linux-kernel - [tip: perf/urgent] perf/x86/intel/lbr: Add static

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 18 Jan 2022 11:17:50 -0000
From:   "tip-bot2 for Peter Zijlstra (Intel)" <tip-bot2@...utronix.de>
To:     linux-tip-commits@...r.kernel.org
Cc:     "Peter Zijlstra (Intel)" <peterz@...radead.org>,
        Kan Liang <kan.liang@...ux.intel.com>, x86@...nel.org,
        linux-kernel@...r.kernel.org
Subject: [tip: perf/urgent] perf/x86/intel/lbr: Add static_branch for LBR INFO flags

The following commit has been merged into the perf/urgent branch of tip:

Commit-ID:     6b19788ddc5937831ffd27525a1b793953fd2d2b
Gitweb:        https://git.kernel.org/tip/6b19788ddc5937831ffd27525a1b793953fd2d2b
Author:        Peter Zijlstra (Intel) <peterz@...radead.org>
AuthorDate:    Tue, 04 Jan 2022 08:51:17 -08:00
Committer:     Peter Zijlstra <peterz@...radead.org>
CommitterDate: Tue, 18 Jan 2022 12:09:49 +01:00

perf/x86/intel/lbr: Add static_branch for LBR INFO flags

Using static_branch to replace the LBR INFO flags to optimize the LBR
INFO parsing.

Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
Signed-off-by: Kan Liang <kan.liang@...ux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
Tested-by: Kan Liang <kan.liang@...ux.intel.com>
Link: https://lkml.kernel.org/r/1641315077-96661-2-git-send-email-peterz@infradead.org
---
 arch/x86/events/intel/lbr.c | 51 ++++++++++++++++++++++++------------
 1 file changed, 34 insertions(+), 17 deletions(-)

diff --git a/arch/x86/events/intel/lbr.c b/arch/x86/events/intel/lbr.c
index b7228a1..f8fd255 100644
--- a/arch/x86/events/intel/lbr.c
+++ b/arch/x86/events/intel/lbr.c
@@ -893,37 +893,40 @@ void intel_pmu_lbr_read_64(struct cpu_hw_events *cpuc)
 	cpuc->lbr_stack.hw_idx = tos;
 }
 
+static DEFINE_STATIC_KEY_FALSE(x86_lbr_mispred);
+static DEFINE_STATIC_KEY_FALSE(x86_lbr_cycles);
+static DEFINE_STATIC_KEY_FALSE(x86_lbr_type);
+
 static __always_inline int get_lbr_br_type(u64 info)
 {
-	if (!static_cpu_has(X86_FEATURE_ARCH_LBR) || !x86_pmu.lbr_br_type)
-		return 0;
+	int type = 0;
 
-	return (info & LBR_INFO_BR_TYPE) >> LBR_INFO_BR_TYPE_OFFSET;
+	if (static_branch_likely(&x86_lbr_type))
+		type = (info & LBR_INFO_BR_TYPE) >> LBR_INFO_BR_TYPE_OFFSET;
+
+	return type;
 }
 
 static __always_inline bool get_lbr_mispred(u64 info)
 {
-	if (static_cpu_has(X86_FEATURE_ARCH_LBR) && !x86_pmu.lbr_mispred)
-		return 0;
+	bool mispred = 0;
 
-	return !!(info & LBR_INFO_MISPRED);
-}
-
-static __always_inline bool get_lbr_predicted(u64 info)
-{
-	if (static_cpu_has(X86_FEATURE_ARCH_LBR) && !x86_pmu.lbr_mispred)
-		return 0;
+	if (static_branch_likely(&x86_lbr_mispred))
+		mispred = !!(info & LBR_INFO_MISPRED);
 
-	return !(info & LBR_INFO_MISPRED);
+	return mispred;
 }
 
 static __always_inline u16 get_lbr_cycles(u64 info)
 {
+	u16 cycles = info & LBR_INFO_CYCLES;
+
 	if (static_cpu_has(X86_FEATURE_ARCH_LBR) &&
-	    !(x86_pmu.lbr_timed_lbr && info & LBR_INFO_CYC_CNT_VALID))
-		return 0;
+	    (!static_branch_likely(&x86_lbr_cycles) ||
+	     !(info & LBR_INFO_CYC_CNT_VALID)))
+		cycles = 0;
 
-	return info & LBR_INFO_CYCLES;
+	return cycles;
 }
 
 static void intel_pmu_store_lbr(struct cpu_hw_events *cpuc,
@@ -951,7 +954,7 @@ static void intel_pmu_store_lbr(struct cpu_hw_events *cpuc,
 		e->from		= from;
 		e->to		= to;
 		e->mispred	= get_lbr_mispred(info);
-		e->predicted	= get_lbr_predicted(info);
+		e->predicted	= !e->mispred;
 		e->in_tx	= !!(info & LBR_INFO_IN_TX);
 		e->abort	= !!(info & LBR_INFO_ABORT);
 		e->cycles	= get_lbr_cycles(info);
@@ -1718,6 +1721,14 @@ void intel_pmu_lbr_init(void)
 		x86_pmu.lbr_to_cycles = 1;
 		break;
 	}
+
+	if (x86_pmu.lbr_has_info) {
+		/*
+		 * Only used in combination with baseline pebs.
+		 */
+		static_branch_enable(&x86_lbr_mispred);
+		static_branch_enable(&x86_lbr_cycles);
+	}
 }
 
 /*
@@ -1779,6 +1790,12 @@ void __init intel_pmu_arch_lbr_init(void)
 	x86_pmu.lbr_br_type = ecx.split.lbr_br_type;
 	x86_pmu.lbr_nr = lbr_nr;
 
+	if (x86_pmu.lbr_mispred)
+		static_branch_enable(&x86_lbr_mispred);
+	if (x86_pmu.lbr_timed_lbr)
+		static_branch_enable(&x86_lbr_cycles);
+	if (x86_pmu.lbr_br_type)
+		static_branch_enable(&x86_lbr_type);
 
 	arch_lbr_xsave = is_arch_lbr_xsave_available();
 	if (arch_lbr_xsave) {