From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: "Alex Bennée" <alex.bennee@linaro.org>,
"Daniel Henrique Barboza" <danielhb413@gmail.com>
Subject: [PULL 34/53] tcg/ppc: Reorg tcg_out_tlb_read
Date: Thu, 11 May 2023 09:04:31 +0100 [thread overview]
Message-ID: <20230511080450.860923-35-richard.henderson@linaro.org> (raw)
In-Reply-To: <20230511080450.860923-1-richard.henderson@linaro.org>
Allocate TCG_REG_TMP2. Use R0, TMP1, TMP2 instead of any of
the normally allocated registers for the tlb load.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/ppc/tcg-target.c.inc | 78 ++++++++++++++++++++++++----------------
1 file changed, 47 insertions(+), 31 deletions(-)
diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc
index 4c479fdece..dbba304e42 100644
--- a/tcg/ppc/tcg-target.c.inc
+++ b/tcg/ppc/tcg-target.c.inc
@@ -68,6 +68,7 @@
#else
# define TCG_REG_TMP1 TCG_REG_R12
#endif
+#define TCG_REG_TMP2 TCG_REG_R11
#define TCG_VEC_TMP1 TCG_REG_V0
#define TCG_VEC_TMP2 TCG_REG_V1
@@ -2015,13 +2016,11 @@ static TCGReg ldst_ra_gen(TCGContext *s, const TCGLabelQemuLdst *l, int arg)
/*
* For the purposes of ppc32 sorting 4 input registers into 4 argument
* registers, there is an outside chance we would require 3 temps.
- * Because of constraints, no inputs are in r3, and env will not be
- * placed into r3 until after the sorting is done, and is thus free.
*/
static const TCGLdstHelperParam ldst_helper_param = {
.ra_gen = ldst_ra_gen,
.ntmp = 3,
- .tmp = { TCG_REG_TMP1, TCG_REG_R0, TCG_REG_R3 }
+ .tmp = { TCG_REG_TMP1, TCG_REG_TMP2, TCG_REG_R0 }
};
static bool tcg_out_qemu_ld_slow_path(TCGContext *s, TCGLabelQemuLdst *lb)
@@ -2135,31 +2134,31 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
/* Load tlb_mask[mmu_idx] and tlb_table[mmu_idx]. */
QEMU_BUILD_BUG_ON(TLB_MASK_TABLE_OFS(0) > 0);
QEMU_BUILD_BUG_ON(TLB_MASK_TABLE_OFS(0) < -32768);
- tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_R3, TCG_AREG0, mask_off);
- tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_R4, TCG_AREG0, table_off);
+ tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP1, TCG_AREG0, mask_off);
+ tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP2, TCG_AREG0, table_off);
/* Extract the page index, shifted into place for tlb index. */
if (TCG_TARGET_REG_BITS == 32) {
- tcg_out_shri32(s, TCG_REG_TMP1, addrlo,
+ tcg_out_shri32(s, TCG_REG_R0, addrlo,
TARGET_PAGE_BITS - CPU_TLB_ENTRY_BITS);
} else {
- tcg_out_shri64(s, TCG_REG_TMP1, addrlo,
+ tcg_out_shri64(s, TCG_REG_R0, addrlo,
TARGET_PAGE_BITS - CPU_TLB_ENTRY_BITS);
}
- tcg_out32(s, AND | SAB(TCG_REG_R3, TCG_REG_R3, TCG_REG_TMP1));
+ tcg_out32(s, AND | SAB(TCG_REG_TMP1, TCG_REG_TMP1, TCG_REG_R0));
- /* Load the TLB comparator. */
+ /* Load the (low part) TLB comparator into TMP2. */
if (cmp_off == 0 && TCG_TARGET_REG_BITS >= TARGET_LONG_BITS) {
uint32_t lxu = (TCG_TARGET_REG_BITS == 32 || TARGET_LONG_BITS == 32
? LWZUX : LDUX);
- tcg_out32(s, lxu | TAB(TCG_REG_TMP1, TCG_REG_R3, TCG_REG_R4));
+ tcg_out32(s, lxu | TAB(TCG_REG_TMP2, TCG_REG_TMP1, TCG_REG_TMP2));
} else {
- tcg_out32(s, ADD | TAB(TCG_REG_R3, TCG_REG_R3, TCG_REG_R4));
+ tcg_out32(s, ADD | TAB(TCG_REG_TMP1, TCG_REG_TMP1, TCG_REG_TMP2));
if (TCG_TARGET_REG_BITS < TARGET_LONG_BITS) {
- tcg_out_ld(s, TCG_TYPE_I32, TCG_REG_TMP1, TCG_REG_R3, cmp_off + 4);
- tcg_out_ld(s, TCG_TYPE_I32, TCG_REG_R4, TCG_REG_R3, cmp_off);
+ tcg_out_ld(s, TCG_TYPE_I32, TCG_REG_TMP2,
+ TCG_REG_TMP1, cmp_off + 4 * HOST_BIG_ENDIAN);
} else {
- tcg_out_ld(s, TCG_TYPE_TL, TCG_REG_TMP1, TCG_REG_R3, cmp_off);
+ tcg_out_ld(s, TCG_TYPE_TL, TCG_REG_TMP2, TCG_REG_TMP1, cmp_off);
}
}
@@ -2167,11 +2166,12 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
* Load the TLB addend for use on the fast path.
