qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: qemu-arm@nongnu.org, qemu-s390x@nongnu.org,
	qemu-riscv@nongnu.org, qemu-ppc@nongnu.org, git@xen0n.name,
	jiaxun.yang@flygoat.com, philmd@linaro.org
Subject: [PATCH v3 42/57] tcg: Introduce atom_and_align_for_opc
Date: Tue, 25 Apr 2023 20:31:31 +0100	[thread overview]
Message-ID: <20230425193146.2106111-43-richard.henderson@linaro.org> (raw)
In-Reply-To: <20230425193146.2106111-1-richard.henderson@linaro.org>

Examine MemOp for atomicity and alignment, adjusting alignment
as required to implement atomicity on the host.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 tcg/tcg.c | 69 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 69 insertions(+)

diff --git a/tcg/tcg.c b/tcg/tcg.c
index c5a0cfd846..d7ff96fd1d 100644
--- a/tcg/tcg.c
+++ b/tcg/tcg.c
@@ -220,6 +220,11 @@ static void * const qemu_st_helpers[MO_SIZE + 1] __attribute__((unused)) = {
 #endif
 };
 
+static MemOp atom_and_align_for_opc(TCGContext *s, MemOp *p_atom_a,
+                                    MemOp *p_atom_u, MemOp opc,
+                                    MemOp host_atom, bool allow_two_ops)
+    __attribute__((unused));
+
 TCGContext tcg_init_ctx;
 __thread TCGContext *tcg_ctx;
 
@@ -5123,6 +5128,70 @@ static void tcg_reg_alloc_call(TCGContext *s, TCGOp *op)
     }
 }
 
+/*
+ * Return the alignment and atomicity to use for the inline fast path
+ * for the given memory operation.  The alignment may be larger than
+ * that specified in @opc, and the correct alignment will be diagnosed
+ * by the slow path helper.
+ */
+static MemOp atom_and_align_for_opc(TCGContext *s, MemOp *p_atom_a,
+                                    MemOp *p_atom_u, MemOp opc,
+                                    MemOp host_atom, bool allow_two_ops)
+{
+    MemOp align = get_alignment_bits(opc);
+    MemOp atom, atmax, atmin, size = opc & MO_SIZE;
+
+    /* When serialized, no further atomicity required.  */
+    if (s->gen_tb->cflags & CF_PARALLEL) {
+        atom = opc & MO_ATOM_MASK;
+    } else {
+        atom = MO_ATOM_NONE;
+    }
+
+    atmax = opc & MO_ATMAX_MASK;
+    if (atmax == MO_ATMAX_SIZE) {
+        atmax = size;
+    } else {
+        atmax = atmax >> MO_ATMAX_SHIFT;
+    }
+
+    switch (atom) {
+    case MO_ATOM_NONE:
+        /* The operation requires no specific atomicity. */
+        atmax = atmin = MO_8;
+        break;
+    case MO_ATOM_IFALIGN:
+        /* If unaligned, the subobjects are bytes. */
+        atmin = MO_8;
+        break;
+    case MO_ATOM_WITHIN16:
+        /* If unaligned, there are subobjects if atmax < size. */
+        atmin = (atmax < size ? atmax : MO_8);
+        atmax = size;
+        break;
+    case MO_ATOM_SUBALIGN:
+        /* If unaligned but not odd, there are subobjects up to atmax - 1. */
+        atmin = (atmax == MO_8 ? MO_8 : atmax - 1);
+        break;
+    default:
+        g_assert_not_reached();
+    }
+
+    /*
+     * If there are subobjects, and the host model does not match, then we
+     * need to raise the initial alignment check.  If the backend is prepared
+     * to double-check alignment and issue two half size ops, we need not
+     * raise initial alignment beyond half.
+     */
+    if (atmin > MO_8 && host_atom != atom) {
+        align = MAX(align, size - allow_two_ops);
+    }
+
+    *p_atom_a = atmax;
+    *p_atom_u = atmin;
+    return align;
+}
+
 /*
  * Similarly for qemu_ld/st slow path helpers.
  * We must re-implement tcg_gen_callN and tcg_reg_alloc_call simultaneously,
-- 
2.34.1



  parent reply	other threads:[~2023-04-25 19:47 UTC|newest]

