_al16_fast likely(have_atomic16)
#else
@@ -48,6 +50,8 @@
#endif
#if defined(CONFIG_ATOMIC128) || defined(CONFIG_CMPXCHG128)
# define HAVE_al16 true
+#elif defined(__aarch64__)
+# define HAVE_al16 true
#else
# define HAVE_al16 false
#endif
@@ -170,6 +174,14 @@ load_atomic16(void *pv)
r.u = qatomic_read__nocheck(p);
return r.s;
+#elif defined(__aarch64__)
+ uint64_t l, h;
+
+ /* Via HAVE_al16_fast, FEAT_LSE2 is present: LDP becomes atomic. */
+ asm("ldp %0, %1, %2" : "=r"(l), "=r"(h) : "m"(*(__uint128_t *)pv));
+
+ qemu_build_assert(!HOST_BIG_ENDIAN);
+ return int128_make128(l, h);
#elif defined(__x86_64__)
Int128Alias r;
@@ -409,6 +421,18 @@ load_atom_extract_al16_or_al8(void *pv, int s)
r = qatomic_read__nocheck(p16);
}
return r >> shr;
+#elif defined(__aarch64__)
+ /*
+ * Via HAVE_al16_fast, FEAT_LSE2 is present.
+ * LDP becomes single-copy atomic if 16-byte aligned, and
+ * single-copy atomic on the parts if 8-byte aligned.
+ */
+ uintptr_t pi = (uintptr_t)pv;
+ int shr = (pi & 7) * 8;
+ uint64_t l, h;
+
+ asm("ldp %0, %1, %2" : "=r"(l), "=r"(h) : "m"(*(__uint128_t *)(pi & ~7)));
+ return (l >> shr) | (h << (-shr & 63));
#elif defined(__x86_64__)
uintptr_t pi = (uintptr_t)pv;
int shr = (pi & 7) * 8;
@@ -764,10 +788,15 @@ store_atomic16(void *pv, Int128Alias val)
l = int128_getlo(val.s);
h = int128_gethi(val.s);
- asm("0: ldxp %0, xzr, %1\n\t"
- "stxp %w0, %2, %3, %1\n\t"
- "cbnz %w0, 0b"
- : "=&r"(t), "=Q"(*(__uint128_t *)pv) : "r"(l), "r"(h));
+ if (HAVE_al16_fast) {
+ /* Via HAVE_al16_fast, FEAT_LSE2 is present: STP becomes atomic. */
+ asm("stp %1, %2, %0" : "=Q"(*(__uint128_t *)pv) : "r"(l), "r"(h));
+ } else {
+ asm("0: ldxp %0, xzr, %1\n\t"
+ "stxp %w0, %2, %3, %1\n\t"
+ "cbnz %w0, 0b"
+ : "=&r"(t), "=Q"(*(__uint128_t *)pv) : "r"(l), "r"(h));
+ }
return;
}
#elif defined(CONFIG_CMPXCHG128)
--
2.34.1
next prev parent reply other threads:[~2023-02-16 3:01 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-02-16 2:57 [PATCH v2 00/30] tcg: Improve atomicity support Richard Henderson
2023-02-16 2:57 ` [PATCH v2 01/30] include/qemu/cpuid: Introduce xgetbv_low Richard Henderson
2023-02-16 2:57 ` [PATCH v2 02/30] include/exec/memop: Add bits describing atomicity Richard Henderson
2023-02-28 17:56 ` Alex Bennée
2023-03-15 17:13 ` Philippe Mathieu-Daudé
2023-02-16 2:57 ` [PATCH v2 03/30] accel/tcg: Add cpu_in_serial_context Richard Henderson
2023-02-28 17:57 ` Alex Bennée
2023-03-15 16:07 ` Philippe Mathieu-Daudé
2023-02-16 2:57 ` [PATCH v2 04/30] accel/tcg: Introduce tlb_read_idx Richard Henderson
2023-02-28 17:59 ` Alex Bennée
2023-02-16 2:57 ` [PATCH v2 05/30] accel/tcg: Reorg system mode load helpers Richard Henderson
2023-02-28 18:08 ` Alex Bennée
2023-02-16 2:57 ` [PATCH v2 06/30] accel/tcg: Reorg system mode store helpers Richard Henderson