* Do this asap to minimize any load use delay.
*/
- h->base = TCG_REG_R3;
- tcg_out_ld(s, TCG_TYPE_PTR, h->base, TCG_REG_R3,
- offsetof(CPUTLBEntry, addend));
+ if (TCG_TARGET_REG_BITS >= TARGET_LONG_BITS) {
+ tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP1, TCG_REG_TMP1,
+ offsetof(CPUTLBEntry, addend));
+ }
- /* Clear the non-page, non-alignment bits from the address */
+ /* Clear the non-page, non-alignment bits from the address in R0. */
if (TCG_TARGET_REG_BITS == 32) {
/*
* We don't support unaligned accesses on 32-bits.
@@ -2204,9 +2204,6 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
if (TARGET_LONG_BITS == 32) {
tcg_out_rlw(s, RLWINM, TCG_REG_R0, t, 0,
(32 - a_bits) & 31, 31 - TARGET_PAGE_BITS);
- /* Zero-extend the address for use in the final address. */
- tcg_out_ext32u(s, TCG_REG_R4, addrlo);
- addrlo = TCG_REG_R4;
} else if (a_bits == 0) {
tcg_out_rld(s, RLDICR, TCG_REG_R0, t, 0, 63 - TARGET_PAGE_BITS);
} else {
@@ -2215,21 +2212,36 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
tcg_out_rld(s, RLDICL, TCG_REG_R0, TCG_REG_R0, TARGET_PAGE_BITS, 0);
}
}
- h->index = addrlo;
if (TCG_TARGET_REG_BITS < TARGET_LONG_BITS) {
- tcg_out_cmp(s, TCG_COND_EQ, TCG_REG_R0, TCG_REG_TMP1,
+ /* Low part comparison into cr7. */
+ tcg_out_cmp(s, TCG_COND_EQ, TCG_REG_R0, TCG_REG_TMP2,
0, 7, TCG_TYPE_I32);
- tcg_out_cmp(s, TCG_COND_EQ, addrhi, TCG_REG_R4, 0, 6, TCG_TYPE_I32);
+
+ /* Load the high part TLB comparator into TMP2. */
+ tcg_out_ld(s, TCG_TYPE_I32, TCG_REG_TMP2, TCG_REG_TMP1,
+ cmp_off + 4 * !HOST_BIG_ENDIAN);
+
+ /* Load addend, deferred for this case. */
+ tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP1, TCG_REG_TMP1,
+ offsetof(CPUTLBEntry, addend));
+
+ /* High part comparison into cr6. */
+ tcg_out_cmp(s, TCG_COND_EQ, addrhi, TCG_REG_TMP2, 0, 6, TCG_TYPE_I32);
+
+ /* Combine comparisons into cr7. */
tcg_out32(s, CRAND | BT(7, CR_EQ) | BA(6, CR_EQ) | BB(7, CR_EQ));
} else {
- tcg_out_cmp(s, TCG_COND_EQ, TCG_REG_R0, TCG_REG_TMP1,
+ /* Full comparison into cr7. */
+ tcg_out_cmp(s, TCG_COND_EQ, TCG_REG_R0, TCG_REG_TMP2,
0, 7, TCG_TYPE_TL);
}
/* Load a pointer into the current opcode w/conditional branch-link. */
ldst->label_ptr[0] = s->code_ptr;
tcg_out32(s, BC | BI(7, CR_EQ) | BO_COND_FALSE | LK);
+
+ h->base = TCG_REG_TMP1;
#else
if (a_bits) {
ldst = new_ldst_label(s);
@@ -2247,13 +2259,16 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h,
}
h->base = guest_base ? TCG_GUEST_BASE_REG : 0;
- h->index = addrlo;
- if (TCG_TARGET_REG_BITS > TARGET_LONG_BITS) {
- tcg_out_ext32u(s, TCG_REG_TMP1, addrlo);
- h->index = TCG_REG_TMP1;
- }
#endif
+ if (TCG_TARGET_REG_BITS > TARGET_LONG_BITS) {
+ /* Zero-extend the guest address for use in the host address. */
