From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: Peter Maydell <peter.maydell@linaro.org>
Subject: [PULL 38/39] tcg/aarch64: Emit BTI insns at jump landing pads
Date: Fri, 15 Sep 2023 20:30:10 -0700 [thread overview]
Message-ID: <20230916033011.479144-39-richard.henderson@linaro.org> (raw)
In-Reply-To: <20230916033011.479144-1-richard.henderson@linaro.org>
The prologue is entered via "call"; the epilogue, each tb,
and each goto_tb continuation point are all reached via "jump".
As tcg_out_goto_long is only used by tcg_out_exit_tb, merge
the two functions. Change the indirect register used to
TCG_REG_TMP1, aka X17, so that the BTI condition created
is "jump" instead of "jump or call".
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/aarch64/tcg-target.c.inc | 54 ++++++++++++++++++++++++++----------
1 file changed, 39 insertions(+), 15 deletions(-)
diff --git a/tcg/aarch64/tcg-target.c.inc b/tcg/aarch64/tcg-target.c.inc
index a0b65029d4..06ea3c7652 100644
--- a/tcg/aarch64/tcg-target.c.inc
+++ b/tcg/aarch64/tcg-target.c.inc
@@ -602,6 +602,10 @@ typedef enum {
DMB_ISH = 0xd50338bf,
DMB_LD = 0x00000100,
DMB_ST = 0x00000200,
+
+ BTI_C = 0xd503245f,
+ BTI_J = 0xd503249f,
+ BTI_JC = 0xd50324df,
} AArch64Insn;
static inline uint32_t tcg_in32(TCGContext *s)
@@ -843,6 +847,17 @@ static void tcg_out_insn_3313(TCGContext *s, AArch64Insn insn,
| rn << 5 | (rd & 0x1f));
}
+static void tcg_out_bti(TCGContext *s, AArch64Insn insn)
+{
+ /*
+ * While BTI insns are nops on hosts without FEAT_BTI,
+ * there is no point in emitting them in that case either.
+ */
+ if (cpuinfo & CPUINFO_BTI) {
+ tcg_out32(s, insn);
+ }
+}
+
/* Register to register move using ORR (shifted register with no shift). */
static void tcg_out_movr(TCGContext *s, TCGType ext, TCGReg rd, TCGReg rm)
{
@@ -1351,18 +1366,6 @@ static void tcg_out_goto(TCGContext *s, const tcg_insn_unit *target)
tcg_out_insn(s, 3206, B, offset);
}
-static void tcg_out_goto_long(TCGContext *s, const tcg_insn_unit *target)
-{
- ptrdiff_t offset = tcg_pcrel_diff(s, target) >> 2;
- if (offset == sextract64(offset, 0, 26)) {
- tcg_out_insn(s, 3206, B, offset);
- } else {
- /* Choose X9 as a call-clobbered non-LR temporary. */
- tcg_out_movi(s, TCG_TYPE_I64, TCG_REG_X9, (intptr_t)target);
- tcg_out_insn(s, 3207, BR, TCG_REG_X9);
- }
-}
-
static void tcg_out_call_int(TCGContext *s, const tcg_insn_unit *target)
{
ptrdiff_t offset = tcg_pcrel_diff(s, target) >> 2;
@@ -1947,12 +1950,28 @@ static const tcg_insn_unit *tb_ret_addr;
static void tcg_out_exit_tb(TCGContext *s, uintptr_t a0)
{
+ const tcg_insn_unit *target;
+ ptrdiff_t offset;
+
/* Reuse the zeroing that exists for goto_ptr. */
if (a0 == 0) {
- tcg_out_goto_long(s, tcg_code_gen_epilogue);
+ target = tcg_code_gen_epilogue;
} else {
tcg_out_movi(s, TCG_TYPE_I64, TCG_REG_X0, a0);
- tcg_out_goto_long(s, tb_ret_addr);
+ target = tb_ret_addr;
+ }
+
+ offset = tcg_pcrel_diff(s, target) >> 2;
+ if (offset == sextract64(offset, 0, 26)) {
+ tcg_out_insn(s, 3206, B, offset);
+ } else {
+ /*
+ * Only x16/x17 generate BTI type Jump (2),
+ * other registers generate BTI type Jump|Call (3).
