Re: [Qemu-devel] [PATCH 4/6] target/arm: Add "-cpu max" support

[Peter Maydell]

On 26 January 2018 at 14:29, Philippe Mathieu-Daudé <address@hidden> wrote:
> Hi Peter,
>
> On 12/07/2017 03:14 PM, Peter Maydell wrote:
>> Add support for "-cpu max" for ARM guests. This CPU type behaves
>> like "-cpu host" when KVM is enabled, and like a system CPU with
>> the maximum possible feature set otherwise. (Note that this means
>> it won't be migratable across versions, as we will likely add
>> features to it in future.)
>>
>> Signed-off-by: Peter Maydell <address@hidden>
>> ---

>> +#ifndef TARGET_AARCH64
>> +/* -cpu max: if KVM is enabled, like -cpu host (best possible with this 
>> host);
>> + * otherwise, a CPU with as many features enabled as our emulation supports.
>> + * The version of '-cpu max' for qemu-system-aarch64 is defined in cpu64.c;
>> + * this only needs to handle 32 bits.
>> + */
>> +static void arm_max_initfn(Object *obj)
>> +{
>> +    ARMCPU *cpu = ARM_CPU(obj);
>> +
>> +    if (kvm_enabled()) {
>> +        kvm_arm_set_cpu_features_from_host(cpu);
>> +    } else {
>> +        cortex_a15_initfn(obj);> +        /* In future we might add feature 
>> bits here even if the
>> +         * real-world A15 doesn't implement them.
>> +         */
>
> Why not use arm_any_initfn() here?

That function (and the 'any' cpu) are deliberately only
included in the linux-user binaries, not the system-emulation binaries.
(Also arm_any_initfn() only initializes userspace-visible stuff, it
doesn't provide ID register values etc for kernel-visible things.)

> Actually what seems cleaner is to move "any" features here, and kill the
> "any" cpu, using "max" for this purpose.

We can't kill 'any', that would break back-compatibility
of command lines.

thanks
-- PMM