+ tcg_out_ext32u(s, TCG_REG_R0, addrlo);
+ h->index = TCG_REG_R0;
+ } else {
+ h->index = addrlo;
+ }
+
return ldst;
}
@@ -3905,7 +3920,8 @@ static void tcg_target_init(TCGContext *s)
#if defined(_CALL_SYSV) || TCG_TARGET_REG_BITS == 64
tcg_regset_set_reg(s->reserved_regs, TCG_REG_R13); /* thread pointer */
#endif
- tcg_regset_set_reg(s->reserved_regs, TCG_REG_TMP1); /* mem temp */
+ tcg_regset_set_reg(s->reserved_regs, TCG_REG_TMP1);
+ tcg_regset_set_reg(s->reserved_regs, TCG_REG_TMP2);
tcg_regset_set_reg(s->reserved_regs, TCG_VEC_TMP1);
tcg_regset_set_reg(s->reserved_regs, TCG_VEC_TMP2);
if (USE_REG_TB) {
--
2.34.1
next prev parent reply other threads:[~2023-05-11 8:12 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-11 8:03 [PULL 00/53] tcg patch queue Richard Henderson
2023-05-11 8:03 ` [PULL 01/53] target/m68k: Fix gen_load_fp for OS_LONG Richard Henderson
2023-05-11 8:03 ` [PULL 02/53] accel/tcg: Fix atomic_mmu_lookup for reads Richard Henderson
2023-05-11 8:04 ` [PULL 03/53] disas: Fix tabs and braces in disas.c Richard Henderson
2023-05-11 8:04 ` [PULL 04/53] disas: Move disas.c to disas/ Richard Henderson
2023-05-11 8:04 ` [PULL 05/53] disas: Remove target_ulong from the interface Richard Henderson
2023-05-11 8:04 ` [PULL 06/53] disas: Remove target-specific headers Richard Henderson
2023-05-11 8:04 ` [PULL 07/53] disas: Move softmmu specific code to separate file Richard Henderson
2023-05-11 8:04 ` [PULL 08/53] disas: Move disas.c into the target-independent source set Richard Henderson
2023-05-11 8:04 ` [PULL 09/53] cpu: expose qemu_cpu_list_lock for lock-guard use Richard Henderson
2023-05-11 8:04 ` [PULL 10/53] accel/tcg/tcg-accel-ops-rr: ensure fairness with icount Richard Henderson
2023-05-11 8:04 ` [PULL 11/53] tcg/i386: Introduce prepare_host_addr Richard Henderson
2023-05-11 8:04 ` [PULL 12/53] tcg/i386: Use indexed addressing for softmmu fast path Richard Henderson
2023-05-11 8:04 ` [PULL 13/53] tcg/aarch64: Introduce prepare_host_addr Richard Henderson
2023-05-11 8:04 ` [PULL 14/53] tcg/arm: " Richard Henderson
2023-05-11 8:04 ` [PULL 15/53] tcg/loongarch64: " Richard Henderson
2023-05-11 8:04 ` [PULL 16/53] tcg/mips: " Richard Henderson
2023-05-11 8:04 ` [PULL 17/53] tcg/ppc: " Richard Henderson
2023-05-11 8:04 ` [PULL 18/53] tcg/riscv: " Richard Henderson
2023-05-11 8:04 ` [PULL 19/53] tcg/s390x: " Richard Henderson
2023-05-11 8:04 ` [PULL 20/53] tcg: Add routines for calling slow-path helpers Richard Henderson
2023-05-11 8:04 ` [PULL 21/53] tcg/i386: Convert tcg_out_qemu_ld_slow_path Richard Henderson
2023-05-11 8:04 ` [PULL 22/53] tcg/i386: Convert tcg_out_qemu_st_slow_path Richard Henderson
2023-05-11 8:04 ` [PULL 23/53] tcg/aarch64: Convert tcg_out_qemu_{ld,st}_slow_path Richard Henderson
