From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: Jordan Niethe <jniethe5@gmail.com>
Subject: [PULL 02/29] tcg/ppc: Enable direct branching tcg_out_goto_tb with TCG_REG_TB
Date: Wed, 18 Oct 2023 15:25:30 -0700 [thread overview]
Message-ID: <20231018222557.1562065-3-richard.henderson@linaro.org> (raw)
In-Reply-To: <20231018222557.1562065-1-richard.henderson@linaro.org>
From: Jordan Niethe <jniethe5@gmail.com>
Direct branch patching was disabled when using TCG_REG_TB in commit
736a1588c1 ("tcg/ppc: Fix race in goto_tb implementation").
The issue with direct branch patching with TCG_REG_TB is the lack of
synchronization between the new TCG_REG_TB being established and the
direct branch being patched in.
If each translation block is responsible for establishing its own
TCG_REG_TB then there can be no synchronization issue.
Make each translation block begin by setting up its own TCG_REG_TB.
Use the preferred 'bcl 20,31,$+4' sequence.
Signed-off-by: Jordan Niethe <jniethe5@gmail.com>
[rth: Split out tcg_out_tb_start, power9 addpcis]
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/ppc/tcg-target.c.inc | 48 ++++++++++++++--------------------------
1 file changed, 17 insertions(+), 31 deletions(-)
diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc
index 5cecc6ed95..9197cfd6c6 100644
--- a/tcg/ppc/tcg-target.c.inc
+++ b/tcg/ppc/tcg-target.c.inc
@@ -2509,9 +2509,6 @@ static void tcg_target_qemu_prologue(TCGContext *s)
tcg_out_mov(s, TCG_TYPE_PTR, TCG_AREG0, tcg_target_call_iarg_regs[0]);
tcg_out32(s, MTSPR | RS(tcg_target_call_iarg_regs[1]) | CTR);
- if (USE_REG_TB) {
- tcg_out_mov(s, TCG_TYPE_PTR, TCG_REG_TB, tcg_target_call_iarg_regs[1]);
- }
tcg_out32(s, BCCTR | BO_ALWAYS);
/* Epilogue */
@@ -2529,7 +2526,13 @@ static void tcg_target_qemu_prologue(TCGContext *s)
static void tcg_out_tb_start(TCGContext *s)
{
- /* nothing to do */
+ /* Load TCG_REG_TB. */
+ if (USE_REG_TB) {
+ /* bcl 20,31,$+4 (preferred form for getting nia) */
+ tcg_out32(s, BC | BO_ALWAYS | BI(7, CR_SO) | 0x4 | LK);
+ tcg_out32(s, MFSPR | RT(TCG_REG_TB) | LR);
+ tcg_out32(s, ADDI | TAI(TCG_REG_TB, TCG_REG_TB, -4));
+ }
}
static void tcg_out_exit_tb(TCGContext *s, uintptr_t arg)
@@ -2542,32 +2545,22 @@ static void tcg_out_goto_tb(TCGContext *s, int which)
{
uintptr_t ptr = get_jmp_target_addr(s, which);
+ /* Direct branch will be patched by tb_target_set_jmp_target. */
+ set_jmp_insn_offset(s, which);
+ tcg_out32(s, NOP);
+
+ /* When branch is out of range, fall through to indirect. */
if (USE_REG_TB) {
ptrdiff_t offset = tcg_tbrel_diff(s, (void *)ptr);
- tcg_out_mem_long(s, LD, LDX, TCG_REG_TB, TCG_REG_TB, offset);
-
- /* TODO: Use direct branches when possible. */
- set_jmp_insn_offset(s, which);
- tcg_out32(s, MTSPR | RS(TCG_REG_TB) | CTR);
-
- tcg_out32(s, BCCTR | BO_ALWAYS);
-
- /* For the unlinked case, need to reset TCG_REG_TB. */
- set_jmp_reset_offset(s, which);
- tcg_out_mem_long(s, ADDI, ADD, TCG_REG_TB, TCG_REG_TB,
- -tcg_current_code_size(s));
+ tcg_out_mem_long(s, LD, LDX, TCG_REG_TMP1, TCG_REG_TB, offset);
} else {
- /* Direct branch will be patched by tb_target_set_jmp_target. */
- set_jmp_insn_offset(s, which);
- tcg_out32(s, NOP);
-
- /* When branch is out of range, fall through to indirect. */
tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_TMP1, ptr - (int16_t)ptr);
tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP1, TCG_REG_TMP1, (int16_t)ptr);
- tcg_out32(s, MTSPR | RS(TCG_REG_TMP1) | CTR);
- tcg_out32(s, BCCTR | BO_ALWAYS);
- set_jmp_reset_offset(s, which);
}
+
+ tcg_out32(s, MTSPR | RS(TCG_REG_TMP1) | CTR);
