Merge tag 'perf-core-2021-02-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull performance event updates from Ingo Molnar:

 - Add CPU-PMU support for Intel Sapphire Rapids CPUs

 - Extend the perf ABI with PERF_SAMPLE_WEIGHT_STRUCT, to offer
   two-parameter sampling event feedback. Not used yet, but is intended
   for Golden Cove CPU-PMU, which can provide both the instruction
   latency and the cache latency information for memory profiling
   events.

 - Remove experimental, default-disabled perfmon-v4 counter_freezing
   support that could only be enabled via a boot option. The hardware is
   hopelessly broken, we'd like to make sure nobody starts relying on
   this, as it would only end in tears.

 - Fix energy/power events on Intel SPR platforms

 - Simplify the uprobes resume_execution() logic

 - Misc smaller fixes.

* tag 'perf-core-2021-02-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/rapl: Fix psys-energy event on Intel SPR platform
  perf/x86/rapl: Only check lower 32bits for RAPL energy counters
  perf/x86/rapl: Add msr mask support
  perf/x86/kvm: Add Cascade Lake Xeon steppings to isolation_ucodes[]
  perf/x86/intel: Support CPUID 10.ECX to disable fixed counters
  perf/x86/intel: Add perf core PMU support for Sapphire Rapids
  perf/x86/intel: Filter unsupported Topdown metrics event
  perf/x86/intel: Factor out intel_update_topdown_event()
  perf/core: Add PERF_SAMPLE_WEIGHT_STRUCT
  perf/intel: Remove Perfmon-v4 counter_freezing support
  x86/perf: Use static_call for x86_pmu.guest_get_msrs
  perf/x86/intel/uncore: With > 8 nodes, get pci bus die id from NUMA info
  perf/x86/intel/uncore: Store the logical die id instead of the physical die id.
  x86/kprobes: Do not decode opcode in resume_execution()

1 files changed, 74 insertions, 94 deletions

diff --git a/arch/x86/kernel/kprobes/core.c b/arch/x86/kernel/kprobes/core.c
index a65e9e97857f..df776cdca327 100644
--- a/arch/x86/kernel/kprobes/core.c
+++ b/arch/x86/kernel/kprobes/core.c
@@ -133,26 +133,6 @@ void synthesize_relcall(void *dest, void *from, void *to)
 NOKPROBE_SYMBOL(synthesize_relcall);
 
 /*
- * Skip the prefixes of the instruction.
- */
-static kprobe_opcode_t *skip_prefixes(kprobe_opcode_t *insn)
-{
-	insn_attr_t attr;
-
-	attr = inat_get_opcode_attribute((insn_byte_t)*insn);
-	while (inat_is_legacy_prefix(attr)) {
-		insn++;
-		attr = inat_get_opcode_attribute((insn_byte_t)*insn);
-	}
-#ifdef CONFIG_X86_64
-	if (inat_is_rex_prefix(attr))
-		insn++;
-#endif
-	return insn;
-}
-NOKPROBE_SYMBOL(skip_prefixes);
-
-/*
  * Returns non-zero if INSN is boostable.
  * RIP relative instructions are adjusted at copying time in 64 bits mode
  */
@@ -312,25 +292,6 @@ static int can_probe(unsigned long paddr)
 }
 
 /*
- * Returns non-zero if opcode modifies the interrupt flag.
- */
-static int is_IF_modifier(kprobe_opcode_t *insn)
-{
-	/* Skip prefixes */
-	insn = skip_prefixes(insn);
-
-	switch (*insn) {
-	case 0xfa:		/* cli */
-	case 0xfb:		/* sti */
-	case 0xcf:		/* iret/iretd */
-	case 0x9d:		/* popf/popfd */
-		return 1;
-	}
-
-	return 0;
-}
-
-/*
  * Copy an instruction with recovering modified instruction by kprobes
  * and adjust the displacement if the instruction uses the %rip-relative
  * addressing mode. Note that since @real will be the final place of copied
@@ -411,9 +372,9 @@ static int prepare_boost(kprobe_opcode_t *buf, struct kprobe *p,
 		synthesize_reljump(buf + len, p->ainsn.insn + len,
 				   p->addr + insn->length);
 		len += JMP32_INSN_SIZE;
-		p->ainsn.boostable = true;
+		p->ainsn.boostable = 1;
 	} else {
-		p->ainsn.boostable = false;
+		p->ainsn.boostable = 0;
 	}
 
 	return len;
@@ -450,6 +411,67 @@ void free_insn_page(void *page)
 	module_memfree(page);
 }
 