+ */
+ QEMU_BUILD_BUG_ON(TCG_REG_TMP0 != TCG_REG_X16);
+ tcg_out_movi(s, TCG_TYPE_I64, TCG_REG_TMP0, (intptr_t)target);
+ tcg_out_insn(s, 3207, BR, TCG_REG_TMP0);
}
}
@@ -1970,6 +1989,7 @@ static void tcg_out_goto_tb(TCGContext *s, int which)
tcg_out32(s, I3206_B);
tcg_out_insn(s, 3207, BR, TCG_REG_TMP0);
set_jmp_reset_offset(s, which);
+ tcg_out_bti(s, BTI_J);
}
void tb_target_set_jmp_target(const TranslationBlock *tb, int n,
@@ -3074,6 +3094,8 @@ static void tcg_target_qemu_prologue(TCGContext *s)
{
TCGReg r;
+ tcg_out_bti(s, BTI_C);
+
/* Push (FP, LR) and allocate space for all saved registers. */
tcg_out_insn(s, 3314, STP, TCG_REG_FP, TCG_REG_LR,
TCG_REG_SP, -PUSH_SIZE, 1, 1);
@@ -3114,10 +3136,12 @@ static void tcg_target_qemu_prologue(TCGContext *s)
* and fall through to the rest of the epilogue.
*/
tcg_code_gen_epilogue = tcg_splitwx_to_rx(s->code_ptr);
+ tcg_out_bti(s, BTI_J);
tcg_out_movi(s, TCG_TYPE_REG, TCG_REG_X0, 0);
/* TB epilogue */
tb_ret_addr = tcg_splitwx_to_rx(s->code_ptr);
+ tcg_out_bti(s, BTI_J);
/* Remove TCG locals stack space. */
tcg_out_insn(s, 3401, ADDI, TCG_TYPE_I64, TCG_REG_SP, TCG_REG_SP,
@@ -3137,7 +3161,7 @@ static void tcg_target_qemu_prologue(TCGContext *s)
static void tcg_out_tb_start(TCGContext *s)
{
- /* nothing to do */
+ tcg_out_bti(s, BTI_J);
}
static void tcg_out_nop_fill(tcg_insn_unit *p, int count)
--
2.34.1
next prev parent reply other threads:[~2023-09-16 3:37 UTC|newest]
Thread overview: 45+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-09-16 3:29 [PULL 00/39] tcg patch queue Richard Henderson
2023-09-16 3:29 ` [PULL 01/39] accel/tcg: mttcg remove false-negative halted assertion Richard Henderson
2023-09-18 6:44 ` Nicholas Piggin
2023-09-18 7:59 ` Alex Bennée
2023-09-18 10:53 ` Nicholas Piggin
2023-09-18 12:19 ` Alex Bennée
2023-09-16 3:29 ` [PULL 02/39] accel/tcg: Fix the comment for CPUTLBEntryFull Richard Henderson
2023-09-16 3:29 ` [PULL 03/39] util: Delete checks for old host definitions Richard Henderson
2023-09-16 3:29 ` [PULL 04/39] softmmu: " Richard Henderson
2023-09-16 3:29 ` [PULL 05/39] thunk: " Richard Henderson
2023-09-16 3:29 ` [PULL 06/39] tcg/loongarch64: Import LSX instructions Richard Henderson
2023-09-16 3:29 ` [PULL 07/39] tcg/loongarch64: Lower basic tcg vec ops to LSX Richard Henderson
2023-09-16 3:29 ` [PULL 08/39] tcg: pass vece to tcg_target_const_match() Richard Henderson
2023-09-16 3:29 ` [PULL 09/39] tcg/loongarch64: Lower cmp_vec to vseq/vsle/vslt Richard Henderson
2023-09-16 3:29 ` [PULL 10/39] tcg/loongarch64: Lower add/sub_vec to vadd/vsub Richard Henderson
2023-09-16 3:29 ` [PULL 11/39] tcg/loongarch64: Lower vector bitwise operations Richard Henderson
2023-09-16 3:29 ` [PULL 12/39] tcg/loongarch64: Lower neg_vec to vneg Richard Henderson
