From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Subject: [PULL 31/80] tcg/riscv: Support softmmu unaligned accesses
Date: Tue, 16 May 2023 12:40:56 -0700 [thread overview]
Message-ID: <20230516194145.1749305-32-richard.henderson@linaro.org> (raw)
In-Reply-To: <20230516194145.1749305-1-richard.henderson@linaro.org>
The system is required to emulate unaligned accesses, even if the
hardware does not support it. The resulting trap may or may not
be more efficient than the qemu slow path. There are linux kernel
patches in flight to allow userspace to query hardware support;
we can re-evaluate whether to enable this by default after that.
In the meantime, softmmu now matches useronly, where we already
assumed that unaligned accesses are supported.
Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/riscv/tcg-target.c.inc | 48 ++++++++++++++++++++++----------------
1 file changed, 28 insertions(+), 20 deletions(-)
diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc
index 19cd4507fb..415e6c6e15 100644
--- a/tcg/riscv/tcg-target.c.inc
+++ b/tcg/riscv/tcg-target.c.inc
@@ -910,12 +910,13 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, TCGReg *pbase,
#ifdef CONFIG_SOFTMMU
unsigned s_bits = opc & MO_SIZE;
+ unsigned s_mask = (1u << s_bits) - 1;
int mem_index = get_mmuidx(oi);
int fast_ofs = TLB_MASK_TABLE_OFS(mem_index);
int mask_ofs = fast_ofs + offsetof(CPUTLBDescFast, mask);
int table_ofs = fast_ofs + offsetof(CPUTLBDescFast, table);
- TCGReg mask_base = TCG_AREG0, table_base = TCG_AREG0;
- tcg_target_long compare_mask;
+ int compare_mask;
+ TCGReg addr_adj;
ldst = new_ldst_label(s);
ldst->is_ld = is_ld;
@@ -924,14 +925,33 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, TCGReg *pbase,
QEMU_BUILD_BUG_ON(TLB_MASK_TABLE_OFS(0) > 0);
QEMU_BUILD_BUG_ON(TLB_MASK_TABLE_OFS(0) < -(1 << 11));
- tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP0, mask_base, mask_ofs);
- tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP1, table_base, table_ofs);
+ tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP0, TCG_AREG0, mask_ofs);
+ tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP1, TCG_AREG0, table_ofs);
tcg_out_opc_imm(s, OPC_SRLI, TCG_REG_TMP2, addr_reg,
TARGET_PAGE_BITS - CPU_TLB_ENTRY_BITS);
tcg_out_opc_reg(s, OPC_AND, TCG_REG_TMP2, TCG_REG_TMP2, TCG_REG_TMP0);
tcg_out_opc_reg(s, OPC_ADD, TCG_REG_TMP2, TCG_REG_TMP2, TCG_REG_TMP1);
+ /*
+ * For aligned accesses, we check the first byte and include the alignment
+ * bits within the address. For unaligned access, we check that we don't
+ * cross pages using the address of the last byte of the access.