2023-05-11 8:04 ` [PULL 24/53] tcg/arm: " Richard Henderson
2023-05-11 8:04 ` [PULL 25/53] tcg/loongarch64: " Richard Henderson
2023-05-11 8:04 ` [PULL 26/53] tcg/mips: " Richard Henderson
2023-05-11 8:04 ` [PULL 27/53] tcg/ppc: " Richard Henderson
2023-05-11 8:04 ` [PULL 28/53] tcg/riscv: " Richard Henderson
2023-05-11 8:04 ` [PULL 29/53] tcg/s390x: " Richard Henderson
2023-05-11 8:04 ` [PULL 30/53] tcg/loongarch64: Simplify constraints on qemu_ld/st Richard Henderson
2023-05-11 8:04 ` [PULL 31/53] tcg/mips: Remove MO_BSWAP handling Richard Henderson
2023-05-11 8:04 ` [PULL 32/53] tcg/mips: Reorg tlb load within prepare_host_addr Richard Henderson
2023-05-11 8:04 ` [PULL 33/53] tcg/mips: Simplify constraints on qemu_ld/st Richard Henderson
2023-05-11 8:04 ` Richard Henderson [this message]
2023-05-11 8:04 ` [PULL 35/53] tcg/ppc: Adjust " Richard Henderson
2023-05-11 8:04 ` [PULL 36/53] tcg/ppc: Remove unused constraints A, B, C, D Richard Henderson
2023-05-11 8:04 ` [PULL 37/53] tcg/ppc: Remove unused constraint J Richard Henderson
2023-05-11 8:04 ` [PULL 38/53] tcg/riscv: Simplify constraints on qemu_ld/st Richard Henderson
2023-05-11 8:04 ` [PULL 39/53] tcg/s390x: Use ALGFR in constructing softmmu host address Richard Henderson
2023-05-11 8:04 ` [PULL 40/53] tcg/s390x: Simplify constraints on qemu_ld/st Richard Henderson
2023-05-11 8:04 ` [PULL 41/53] target/mips: Add MO_ALIGN to gen_llwp, gen_scwp Richard Henderson
2023-05-11 8:04 ` [PULL 42/53] target/mips: Add missing default_tcg_memop_mask Richard Henderson
2023-05-11 8:04 ` [PULL 43/53] target/mips: Use MO_ALIGN instead of 0 Richard Henderson
2023-05-11 8:04 ` [PULL 44/53] target/mips: Remove TARGET_ALIGNED_ONLY Richard Henderson
2023-05-11 8:04 ` [PULL 45/53] target/nios2: " Richard Henderson
2023-05-11 8:04 ` [PULL 46/53] target/sh4: Use MO_ALIGN where required Richard Henderson
2023-05-11 8:04 ` [PULL 47/53] target/sh4: Remove TARGET_ALIGNED_ONLY Richard Henderson
2023-05-11 8:04 ` [PULL 48/53] tcg: " Richard Henderson
2023-05-11 8:04 ` [PULL 49/53] accel/tcg: Add cpu_in_serial_context Richard Henderson
2023-05-11 8:04 ` [PULL 50/53] accel/tcg: Introduce tlb_read_idx Richard Henderson
2023-05-11 8:04 ` [PULL 51/53] accel/tcg: Reorg system mode load helpers Richard Henderson
2023-05-11 8:04 ` [PULL 52/53] accel/tcg: Reorg system mode store helpers Richard Henderson
2023-05-11 8:04 ` [PULL 53/53] target/loongarch: Do not include tcg-ldst.h 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=20230511080450.860923-35-richard.henderson@linaro.org \
--to=richard.henderson@linaro.org \
--cc=alex.bennee@linaro.org \
--cc=danielhb413@gmail.com \
--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).