linux-kernel - [PATCH -v2] x86/microcode: Do not taint when late loading on AMD

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: <20230411125056.2333-1-bp@alien8.de>
Date:   Tue, 11 Apr 2023 14:50:55 +0200
From:   Borislav Petkov <bp@...en8.de>
To:     X86 ML <x86@...nel.org>
Cc:     LKML <linux-kernel@...r.kernel.org>
Subject: [PATCH -v2] x86/microcode: Do not taint when late loading on AMD

From: "Borislav Petkov (AMD)" <bp@...en8.de>

Describe why the concurrency issues which late loading poses are not
affecting AMD hardware, after discussing it with hw folks. Thus, do not
taint when late loading on it.

Signed-off-by: Borislav Petkov (AMD) <bp@...en8.de>
---
 Documentation/x86/microcode.rst      | 10 ++++++++++
 arch/x86/kernel/cpu/microcode/core.c |  9 +++++----
 2 files changed, 15 insertions(+), 4 deletions(-)

diff --git a/Documentation/x86/microcode.rst b/Documentation/x86/microcode.rst
index b627c6f36bcf..15b52e2b181d 100644
--- a/Documentation/x86/microcode.rst
+++ b/Documentation/x86/microcode.rst
@@ -208,6 +208,16 @@ Basically there is no way to declare a new microcode update suitable
 for late-loading. This is another one of the problems that caused late
 loading to be not enabled by default.
 
+AMD
+---
+
+Late loading on AMD does not have the concurrency issues described
+above: when loading is attempted on T0, the T1 is quiesced and does not
+execute instructions. Therefore, even if a higher priority interrupt or
+a fault happens, the whole core will see it either before the microcode
+patch has been applied or after. In either case, T0 and T1 will have the
+same microcode revision and nothing intermediate.
+
 Builtin microcode
 =================
 
diff --git a/arch/x86/kernel/cpu/microcode/core.c b/arch/x86/kernel/cpu/microcode/core.c
index 7a329e561354..22cf57c899b6 100644
--- a/arch/x86/kernel/cpu/microcode/core.c
+++ b/arch/x86/kernel/cpu/microcode/core.c
@@ -440,9 +440,6 @@ static int microcode_reload_late(void)
 	int old = boot_cpu_data.microcode, ret;
 	struct cpuinfo_x86 prev_info;
 
-	pr_err("Attempting late microcode loading - it is dangerous and taints the kernel.\n");
-	pr_err("You should switch to early loading, if possible.\n");
-
 	atomic_set(&late_cpus_in,  0);
 	atomic_set(&late_cpus_out, 0);
 
@@ -498,7 +495,11 @@ static ssize_t reload_store(struct device *dev,
 	if (ret == 0)
 		ret = size;
 
-	add_taint(TAINT_CPU_OUT_OF_SPEC, LOCKDEP_STILL_OK);
+	if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) {
+		pr_err("Late microcode loading is dangerous and taints the kernel.\n");
+		pr_err("You should switch to early loading if possible.\n");
+		add_taint(TAINT_CPU_OUT_OF_SPEC, LOCKDEP_STILL_OK);
+	}
 
 	return ret;
 }
-- 
2.35.1