+ */
+ addr_adj = addr_reg;
+ if (a_bits < s_bits) {
+ addr_adj = TCG_REG_TMP0;
+ tcg_out_opc_imm(s, TARGET_LONG_BITS == 32 ? OPC_ADDIW : OPC_ADDI,
+ addr_adj, addr_reg, s_mask - a_mask);
+ }
+ compare_mask = TARGET_PAGE_MASK | a_mask;
+ if (compare_mask == sextreg(compare_mask, 0, 12)) {
+ tcg_out_opc_imm(s, OPC_ANDI, TCG_REG_TMP1, addr_adj, compare_mask);
+ } else {
+ tcg_out_movi(s, TCG_TYPE_TL, TCG_REG_TMP1, compare_mask);
+ tcg_out_opc_reg(s, OPC_AND, TCG_REG_TMP1, TCG_REG_TMP1, addr_adj);
+ }
+
/* Load the tlb comparator and the addend. */
tcg_out_ld(s, TCG_TYPE_TL, TCG_REG_TMP0, TCG_REG_TMP2,
is_ld ? offsetof(CPUTLBEntry, addr_read)
@@ -939,29 +959,17 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, TCGReg *pbase,
tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP2, TCG_REG_TMP2,
offsetof(CPUTLBEntry, addend));
- /* We don't support unaligned accesses. */
- if (a_bits < s_bits) {
- a_bits = s_bits;
- }
- /* Clear the non-page, non-alignment bits from the address. */
- compare_mask = (tcg_target_long)TARGET_PAGE_MASK | a_mask;
- if (compare_mask == sextreg(compare_mask, 0, 12)) {
- tcg_out_opc_imm(s, OPC_ANDI, TCG_REG_TMP1, addr_reg, compare_mask);
- } else {
- tcg_out_movi(s, TCG_TYPE_TL, TCG_REG_TMP1, compare_mask);
- tcg_out_opc_reg(s, OPC_AND, TCG_REG_TMP1, TCG_REG_TMP1, addr_reg);
- }
-
/* Compare masked address with the TLB entry. */
ldst->label_ptr[0] = s->code_ptr;
tcg_out_opc_branch(s, OPC_BNE, TCG_REG_TMP0, TCG_REG_TMP1, 0);
/* TLB Hit - translate address using addend. */
+ addr_adj = addr_reg;
if (TARGET_LONG_BITS == 32) {
- tcg_out_ext32u(s, TCG_REG_TMP0, addr_reg);
- addr_reg = TCG_REG_TMP0;
+ addr_adj = TCG_REG_TMP0;
+ tcg_out_ext32u(s, addr_adj, addr_reg);
}
- tcg_out_opc_reg(s, OPC_ADD, TCG_REG_TMP0, TCG_REG_TMP2, addr_reg);
+ tcg_out_opc_reg(s, OPC_ADD, TCG_REG_TMP0, TCG_REG_TMP2, addr_adj);
*pbase = TCG_REG_TMP0;
#else
if (a_mask) {
--
2.34.1
next prev parent reply other threads:[~2023-05-16 19:57 UTC|newest]
Thread overview: 84+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-16 19:40 [PULL 00/80] tcg patch queue Richard Henderson
2023-05-16 19:40 ` [PULL 01/80] tcg/i386: Set P_REXW in tcg_out_addi_ptr Richard Henderson
2023-05-16 19:40 ` [PULL 02/80] include/exec/memop: Add MO_ATOM_* Richard Henderson
2023-05-16 19:40 ` [PULL 03/80] accel/tcg: Honor atomicity of loads Richard Henderson
2023-05-16 19:40 ` [PULL 04/80] accel/tcg: Honor atomicity of stores Richard Henderson
2023-05-16 19:40 ` [PULL 05/80] tcg: Unify helper_{be,le}_{ld,st}* Richard Henderson
2023-05-16 19:40 ` [PULL 06/80] accel/tcg: Implement helper_{ld,st}*_mmu for user-only Richard Henderson
2023-05-16 19:40 ` [PULL 07/80] tcg/tci: Use " Richard Henderson
2023-05-16 19:40 ` [PULL 08/80] tcg: Add 128-bit guest memory primitives Richard Henderson
2023-05-16 19:40 ` [PULL 09/80] meson: Detect atomic128 support with optimization Richard Henderson
2023-05-16 19:40 ` [PULL 10/80] tcg/i386: Add have_atomic16 Richard Henderson
2023-05-16 19:40 ` [PULL 11/80] tcg/aarch64: Detect have_lse, have_lse2 for linux Richard Henderson