2023-09-16 3:29 ` [PULL 13/39] tcg/loongarch64: Lower mul_vec to vmul Richard Henderson
2023-09-16 3:29 ` [PULL 14/39] tcg/loongarch64: Lower vector min max ops Richard Henderson
2023-09-16 3:29 ` [PULL 15/39] tcg/loongarch64: Lower vector saturated ops Richard Henderson
2023-09-16 3:29 ` [PULL 16/39] tcg/loongarch64: Lower vector shift vector ops Richard Henderson
2023-09-16 3:29 ` [PULL 17/39] tcg/loongarch64: Lower bitsel_vec to vbitsel Richard Henderson
2023-09-16 3:29 ` [PULL 18/39] tcg/loongarch64: Lower vector shift integer ops Richard Henderson
2023-09-16 3:29 ` [PULL 19/39] tcg/loongarch64: Lower rotv_vec ops to LSX Richard Henderson
2023-09-16 3:29 ` [PULL 20/39] tcg/loongarch64: Lower rotli_vec to vrotri Richard Henderson
2023-09-16 3:29 ` [PULL 21/39] tcg/loongarch64: Implement 128-bit load & store Richard Henderson
2023-09-16 3:29 ` [PULL 22/39] tcg: Add gvec compare with immediate and scalar operand Richard Henderson
2023-09-16 3:29 ` [PULL 23/39] target/arm: Use tcg_gen_gvec_cmpi for compare vs 0 Richard Henderson
2023-09-16 3:29 ` [PULL 24/39] accel/tcg: Simplify tlb_plugin_lookup Richard Henderson
2023-09-16 3:29 ` [PULL 25/39] accel/tcg: Split out io_prepare and io_failed Richard Henderson
2023-09-16 3:29 ` [PULL 26/39] accel/tcg: Use CPUTLBEntryFull.phys_addr in io_failed Richard Henderson
2023-09-16 3:29 ` [PULL 27/39] plugin: Simplify struct qemu_plugin_hwaddr Richard Henderson
2023-09-16 3:30 ` [PULL 28/39] accel/tcg: Merge cpu_transaction_failed into io_failed Richard Henderson
2023-09-16 3:30 ` [PULL 29/39] accel/tcg: Replace direct use of io_readx/io_writex in do_{ld, st}_1 Richard Henderson
2023-09-16 3:30 ` [PULL 30/39] accel/tcg: Merge io_readx into do_ld_mmio_beN Richard Henderson
2023-09-16 3:30 ` [PULL 31/39] accel/tcg: Merge io_writex into do_st_mmio_leN Richard Henderson
2023-09-16 3:30 ` [PULL 32/39] accel/tcg: Introduce do_ld16_mmio_beN Richard Henderson
2023-09-16 3:30 ` [PULL 33/39] accel/tcg: Introduce do_st16_mmio_leN Richard Henderson
2023-09-16 3:30 ` [PULL 34/39] fpu: Add conversions between bfloat16 and [u]int8 Richard Henderson
2023-09-16 3:30 ` [PULL 35/39] fpu: Handle m68k extended precision denormals properly Richard Henderson
2023-09-16 3:30 ` [PULL 36/39] tcg: Add tcg_out_tb_start backend hook Richard Henderson
2023-09-16 3:30 ` [PULL 37/39] util/cpuinfo-aarch64: Add CPUINFO_BTI Richard Henderson
2023-09-16 3:30 ` Richard Henderson [this message]
2023-09-16 3:30 ` [PULL 39/39] tcg: Map code_gen_buffer with PROT_BTI Richard Henderson
2023-09-16 4:07 ` [PULL 00/39] tcg patch queue 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=20230916033011.479144-39-richard.henderson@linaro.org \
--to=richard.henderson@linaro.org \
--cc=peter.maydell@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).