linux-kernel - Re: [PATCH tip] x86/microcode/AMD: Read from MSR_AMD64_PATCH_LEVEL to get base

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 for Android: free password hash cracker in your pocket

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

Message-ID: <a9bdc5cb-28f2-4d0b-a823-4039fdaf0aa1@redhat.com>
Date: Mon, 17 Nov 2025 14:58:30 -0500
From: Waiman Long <llong@...hat.com>
To: Borislav Petkov <bp@...en8.de>
Cc: Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...hat.com>,
 Dave Hansen <dave.hansen@...ux.intel.com>, "H. Peter Anvin" <hpa@...or.com>,
 x86@...nel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH tip] x86/microcode/AMD: Read from MSR_AMD64_PATCH_LEVEL to
 get base_rev if not defined

On 11/17/25 2:37 PM, Borislav Petkov wrote:
> On Mon, Nov 17, 2025 at 02:15:27PM -0500, Waiman Long wrote:
>> AMD microcode is being handled inconsistenly depending on whether the
>> CONFIG_MICROCODE_DBG is on or off.
> Did you read the help text to CONFIG_MICROCODE_DBG?
>
I am fine if "microcode,base_rev=" is specified in the boot command line 
to force a base_rev for testing purpose. But if that is not specified, 
get_patch_level() should return the right value. The problem I have on 
that particular test machine mentioned in the commit is that the kernel 
is trying to load an incompatible microcode blob causing error like the 
following when CONFIG_MICROCODE_DBG is on.

[    0.000000] unchecked MSR access error: WRMSR to 0xc0010020 (tried to 
write 0xff11000026d3e740) at rIP: 0xffffffffa72cc63b 
(__apply_microcode_amd+0x3b/0xb0)
[    0.000000] Call Trace:
[    0.000000]  <TASK>
[    0.000000]  ? show_trace_log_lvl+0x1b0/0x2f0
[    0.000000]  ? show_trace_log_lvl+0x1b0/0x2f0
[    0.000000]  ? ex_handler_msr.isra.0.cold+0x5b/0x60
[    0.000000]  ? fixup_exception+0x8f/0x380
[    0.000000]  ? early_fixup_exception+0x45/0xb0
[    0.000000]  ? early_idt_handler_common+0x2f/0x3a
[    0.000000]  ? __apply_microcode_amd+0x3b/0xb0
[    0.000000]  ? load_ucode_amd_bsp+0x10b/0x140
[    0.000000]  ? x86_64_start_kernel+0x84/0xa0
[    0.000000]  ? common_startup_64+0x13e/0x141
[    0.000000]  </TASK>
[    0.000000] microcode: updated rev: 0xa101148

I can add a boot_cpu_has(X86_FEATURE_HYPERVISOR) check and use 
native_rdmsr() only if there is no hypervisor running. Do you think that 
is acceptable?

Regards,
Longman