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;
MIME-Version: 1.0
In-Reply-To: <3b274c9a58ddfba8bd37827519aec541d1838697.camel@intel.com>
References: <20220614210217.1940563-1-martin.fernandez@eclypsium.com> <3b274c9a58ddfba8bd37827519aec541d1838697.camel@intel.com>
From:   Martin Fernandez <martin.fernandez@eclypsium.com>
Date:   Thu, 16 Jun 2022 11:23:16 -0300
Message-ID: <CAKgze5YXqi_+bbxCKd_yped5Vx=H8u0dc7uXH3WdF4T5oTOPFQ@mail.gmail.com>
Subject: Re: [PATCH] x86/cpuinfo: Clear X86_FEATURE_TME if TME/MKTME is
 disabled by BIOS
To:     Kai Huang <kai.huang@intel.com>
Cc:     linux-kernel@vger.kernel.org, bp@alien8.de,
        dave.hansen@linux.intel.com, x86@kernel.org, mingo@redhat.com,
        tglx@linutronix.de, daniel.gutson@eclypsium.com,
        hughsient@gmail.com, alex.bazhaniuk@eclypsium.com
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On 6/15/22, Kai Huang <kai.huang@intel.com> wrote:
> On Tue, 2022-06-14 at 18:02 -0300, Martin Fernandez wrote:
>> Right now the only way to check this is by greping the kernel logs,
>> which is inconvinient. This is currently checked for fwupd for
>> example.
>>
>> I understand that cpuinfo is supposed to report every feature in the
>> cpu but since AMD is doing the same for sme/sev I think is good to
>> have this for Intel too.
>>
>> Signed-off-by: Martin Fernandez <martin.fernandez@eclypsium.com>
>> ---
>>  arch/x86/kernel/cpu/intel.c | 2 ++
>>  1 file changed, 2 insertions(+)
>>
>> diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c
>> index fd5dead8371c..7311172aceaf 100644
>> --- a/arch/x86/kernel/cpu/intel.c
>> +++ b/arch/x86/kernel/cpu/intel.c
>> @@ -570,6 +570,8 @@ static void detect_tme(struct cpuinfo_x86 *c)
>>
>>  	if (!TME_ACTIVATE_LOCKED(tme_activate) ||
>> !TME_ACTIVATE_ENABLED(tme_activate)) {
>>  		pr_info_once("x86/tme: not enabled by BIOS\n");
>> +		if (mktme_status == MKTME_UNINITIALIZED)
>> +			clear_cpu_cap(c, X86_FEATURE_TME);
>
> The above code seems only clear TME on BSP, if I read correctly.  Why not
> unconditionally clear this flag?  If TME is detected as not enabled by BIOS
> on
> *this* cpu, then the flag for *this* cpu should be cleared.

Yes, I think you're right. Thanks.

> Also could you CC Kirill A. Shutemov <kirill.shutemov@linux.intel.com>, who
> is
> the original author of this code?

I'll add Kirill for the next version, thank you for mention it.

>>  		mktme_status = MKTME_DISABLED;
>>  		return;
>>  	}
>> --
>> 2.30.2
>>
>
> --
> Thanks,
> -Kai
>
>
>