Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Mihai Carabas <mihai.carabas@oracle.com>
To:     linux-kernel@vger.kernel.org
Cc:     tglx@linutronix.de, mingo@redhat.com, bp@alien8.de, x86@kernel.org,
        boris.ostrovsky@oracle.com, konrad.wilk@oracle.com,
        Mihai Carabas <mihai.carabas@oracle.com>
Subject: [PATCH RFC 4/7] x86: cpu: bugs.c: update cpu_smt_disable to be callable at runtime
Date:   Thu,  2 Jul 2020 18:18:24 +0300
Message-Id: <1593703107-8852-5-git-send-email-mihai.carabas@oracle.com>
In-Reply-To: <1593703107-8852-1-git-send-email-mihai.carabas@oracle.com>
References: <1593703107-8852-1-git-send-email-mihai.carabas@oracle.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

If the microcode late loading and bug mitigation logic needs to turn
off SMT, it must use the hot plug infrastructure, not the boot time call
"cpu_smt_disable".

Update "cpu_smt_disable" to use the hot plug infrastructure to turn off
SMT when the system is in state running.

Signed-off-by: Mihai Carabas <mihai.carabas@oracle.com>
---
 kernel/cpu.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/kernel/cpu.c b/kernel/cpu.c
index fe67a01..719670f 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -393,9 +393,25 @@ void __weak arch_smt_update(void) { }
 
 void cpu_smt_disable(bool force)
 {
+	int ret;
+
 	if (!cpu_smt_possible())
 		return;
 
+	if (system_state == SYSTEM_RUNNING) {
+		if (force)
+			ret = cpuhp_smt_disable(CPU_SMT_FORCE_DISABLED);
+		else
+			ret = cpuhp_smt_disable(CPU_SMT_DISABLED);
+
+		/* If SMT disable did not succeed, print a warning*/
+		if (ret)
+			pr_warn("SMT: not disabled %d\n", ret);
+		else
+			pr_info("SMT:%s disabled\n", force ? " Force" : "");
+		return;
+	}
+
 	if (force) {
 		pr_info("SMT: Force disabled\n");
 		cpu_smt_control = CPU_SMT_FORCE_DISABLED;
-- 
1.8.3.1