Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH v5 03/10] arm64: add sysfs vulnerability show for meltdown
To:     Andre Przywara <andre.przywara@arm.com>,
        linux-arm-kernel@lists.infradead.org
Cc:     catalin.marinas@arm.com, will.deacon@arm.com, marc.zyngier@arm.com,
        suzuki.poulose@arm.com, Dave.Martin@arm.com,
        shankerd@codeaurora.org, julien.thierry@arm.com,
        mlangsdo@redhat.com, stefan.wahren@i2e.com,
        linux-kernel@vger.kernel.org
References: <20190227010544.597579-1-jeremy.linton@arm.com>
 <20190227010544.597579-4-jeremy.linton@arm.com>
 <c98ff98e-30e2-596b-39c5-8505ddef9496@foss.arm.com>
From:   Jeremy Linton <jeremy.linton@arm.com>
Message-ID: <9cfb9cff-6a26-fff7-9374-82eea0f63a21@arm.com>
Date:   Fri, 1 Mar 2019 10:12:09 -0600
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.4.0
MIME-Version: 1.0
In-Reply-To: <c98ff98e-30e2-596b-39c5-8505ddef9496@foss.arm.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi,

On 3/1/19 1:11 AM, Andre Przywara wrote:
> Hi,
> 
> On 2/26/19 7:05 PM, Jeremy Linton wrote:
>> Display the mitigation status if active, otherwise
>> assume the cpu is safe unless it doesn't have CSV3
>> and isn't in our whitelist.
>>
>> Signed-off-by: Jeremy Linton <jeremy.linton@arm.com>
>> ---
>>   arch/arm64/kernel/cpufeature.c | 47 ++++++++++++++++++++++++++--------
>>   1 file changed, 37 insertions(+), 10 deletions(-)
>>
>> diff --git a/arch/arm64/kernel/cpufeature.c 
>> b/arch/arm64/kernel/cpufeature.c
>> index f6d84e2c92fe..d31bd770acba 100644
>> --- a/arch/arm64/kernel/cpufeature.c
>> +++ b/arch/arm64/kernel/cpufeature.c
>> @@ -944,7 +944,7 @@ has_useable_cnp(const struct 
>> arm64_cpu_capabilities *entry, int scope)
>>       return has_cpuid_feature(entry, scope);
>>   }
>> -#ifdef CONFIG_UNMAP_KERNEL_AT_EL0
>> +static bool __meltdown_safe = true;
>>   static int __kpti_forced; /* 0: not forced, >0: forced on, <0: 
>> forced off */
>>   static bool unmap_kernel_at_el0(const struct arm64_cpu_capabilities 
>> *entry,
>> @@ -963,6 +963,16 @@ static bool unmap_kernel_at_el0(const struct 
>> arm64_cpu_capabilities *entry,
>>           { /* sentinel */ }
>>       };
>>       char const *str = "command line option";
>> +    bool meltdown_safe;
>> +
>> +    meltdown_safe = is_midr_in_range_list(read_cpuid_id(), 
>> kpti_safe_list);
>> +
>> +    /* Defer to CPU feature registers */
>> +    if (has_cpuid_feature(entry, scope))
>> +        meltdown_safe = true;
>> +
>> +    if (!meltdown_safe)
>> +        __meltdown_safe = false;
>>       /*
>>        * For reasons that aren't entirely clear, enabling KPTI on Cavium
>> @@ -974,6 +984,11 @@ static bool unmap_kernel_at_el0(const struct 
>> arm64_cpu_capabilities *entry,
>>           __kpti_forced = -1;
>>       }
>> +    if (!IS_ENABLED(CONFIG_UNMAP_KERNEL_AT_EL0)) {
>> +        pr_info_once("kernel page table isolation disabled by 
>> CONFIG\n");
>> +        return false;
>> +    }
>> +
>>       /* Forced? */
>>       if (__kpti_forced) {
>>           pr_info_once("kernel page table isolation forced %s by %s\n",
>> @@ -985,14 +1000,10 @@ static bool unmap_kernel_at_el0(const struct 
>> arm64_cpu_capabilities *entry,
>>       if (IS_ENABLED(CONFIG_RANDOMIZE_BASE))
>>           return kaslr_offset() > 0;
>> -    /* Don't force KPTI for CPUs that are not vulnerable */
>> -    if (is_midr_in_range_list(read_cpuid_id(), kpti_safe_list))
>> -        return false;
>> -
>> -    /* Defer to CPU feature registers */
>> -    return !has_cpuid_feature(entry, scope);
>> +    return !meltdown_safe;
>>   }
>> +#ifdef CONFIG_UNMAP_KERNEL_AT_EL0
>>   static void
>>   kpti_install_ng_mappings(const struct arm64_cpu_capabilities *__unused)
>>   {
>> @@ -1022,6 +1033,13 @@ kpti_install_ng_mappings(const struct 
>> arm64_cpu_capabilities *__unused)
>>       return;
>>   }
>> +#else
>> +static void
>> +kpti_install_ng_mappings(const struct arm64_cpu_capabilities *__unused)
>> +{
>> +}
>> +#endif    /* CONFIG_UNMAP_KERNEL_AT_EL0 */
>> +
>>   static int __init parse_kpti(char *str)
>>   {
>> @@ -1035,7 +1053,6 @@ static int __init parse_kpti(char *str)
>>       return 0;
>>   }
>>   early_param("kpti", parse_kpti);
>> -#endif    /* CONFIG_UNMAP_KERNEL_AT_EL0 */
>>   #ifdef CONFIG_ARM64_HW_AFDBM
>>   static inline void __cpu_enable_hw_dbm(void)
>> @@ -1286,7 +1303,6 @@ static const struct arm64_cpu_capabilities 
>> arm64_features[] = {
>>           .field_pos = ID_AA64PFR0_EL0_SHIFT,
>>           .min_field_value = ID_AA64PFR0_EL0_32BIT_64BIT,
>>       },
>> -#ifdef CONFIG_UNMAP_KERNEL_AT_EL0
>>       {
>>           .desc = "Kernel page table isolation (KPTI)",
>>           .capability = ARM64_UNMAP_KERNEL_AT_EL0,
>> @@ -1302,7 +1318,6 @@ static const struct arm64_cpu_capabilities 
>> arm64_features[] = {
>>           .matches = unmap_kernel_at_el0,
>>           .cpu_enable = kpti_install_ng_mappings,
>>       },
>> -#endif
>>       {
>>           /* FP/SIMD is not implemented */
>>           .capability = ARM64_HAS_NO_FPSIMD,
>> @@ -2063,3 +2078,15 @@ static int __init enable_mrs_emulation(void)
>>   }
>>   core_initcall(enable_mrs_emulation);
>> +
>> +ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute 
>> *attr,
>> +        char *buf)
>> +{
>> +    if (arm64_kernel_unmapped_at_el0())
>> +        return sprintf(buf, "Mitigation: KPTI\n");
>> +
>> +    if (__meltdown_safe)
>> +        return sprintf(buf, "Not affected\n");
> 
> Shall those two checks be swapped? So it doesn't report about a KPTI 
> mitigation if the CPU is safe, but we enable KPTI because of KASLR 
> having enabled it? Or is that a different knob?

Hmmm, I think having it this way reflects the fact that the machine is 
mitigated independent of whether it needed it. The force on case is 
similar. The machine may not have needed the mitigation but it was 
forced on.


> 
> Cheers,
> Andre.
>  
>> +
>> +    return sprintf(buf, "Vulnerable\n");
>> +}
>>