Re: [PATCH v7 2/4] arm64: Add BBM Level 2 cpu feature

[Ryan Roberts <ryan.roberts@arm.com>]

On 19/06/2025 09:57, Mikołaj Lenczewski wrote:
> On Wed, Jun 18, 2025 at 03:07:41PM +0100, Ryan Roberts wrote:
>> On 17/06/2025 10:51, Mikołaj Lenczewski wrote:
>>>  #ifdef CONFIG_ARM64_PAN
>>>  static void cpu_enable_pan(const struct arm64_cpu_capabilities *__unused)
>>>  {
>>> @@ -2980,6 +3015,11 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
>>>  		.matches = has_cpuid_feature,
>>>  		ARM64_CPUID_FIELDS(ID_AA64MMFR2_EL1, EVT, IMP)
>>>  	},
>>> +	{
>>> +		.capability = ARM64_HAS_BBML2_NOABORT,
>>> +		.type = ARM64_CPUCAP_EARLY_LOCAL_CPU_FEATURE,
>>> +		.matches = has_bbml2_noabort,
>>
>> Is there a reason you have removed the .desc from this? Without it, the kernel
>> won't print a "detected" line when it enables the feature. Previously you had:
>>
>> 		.desc = "BBM Level 2 without conflict abort",
>>
>>> +	},
>>>  	{
>>>  		.desc = "52-bit Virtual Addressing for KVM (LPA2)",
>>>  		.capability = ARM64_HAS_LPA2, 
>>
> 
> Damn it! I completely forgot to birng that back. Yet another remenant of
> the NO_BBML2_NOABORT version we had discussed in v6, because a
> description field would have been confusing due to the double negation:
> 
> https://lore.kernel.org/all/83d1f7af-3dc7-45f9-94f3-8a0917c051d2@arm.com/
> 
> Thank you very much for spotting this Ryan! Ill return the description,
> and respin on Monday (unless it is possible to do this as a RESEND or
> similar, but I doubt it since this is technically a functional change) :/
> 

I think probably just send out a new version.