2023-05-16 19:40 ` [PULL 12/80] tcg/aarch64: Detect have_lse, have_lse2 for darwin Richard Henderson
2023-05-16 19:40 ` [PULL 13/80] tcg/i386: Use full load/store helpers in user-only mode Richard Henderson
2023-05-16 19:40 ` [PULL 14/80] tcg/aarch64: " Richard Henderson
2023-05-16 19:40 ` [PULL 15/80] tcg/ppc: " Richard Henderson
2023-05-16 19:40 ` [PULL 16/80] tcg/loongarch64: " Richard Henderson
2023-05-16 19:40 ` [PULL 17/80] tcg/riscv: " Richard Henderson
2023-05-16 19:40 ` [PULL 18/80] tcg/arm: Adjust constraints on qemu_ld/st Richard Henderson
2023-05-16 19:40 ` [PULL 19/80] tcg/arm: Use full load/store helpers in user-only mode Richard Henderson
2023-05-16 19:40 ` [PULL 20/80] tcg/mips: " Richard Henderson
2023-05-16 19:40 ` [PULL 21/80] tcg/s390x: " Richard Henderson
2023-05-16 19:40 ` [PULL 22/80] tcg/sparc64: Allocate %g2 as a third temporary Richard Henderson
2023-05-16 19:40 ` [PULL 23/80] tcg/sparc64: Rename tcg_out_movi_imm13 to tcg_out_movi_s13 Richard Henderson
2023-05-16 19:40 ` [PULL 24/80] target/sparc64: Remove tcg_out_movi_s13 case from tcg_out_movi_imm32 Richard Henderson
2023-05-16 19:40 ` [PULL 25/80] tcg/sparc64: Rename tcg_out_movi_imm32 to tcg_out_movi_u32 Richard Henderson
2023-05-16 19:40 ` [PULL 26/80] tcg/sparc64: Split out tcg_out_movi_s32 Richard Henderson
2023-05-16 19:40 ` [PULL 27/80] tcg/sparc64: Use standard slow path for softmmu Richard Henderson
2023-05-16 19:40 ` [PULL 28/80] accel/tcg: Remove helper_unaligned_{ld,st} Richard Henderson
2023-05-16 19:40 ` [PULL 29/80] tcg/loongarch64: Check the host supports unaligned accesses Richard Henderson
2023-05-16 19:40 ` [PULL 30/80] tcg/loongarch64: Support softmmu " Richard Henderson
2023-05-16 19:40 ` Richard Henderson [this message]
2023-05-16 19:40 ` [PULL 32/80] tcg: Introduce tcg_target_has_memory_bswap Richard Henderson
2023-05-16 19:40 ` [PULL 33/80] tcg: Add INDEX_op_qemu_{ld,st}_i128 Richard Henderson
2023-05-16 19:40 ` [PULL 34/80] tcg: Introduce tcg_out_movext3 Richard Henderson
2023-05-16 19:41 ` [PULL 35/80] tcg: Merge tcg_out_helper_load_regs into caller Richard Henderson
2023-05-16 19:41 ` [PULL 36/80] tcg: Support TCG_TYPE_I128 in tcg_out_{ld, st}_helper_{args, ret} Richard Henderson
2023-05-16 19:41 ` [PULL 37/80] tcg: Introduce atom_and_align_for_opc Richard Henderson
2023-05-16 19:41 ` [PULL 38/80] tcg/i386: Use atom_and_align_for_opc Richard Henderson
2023-05-16 19:41 ` [PULL 39/80] tcg/aarch64: " Richard Henderson
2023-05-16 19:41 ` [PULL 40/80] tcg/arm: " Richard Henderson
2023-05-16 19:41 ` [PULL 41/80] tcg/loongarch64: " Richard Henderson
2023-05-16 19:41 ` [PULL 42/80] tcg/mips: " Richard Henderson
2023-05-16 19:41 ` [PULL 43/80] tcg/ppc: " Richard Henderson
2023-05-16 19:41 ` [PULL 44/80] tcg/riscv: " Richard Henderson
2023-05-16 19:41 ` [PULL 45/80] tcg/s390x: " Richard Henderson
2023-05-16 19:41 ` [PULL 46/80] tcg/sparc64: " Richard Henderson
2023-05-16 19:41 ` [PULL 47/80] tcg/i386: Honor 64-bit atomicity in 32-bit mode Richard Henderson
2023-05-16 19:41 ` [PULL 48/80] tcg/i386: Support 128-bit load/store with have_atomic16 Richard Henderson
2023-05-16 19:41 ` [PULL 49/80] tcg/aarch64: Rename temporaries Richard Henderson