+static void set_resume_flags(struct kprobe *p, struct insn *insn)
+{
+	insn_byte_t opcode = insn->opcode.bytes[0];
+
+	switch (opcode) {
+	case 0xfa:		/* cli */
+	case 0xfb:		/* sti */
+	case 0x9d:		/* popf/popfd */
+		/* Check whether the instruction modifies Interrupt Flag or not */
+		p->ainsn.if_modifier = 1;
+		break;
+	case 0x9c:	/* pushfl */
+		p->ainsn.is_pushf = 1;
+		break;
+	case 0xcf:	/* iret */
+		p->ainsn.if_modifier = 1;
+		fallthrough;
+	case 0xc2:	/* ret/lret */
+	case 0xc3:
+	case 0xca:
+	case 0xcb:
+	case 0xea:	/* jmp absolute -- ip is correct */
+		/* ip is already adjusted, no more changes required */
+		p->ainsn.is_abs_ip = 1;
+		/* Without resume jump, this is boostable */
+		p->ainsn.boostable = 1;
+		break;
+	case 0xe8:	/* call relative - Fix return addr */
+		p->ainsn.is_call = 1;
+		break;
+#ifdef CONFIG_X86_32
+	case 0x9a:	/* call absolute -- same as call absolute, indirect */
+		p->ainsn.is_call = 1;
+		p->ainsn.is_abs_ip = 1;
+		break;
+#endif
+	case 0xff:
+		opcode = insn->opcode.bytes[1];
+		if ((opcode & 0x30) == 0x10) {
+			/*
+			 * call absolute, indirect
+			 * Fix return addr; ip is correct.
+			 * But this is not boostable
+			 */
+			p->ainsn.is_call = 1;
+			p->ainsn.is_abs_ip = 1;
+			break;
+		} else if (((opcode & 0x31) == 0x20) ||
+			   ((opcode & 0x31) == 0x21)) {
+			/*
+			 * jmp near and far, absolute indirect
+			 * ip is correct.
+			 */
+			p->ainsn.is_abs_ip = 1;
+			/* Without resume jump, this is boostable */
+			p->ainsn.boostable = 1;
+		}
+		break;
+	}
+}
+
 static int arch_copy_kprobe(struct kprobe *p)
 {
 	struct insn insn;
@@ -467,8 +489,8 @@ static int arch_copy_kprobe(struct kprobe *p)
 	 */
 	len = prepare_boost(buf, p, &insn);
 
-	/* Check whether the instruction modifies Interrupt Flag or not */
-	p->ainsn.if_modifier = is_IF_modifier(buf);
+	/* Analyze the opcode and set resume flags */
+	set_resume_flags(p, &insn);
 
 	/* Also, displacement change doesn't affect the first byte */
 	p->opcode = buf[0];
@@ -491,6 +513,9 @@ int arch_prepare_kprobe(struct kprobe *p)
 
 	if (!can_probe((unsigned long)p->addr))
 		return -EILSEQ;
+
+	memset(&p->ainsn, 0, sizeof(p->ainsn));
+
 	/* insn: must be on special executable page on x86. */
 	p->ainsn.insn = get_insn_slot();
 	if (!p->ainsn.insn)
@@ -806,11 +831,6 @@ NOKPROBE_SYMBOL(trampoline_handler);
  * 2) If the single-stepped instruction was a call, the return address
  * that is atop the stack is the address following the copied instruction.
  * We need to make it the address following the original instruction.
- *
- * If this is the first time we've single-stepped the instruction at
- * this probepoint, and the instruction is boostable, boost it: add a
- * jump instruction after the copied instruction, that jumps to the next
- * instruction after the probepoint.
  */
 static void resume_execution(struct kprobe *p, struct pt_regs *regs,
 			     struct kprobe_ctlblk *kcb)
@@ -818,60 +838,20 @@ static void resume_execution(struct kprobe *p, struct pt_regs *regs,
 	unsigned long *tos = stack_addr(regs);
 	unsigned long copy_ip = (unsigned long)p->ainsn.insn;
 	unsigned long orig_ip = (unsigned long)p->addr;
-	kprobe_opcode_t *insn = p->ainsn.insn;
-
-	/* Skip prefixes */
-	insn = skip_prefixes(insn);
 
 	regs->flags &= ~X86_EFLAGS_TF;
-	switch (*insn) {
-	case 0x9c:	/* pushfl */
+
+	/* Fixup the contents of top of stack */
+	if (p->ainsn.is_pushf) {
 		*tos &= ~(X86_EFLAGS_TF | X86_EFLAGS_IF);
 		*tos |= kcb->kprobe_old_flags;
-		break;
-	case 0xc2:	/* iret/ret/lret */
-	case 0xc3:
-	case 0xca:
-	case 0xcb:
-	case 0xcf:
-	case 0xea:	/* jmp absolute -- ip is correct */
-		/* ip is already adjusted, no more changes required */
-		p->ainsn.boostable = true;
-		goto no_change;
-	case 0xe8:	/* call relative - Fix return addr */
+	} else if (p->ainsn.is_call) {
 		*tos = orig_ip + (*tos - copy_ip);
-		break;
-#ifdef CONFIG_X86_32
-	case 0x9a:	/* call absolute -- same as call absolute, indirect */
-		*tos = orig_ip + (*tos - copy_ip);
-		goto no_change;
-#endif
-	case 0xff:
-		if ((insn[1] & 0x30) == 0x10) {
-			/*
-			 * call absolute, indirect
-			 * Fix return addr; ip is correct.
-			 * But this is not boostable
-			 */
-			*tos = orig_ip + (*tos - copy_ip);
-			goto no_change;
-		} else if (((insn[1] & 0x31) == 0x20) ||
-			   ((insn[1] & 0x31) == 0x21)) {
-			/*
-			 * jmp near and far, absolute indirect
-			 * ip is correct. And this is boostable
-			 */
-			p->ainsn.boostable = true;
-			goto no_change;
-		}
-		break;
-	default:
-		break;
 	}
 
-	regs->ip += orig_ip - copy_ip;
+	if (!p->ainsn.is_abs_ip)
+		regs->ip += orig_ip - copy_ip;
 
-no_change:
 	restore_btf();
 }
 NOKPROBE_SYMBOL(resume_execution);