linux-kernel - [PATCH v1 4/9] x86/resctrl: Flush MBM event counts on soft RMID change

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]

Message-ID: <20230421141723.2405942-5-peternewman@google.com>
Date:   Fri, 21 Apr 2023 16:17:18 +0200
From:   Peter Newman <peternewman@...gle.com>
To:     Fenghua Yu <fenghua.yu@...el.com>,
        Reinette Chatre <reinette.chatre@...el.com>
Cc:     Babu Moger <babu.moger@....com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        Dave Hansen <dave.hansen@...ux.intel.com>, x86@...nel.org,
        "H. Peter Anvin" <hpa@...or.com>,
        Stephane Eranian <eranian@...gle.com>,
        James Morse <james.morse@....com>,
        linux-kernel@...r.kernel.org, linux-kselftest@...r.kernel.org,
        Peter Newman <peternewman@...gle.com>
Subject: [PATCH v1 4/9] x86/resctrl: Flush MBM event counts on soft RMID change

To implement soft RMIDs, context switch must detect when the current
soft RMID is changing and if so, flush the CPU's MBM event counts to the
outgoing soft RMID.

To avoid impacting context switch performance in the non-soft RMID case,
protect the new logic with a static branch.

Co-developed-by: Stephane Eranian <eranian@...gle.com>
Signed-off-by: Stephane Eranian <eranian@...gle.com>
Signed-off-by: Peter Newman <peternewman@...gle.com>
---
 arch/x86/include/asm/resctrl.h         | 27 +++++++++++++++++++++++++-
 arch/x86/kernel/cpu/resctrl/rdtgroup.c |  1 +
 2 files changed, 27 insertions(+), 1 deletion(-)

diff --git a/arch/x86/include/asm/resctrl.h b/arch/x86/include/asm/resctrl.h
index e7acf118d770..50d05e883dbb 100644
--- a/arch/x86/include/asm/resctrl.h
+++ b/arch/x86/include/asm/resctrl.h
@@ -36,6 +36,9 @@ DECLARE_PER_CPU(struct resctrl_pqr_state, pqr_state);
 DECLARE_STATIC_KEY_FALSE(rdt_enable_key);
 DECLARE_STATIC_KEY_FALSE(rdt_alloc_enable_key);
 DECLARE_STATIC_KEY_FALSE(rdt_mon_enable_key);
+DECLARE_STATIC_KEY_FALSE(rdt_soft_rmid_enable_key);
+
+void resctrl_mbm_flush_cpu(void);
 
 /*
  * __resctrl_sched_in() - Writes the task's CLOSid/RMID to IA32_PQR_MSR
@@ -75,9 +78,31 @@ static inline void __resctrl_sched_in(struct task_struct *tsk)
 	}
 
 	if (closid != state->cur_closid || rmid != state->cur_rmid) {
+		if (static_branch_likely(&rdt_soft_rmid_enable_key)) {
+			/*
+			 * Flush current event counts to outgoing soft rmid
+			 * when it changes.
+			 */
+			if (rmid != state->cur_rmid)
+				resctrl_mbm_flush_cpu();
+
+			/*
+			 * rmid never changes in this mode, so skip wrmsr if the
+			 * closid is not changing.
+			 */
+			if (closid != state->cur_closid)
+				wrmsr(MSR_IA32_PQR_ASSOC, state->hw_rmid,
+				      closid);
+		} else {
+			wrmsr(MSR_IA32_PQR_ASSOC, rmid, closid);
+		}
+
+		/*
+		 * Record new closid/rmid last so soft rmid case can detect
+		 * changes.
+		 */
 		state->cur_closid = closid;
 		state->cur_rmid = rmid;
-		wrmsr(MSR_IA32_PQR_ASSOC, rmid, closid);
 	}
 }
 
diff --git a/arch/x86/kernel/cpu/resctrl/rdtgroup.c b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
index 6ad33f355861..c10f4798156a 100644
--- a/arch/x86/kernel/cpu/resctrl/rdtgroup.c
+++ b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
@@ -35,6 +35,7 @@
 DEFINE_STATIC_KEY_FALSE(rdt_enable_key);
 DEFINE_STATIC_KEY_FALSE(rdt_mon_enable_key);
 DEFINE_STATIC_KEY_FALSE(rdt_alloc_enable_key);
+DEFINE_STATIC_KEY_FALSE(rdt_soft_rmid_enable_key);
 static struct kernfs_root *rdt_root;
 struct rdtgroup rdtgroup_default;
 LIST_HEAD(rdt_all_groups);
-- 
2.40.0.634.g4ca3ef3211-goog