2023-02-16 2:57 ` [PATCH v2 07/30] accel/tcg: Honor atomicity of loads Richard Henderson
2023-02-28 18:19 ` Alex Bennée
2023-02-16 2:57 ` [PATCH v2 08/30] accel/tcg: Honor atomicity of stores Richard Henderson
2023-02-16 2:57 ` [PATCH v2 09/30] tcg/tci: Use cpu_{ld,st}_mmu Richard Henderson
2023-02-16 2:57 ` [PATCH v2 10/30] tcg: Unify helper_{be,le}_{ld,st}* Richard Henderson
2023-02-16 2:57 ` [PATCH v2 11/30] accel/tcg: Implement helper_{ld, st}*_mmu for user-only Richard Henderson
2023-02-16 2:57 ` [PATCH v2 12/30] tcg: Add 128-bit guest memory primitives Richard Henderson
2023-02-16 2:57 ` [PATCH v2 13/30] meson: Detect atomic128 support with optimization Richard Henderson
2023-02-16 2:57 ` [PATCH v2 14/30] tcg/i386: Add have_atomic16 Richard Henderson
2023-03-15 17:06 ` Philippe Mathieu-Daudé
2023-02-16 2:57 ` [PATCH v2 15/30] accel/tcg: Use have_atomic16 in ldst_atomicity.c.inc Richard Henderson
2023-02-16 2:57 ` [PATCH v2 16/30] accel/tcg: Add aarch64 specific support in ldst_atomicity Richard Henderson
2023-02-16 2:57 ` [PATCH v2 17/30] tcg/aarch64: Detect have_lse, have_lse2 for linux Richard Henderson
2023-03-15 17:00 ` Philippe Mathieu-Daudé
2023-02-16 2:57 ` [PATCH v2 18/30] tcg/aarch64: Detect have_lse, have_lse2 for darwin Richard Henderson
2023-03-15 16:59 ` Philippe Mathieu-Daudé
2023-02-16 2:57 ` Richard Henderson [this message]
2023-02-16 2:57 ` [PATCH v2 20/30] tcg: Introduce TCG_OPF_TYPE_MASK Richard Henderson
2023-02-16 2:57 ` [PATCH v2 21/30] tcg: Add INDEX_op_qemu_{ld,st}_i128 Richard Henderson
2023-02-16 2:57 ` [PATCH v2 22/30] tcg/i386: Introduce tcg_out_mov2 Richard Henderson
2023-02-16 2:57 ` [PATCH v2 23/30] tcg/i386: Introduce tcg_out_testi Richard Henderson
2023-02-16 2:57 ` [PATCH v2 24/30] tcg/i386: Use full load/store helpers in user-only mode Richard Henderson
2023-02-16 2:57 ` [PATCH v2 25/30] tcg/i386: Replace is64 with type in qemu_ld/st routines Richard Henderson
2023-02-16 2:57 ` [PATCH v2 26/30] tcg/i386: Mark Win64 call-saved vector regs as reserved Richard Henderson
2023-02-16 2:57 ` [PATCH v2 27/30] tcg/i386: Examine MemOp for atomicity and alignment Richard Henderson
2023-02-16 2:57 ` [PATCH v2 28/30] tcg/i386: Support 128-bit load/store with have_atomic16 Richard Henderson
2023-02-16 2:57 ` [PATCH v2 29/30] tcg/i386: Add vex_v argument to tcg_out_vex_modrm_pool Richard Henderson
2023-03-15 16:43 ` Philippe Mathieu-Daudé
2023-02-16 2:57 ` [PATCH v2 30/30] tcg/i386: Honor 64-bit atomicity in 32-bit mode Richard Henderson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230216025739.1211680-20-richard.henderson@linaro.org \
--to=richard.henderson@linaro.org \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).