Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Muhammad Usama Anjum <usama.anjum@collabora.com>
To:     Jonathan Corbet <corbet@lwn.net>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        x86@kernel.org (maintainer:X86 ARCHITECTURE (32-BIT AND 64-BIT)),
        "H. Peter Anvin" <hpa@zytor.com>,
        linux-doc@vger.kernel.org (open list:DOCUMENTATION),
        linux-kernel@vger.kernel.org (open list)
Cc:     Steven Noonan <steven@uplinklabs.net>, usama.anjum@collabora.com,
        kernel@collabora.com
Subject: [PATCH 3/3] x86/tsc: don't check for random warps if using direct sync
Date:   Mon,  8 Aug 2022 16:39:54 +0500
Message-Id: <20220808113954.345579-3-usama.anjum@collabora.com>
In-Reply-To: <20220808113954.345579-1-usama.anjum@collabora.com>
References: <20220808113954.345579-1-usama.anjum@collabora.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Steven Noonan <steven@uplinklabs.net>

There's some overhead in writing and reading MSR_IA32_TSC. We try to
account for it. But sometimes overhead gets under or over estimated.
When we retry syncing, it sees the clock "go backwards". Hence,
ignore random wrap if using direct sync.

Signed-off-by: Steven Noonan <steven@uplinklabs.net>
Signed-off-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
---
 arch/x86/kernel/tsc_sync.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/kernel/tsc_sync.c b/arch/x86/kernel/tsc_sync.c
index 2a855991f982..1fc751212a0e 100644
--- a/arch/x86/kernel/tsc_sync.c
+++ b/arch/x86/kernel/tsc_sync.c
@@ -405,7 +405,7 @@ void check_tsc_sync_source(int cpu)
 		pr_debug("TSC synchronization [CPU#%d -> CPU#%d]: passed\n",
 			smp_processor_id(), cpu);
 
-	} else if (atomic_dec_and_test(&test_runs) || random_warps) {
+	} else if (atomic_dec_and_test(&test_runs) || (random_warps && !tsc_allow_direct_sync)) {
 		/* Force it to 0 if random warps brought us here */
 		atomic_set(&test_runs, 0);
 
-- 
2.30.2