[PATCH 3/3] arm64: Introduce command line parameter to disable CNP

[Vladimir Murzin]

Hi Julien,

On 10/10/17 15:36, Julien Thierry wrote:
> Hi Vladimir,
> 
> On 09/10/17 13:55, Vladimir Murzin wrote:
>> There are cases when activating of Common Not Private (CNP) feature
>> might not be desirable; this patch allows to forcefully disable CNP
>> even it is supported by hardware.
>>
>> Signed-off-by: Vladimir Murzin <vladimir.murzin at arm.com>
>> ---
>>   Documentation/admin-guide/kernel-parameters.txt |  4 ++++
>>   arch/arm64/kernel/cpufeature.c                  | 20 +++++++++++++++++++-
>>   2 files changed, 23 insertions(+), 1 deletion(-)
>>
>> diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
>> index 0549662..3c1e45d 100644
>> --- a/Documentation/admin-guide/kernel-parameters.txt
>> +++ b/Documentation/admin-guide/kernel-parameters.txt
>> @@ -2560,6 +2560,10 @@
>>         noclflush    [BUGS=X86] Don't use the CLFLUSH instruction
>>   +    nocnp        [ARM64]
>> +            Disable CNP (Common not Private translations)
>> +            even if it is supported by processor.
>> +
>>       nodelayacct    [KNL] Disable per-task delay accounting
>>         nodsp        [SH] Disable hardware DSP at boot time.
>> diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
>> index 8d098a1..724fd93 100644
>> --- a/arch/arm64/kernel/cpufeature.c
>> +++ b/arch/arm64/kernel/cpufeature.c
>> @@ -771,6 +771,24 @@ static bool has_no_hw_prefetch(const struct arm64_cpu_capabilities *entry, int _
>>           MIDR_CPU_VAR_REV(1, MIDR_REVISION_MASK));
>>   }
>>   +static bool nocnp;
>> +
>> +static int __init early_nocnp(char *p)
>> +{
>> +    nocnp = true;
>> +    return 0;
>> +}
>> +early_param("nocnp", early_nocnp);
>> +
>> +static bool has_useable_cnp(const struct arm64_cpu_capabilities *entry, int scope)
>> +{
>> +    if (!has_cpuid_feature(entry, scope))
>> +        return false;
>> +
>> +    return nocnp ? false : true;
> 
> This feels a bit odd.
> 
> Wouldn't the following be better?
> return !nocnp;
> 
> Or simply the whole function as:
> return has_cpuid_feature(entry, scope) && !nocnp;
> 

I have no strong opinion on that, so I'll change per your suggestion.

Cheers
Vladimir

> Thanks,
> 
>> +}
>> +
>> +
>>   static bool runs_at_el2(const struct arm64_cpu_capabilities *entry, int __unused)
>>   {
>>       return is_kernel_in_hyp_mode();
>> @@ -905,7 +923,7 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
>>           .desc = "Common not Private translations",
>>           .capability = ARM64_HAS_CNP,
>>           .def_scope = SCOPE_SYSTEM,
>> -        .matches = has_cpuid_feature,
>> +        .matches = has_useable_cnp,
>>           .sys_reg = SYS_ID_AA64MMFR2_EL1,
>>           .sign = FTR_UNSIGNED,
>>           .field_pos = ID_AA64MMFR2_CNP_SHIFT,
>>
>