+ tcg_out32(s, BCCTR | BO_ALWAYS);
+ set_jmp_reset_offset(s, which);
}
void tb_target_set_jmp_target(const TranslationBlock *tb, int n,
@@ -2577,10 +2570,6 @@ void tb_target_set_jmp_target(const TranslationBlock *tb, int n,
intptr_t diff = addr - jmp_rx;
tcg_insn_unit insn;
- if (USE_REG_TB) {
- return;
- }
-
if (in_range_b(diff)) {
insn = B | (diff & 0x3fffffc);
} else {
@@ -2600,9 +2589,6 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
switch (opc) {
case INDEX_op_goto_ptr:
tcg_out32(s, MTSPR | RS(args[0]) | CTR);
- if (USE_REG_TB) {
- tcg_out_mov(s, TCG_TYPE_PTR, TCG_REG_TB, args[0]);
- }
tcg_out32(s, ADDI | TAI(TCG_REG_R3, 0, 0));
tcg_out32(s, BCCTR | BO_ALWAYS);
break;
--
2.34.1
next prev parent reply other threads:[~2023-10-18 22:28 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-18 22:25 [PULL 00/29] tcg patch queue Richard Henderson
2023-10-18 22:25 ` [PULL 01/29] tcg/ppc: Untabify tcg-target.c.inc Richard Henderson
2023-10-18 22:25 ` Richard Henderson [this message]
2023-10-18 22:25 ` [PULL 03/29] tcg/ppc: Reinterpret tb-relative to TB+4 Richard Henderson
2023-10-18 22:25 ` [PULL 04/29] tcg/ppc: Use ADDPCIS in tcg_out_tb_start Richard Henderson
2023-10-18 22:25 ` [PULL 05/29] tcg/ppc: Use ADDPCIS in tcg_out_movi_int Richard Henderson
2023-10-18 22:25 ` [PULL 06/29] tcg/ppc: Use ADDPCIS for the constant pool Richard Henderson
2023-10-18 22:25 ` [PULL 07/29] tcg/ppc: Use ADDPCIS in tcg_out_goto_tb Richard Henderson
2023-10-18 22:25 ` [PULL 08/29] tcg/ppc: Use PADDI in tcg_out_movi Richard Henderson
2023-10-18 22:25 ` [PULL 09/29] tcg/ppc: Use prefixed instructions in tcg_out_mem_long Richard Henderson
2023-10-18 22:25 ` [PULL 10/29] tcg/ppc: Use PLD in tcg_out_movi for constant pool Richard Henderson
2023-10-18 22:25 ` [PULL 11/29] tcg/ppc: Use prefixed instructions in tcg_out_dupi_vec Richard Henderson
2023-10-18 22:25 ` [PULL 12/29] tcg/ppc: Use PLD in tcg_out_goto_tb Richard Henderson
2023-10-18 22:25 ` [PULL 13/29] tcg/ppc: Disable TCG_REG_TB for Power9/Power10 Richard Henderson
2023-10-18 22:25 ` [PULL 14/29] tcg: Introduce tcg_use_softmmu Richard Henderson
2023-10-18 22:25 ` [PULL 15/29] tcg: Provide guest_base fallback for system mode Richard Henderson
2023-10-19 16:37 ` Stefan Hajnoczi
2023-10-18 22:25 ` [PULL 16/29] tcg/arm: Use tcg_use_softmmu Richard Henderson
2023-10-18 22:25 ` [PULL 17/29] tcg/aarch64: " Richard Henderson
2023-10-18 22:25 ` [PULL 18/29] tcg/i386: " Richard Henderson
2023-10-18 22:25 ` [PULL 19/29] tcg/loongarch64: " Richard Henderson
2023-10-18 22:25 ` [PULL 20/29] tcg/mips: " Richard Henderson
2023-10-18 22:25 ` [PULL 21/29] tcg/ppc: " Richard Henderson
2023-10-18 22:25 ` [PULL 22/29] tcg/riscv: Do not reserve TCG_GUEST_BASE_REG for guest_base zero Richard Henderson
2023-10-18 22:25 ` [PULL 23/29] tcg/riscv: Use tcg_use_softmmu Richard Henderson
2023-10-18 22:25 ` [PULL 24/29] tcg/s390x: " Richard Henderson
2023-10-18 22:25 ` [PULL 25/29] tcg: drop unused tcg_temp_free define Richard Henderson
2023-10-18 22:25 ` [PULL 26/29] tcg: Use constant zero when expanding with divu2 Richard Henderson
2023-10-18 22:25 ` [PULL 27/29] tcg: Optimize past conditional branches Richard Henderson
2023-10-18 22:25 ` [PULL 28/29] tcg: Add tcg_gen_{ld,st}_i128 Richard Henderson
2023-10-18 22:25 ` [PULL 29/29] target/i386: Use i128 for 128 and 256-bit loads and stores 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=20231018222557.1562065-3-richard.henderson@linaro.org \
--to=richard.henderson@linaro.org \
--cc=jniethe5@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).