Thread overview: 62+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-04-25 19:30 [PATCH v3 00/57] tcg: Improve atomicity support Richard Henderson
2023-04-25 19:30 ` [PATCH v3 01/57] include/exec/memop: Add bits describing atomicity Richard Henderson
2023-04-25 19:30 ` [PATCH v3 02/57] accel/tcg: Add cpu_in_serial_context Richard Henderson
2023-05-04 14:52   ` Peter Maydell
2023-04-25 19:30 ` [PATCH v3 03/57] accel/tcg: Introduce tlb_read_idx Richard Henderson
2023-04-25 19:30 ` [PATCH v3 04/57] accel/tcg: Reorg system mode load helpers Richard Henderson
2023-04-25 19:30 ` [PATCH v3 05/57] accel/tcg: Reorg system mode store helpers Richard Henderson
2023-04-25 19:30 ` [PATCH v3 06/57] accel/tcg: Honor atomicity of loads Richard Henderson
2023-04-25 19:30 ` [PATCH v3 07/57] accel/tcg: Honor atomicity of stores Richard Henderson
2023-04-25 19:30 ` [PATCH v3 08/57] target/loongarch: Do not include tcg-ldst.h Richard Henderson
2023-04-26  0:45   ` Song Gao
2023-04-25 19:30 ` [PATCH v3 09/57] tcg: Unify helper_{be,le}_{ld,st}* Richard Henderson
2023-04-25 19:30 ` [PATCH v3 10/57] accel/tcg: Implement helper_{ld, st}*_mmu for user-only Richard Henderson
2023-04-25 19:31 ` [PATCH v3 11/57] tcg/tci: Use helper_{ld,st}*_mmu " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 12/57] tcg: Add 128-bit guest memory primitives Richard Henderson
2023-04-25 19:31 ` [PATCH v3 13/57] meson: Detect atomic128 support with optimization Richard Henderson
2023-04-25 19:31 ` [PATCH v3 14/57] tcg/i386: Add have_atomic16 Richard Henderson
2023-04-25 19:31 ` [PATCH v3 15/57] accel/tcg: Use have_atomic16 in ldst_atomicity.c.inc Richard Henderson
2023-04-25 19:31 ` [PATCH v3 16/57] accel/tcg: Add aarch64 specific support in ldst_atomicity Richard Henderson
2023-04-25 19:31 ` [PATCH v3 17/57] tcg/aarch64: Detect have_lse, have_lse2 for linux Richard Henderson
2023-04-25 19:31 ` [PATCH v3 18/57] tcg/aarch64: Detect have_lse, have_lse2 for darwin Richard Henderson
2023-04-25 19:31 ` [PATCH v3 19/57] accel/tcg: Add have_lse2 support in ldst_atomicity Richard Henderson
2023-04-25 19:31 ` [PATCH v3 20/57] tcg: Introduce TCG_OPF_TYPE_MASK Richard Henderson
2023-04-25 19:31 ` [PATCH v3 21/57] tcg/i386: Use full load/store helpers in user-only mode Richard Henderson
2023-04-25 19:31 ` [PATCH v3 22/57] tcg/aarch64: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 23/57] tcg/ppc: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 24/57] tcg/loongarch64: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 25/57] tcg/riscv: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 26/57] tcg/arm: Adjust constraints on qemu_ld/st Richard Henderson
2023-04-25 19:31 ` [PATCH v3 27/57] tcg/arm: Use full load/store helpers in user-only mode Richard Henderson
2023-04-25 19:31 ` [PATCH v3 28/57] tcg/mips: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 29/57] tcg/s390x: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 30/57] tcg/sparc64: Allocate %g2 as a third temporary Richard Henderson
2023-04-25 19:31 ` [PATCH v3 31/57] tcg/sparc64: Rename tcg_out_movi_imm13 to tcg_out_movi_s13 Richard Henderson
2023-04-25 19:31 ` [PATCH v3 32/57] tcg/sparc64: Rename tcg_out_movi_imm32 to tcg_out_movi_u32 Richard Henderson
2023-04-25 19:31 ` [PATCH v3 33/57] tcg/sparc64: Split out tcg_out_movi_s32 Richard Henderson
2023-04-25 19:31 ` [PATCH v3 34/57] tcg/sparc64: Use standard slow path for softmmu Richard Henderson
2023-04-25 19:31 ` [PATCH v3 35/57] accel/tcg: Remove helper_unaligned_{ld,st} Richard Henderson
2023-04-25 19:31 ` [PATCH v3 36/57] tcg/loongarch64: Assert the host supports unaligned accesses Richard Henderson
2023-04-25 19:31 ` [PATCH v3 37/57] tcg/loongarch64: Support softmmu " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 38/57] tcg/riscv: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 39/57] tcg: Introduce tcg_target_has_memory_bswap Richard Henderson
2023-04-25 19:31 ` [PATCH v3 40/57] tcg: Add INDEX_op_qemu_{ld,st}_i128 Richard Henderson
2023-04-25 19:31 ` [PATCH v3 41/57] tcg: Support TCG_TYPE_I128 in tcg_out_{ld, st}_helper_{args, ret} Richard Henderson
2023-04-25 19:31 ` Richard Henderson [this message]
2023-04-25 19:31 ` [PATCH v3 43/57] tcg/i386: Use atom_and_align_for_opc Richard Henderson
2023-04-25 19:31 ` [PATCH v3 44/57] tcg/aarch64: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 45/57] tcg/arm: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 46/57] tcg/loongarch64: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 47/57] tcg/mips: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 48/57] tcg/ppc: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 49/57] tcg/riscv: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 50/57] tcg/s390x: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 51/57] tcg/sparc64: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 52/57] tcg/i386: Honor 64-bit atomicity in 32-bit mode Richard Henderson
2023-04-25 19:31 ` [PATCH v3 53/57] tcg/i386: Support 128-bit load/store with have_atomic16 Richard Henderson
2023-04-25 19:31 ` [PATCH v3 54/57] tcg/aarch64: Rename temporaries Richard Henderson
2023-04-25 19:31 ` [PATCH v3 55/57] tcg/aarch64: Support 128-bit load/store Richard Henderson
2023-04-25 19:31 ` [PATCH v3 56/57] tcg/ppc: " Richard Henderson
2023-04-25 19:31 ` [PATCH v3 57/57] tcg/s390x: " Richard Henderson
2023-05-02 16:11 ` [PATCH v3 00/57] tcg: Improve atomicity support Peter Maydell
2023-05-02 19:19   ` 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=20230425193146.2106111-43-richard.henderson@linaro.org \
    --to=richard.henderson@linaro.org \
    --cc=git@xen0n.name \
    --cc=jiaxun.yang@flygoat.com \
    --cc=philmd@linaro.org \
    --cc=qemu-arm@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-ppc@nongnu.org \
    --cc=qemu-riscv@nongnu.org \
    --cc=qemu-s390x@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).