2023-05-16 19:41 ` [PULL 50/80] tcg/aarch64: Support 128-bit load/store Richard Henderson
2023-05-16 19:41 ` [PULL 51/80] tcg/ppc: " Richard Henderson
2023-05-16 19:41 ` [PULL 52/80] tcg/s390x: " Richard Henderson
2023-05-16 19:41 ` [PULL 53/80] tcg: Split out memory ops to tcg-op-ldst.c Richard Henderson
2023-05-16 19:41 ` [PULL 54/80] tcg: Widen gen_insn_data to uint64_t Richard Henderson
2023-05-16 19:41 ` [PULL 55/80] accel/tcg: Widen tcg-ldst.h addresses " Richard Henderson
2023-05-16 19:41 ` [PULL 56/80] tcg: Widen helper_{ld,st}_i128 " Richard Henderson
2023-05-16 19:41 ` [PULL 57/80] tcg: Widen helper_atomic_* " Richard Henderson
2023-05-16 19:41 ` [PULL 58/80] tcg: Widen tcg_gen_code pc_start argument " Richard Henderson
2023-05-16 19:41 ` [PULL 59/80] accel/tcg: Merge gen_mem_wrapped with plugin_gen_empty_mem_callback Richard Henderson
2023-05-16 19:41 ` [PULL 60/80] accel/tcg: Merge do_gen_mem_cb into caller Richard Henderson
2023-05-16 19:41 ` [PULL 61/80] tcg: Reduce copies for plugin_gen_mem_callbacks Richard Henderson
2023-05-16 19:41 ` [PULL 62/80] accel/tcg: Widen plugin_gen_empty_mem_callback to i64 Richard Henderson
2023-05-18 13:22 ` Peter Maydell
2023-05-16 19:41 ` [PULL 63/80] tcg: Add addr_type to TCGContext Richard Henderson
2023-05-16 19:41 ` [PULL 64/80] tcg: Remove TCGv from tcg_gen_qemu_{ld,st}_* Richard Henderson
2023-05-16 19:41 ` [PULL 65/80] tcg: Remove TCGv from tcg_gen_atomic_* Richard Henderson
2023-05-16 19:41 ` [PULL 66/80] tcg: Split INDEX_op_qemu_{ld, st}* for guest address size Richard Henderson
2023-05-16 19:41 ` [PULL 67/80] tcg/tci: Elimnate TARGET_LONG_BITS, target_ulong Richard Henderson
2023-05-16 19:41 ` [PULL 68/80] tcg/i386: Always enable TCG_TARGET_HAS_extr[lh]_i64_i32 Richard Henderson
2023-05-16 19:41 ` [PULL 69/80] tcg/i386: Conditionalize tcg_out_extu_i32_i64 Richard Henderson
2023-05-16 19:41 ` [PULL 70/80] tcg/i386: Adjust type of tlb_mask Richard Henderson
2023-05-16 19:41 ` [PULL 71/80] tcg/i386: Remove TARGET_LONG_BITS, TCG_TYPE_TL Richard Henderson
2023-05-16 19:41 ` [PULL 72/80] tcg/arm: Remove TARGET_LONG_BITS Richard Henderson
2023-05-16 19:41 ` [PULL 73/80] tcg/aarch64: Remove USE_GUEST_BASE Richard Henderson
2023-05-16 19:41 ` [PULL 74/80] tcg/aarch64: Remove TARGET_LONG_BITS, TCG_TYPE_TL Richard Henderson
2023-05-16 19:41 ` [PULL 75/80] tcg/loongarch64: " Richard Henderson
2023-05-16 19:41 ` [PULL 76/80] tcg/mips: " Richard Henderson
2023-05-16 19:41 ` [PULL 77/80] tcg: " Richard Henderson
2023-05-16 19:41 ` [PULL 78/80] tcg: Add page_bits and page_mask to TCGContext Richard Henderson
2023-05-16 19:41 ` [PULL 79/80] tcg: Add tlb_dyn_max_bits " Richard Henderson
2023-05-16 19:41 ` [PULL 80/80] tcg: Split out exec/user/guest-base.h Richard Henderson
2023-05-17 15:11 ` [PULL 00/80] tcg patch queue Peter Maydell
2023-05-17 19:16 ` 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=20230516194145.1749305-32-richard.henderson@linaro.org \
--to=richard.henderson@linaro.org \
--cc=qemu-devel@nongnu.org \
--cc=zhiwei_liu@linux.alibaba.com \
/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).