Re: [Qemu-devel] [PATCH 4/8] [PATCH RFC v3] s390-qemu: cpu hotplug - Storage key global access

["Jason J. Herne" <jjherne@linux.vnet.ibm.com>]

On 09/05/2013 07:46 AM, Andreas Färber wrote:
> Am 01.08.2013 16:12, schrieb Jason J. Herne:
>> From: "Jason J. Herne" <jjherne@us.ibm.com>
>>
>> diff --git a/hw/s390x/s390-virtio.c b/hw/s390x/s390-virtio.c
>> index 439d732..21e9124 100644
>> --- a/hw/s390x/s390-virtio.c
>> +++ b/hw/s390x/s390-virtio.c
>> @@ -123,6 +123,18 @@ static void s390_virtio_register_hcalls(void)
>>                                      s390_virtio_hcall_set_status);
>>   }
>>
>> +static uint8_t *storage_keys;
>> +
>> +uint8_t *s390_get_storage_keys(void)
>> +{
>> +    return storage_keys;
>> +}
>> +
>> +void s390_alloc_storage_keys(ram_addr_t ram_size)
>> +{
>> +    storage_keys = g_malloc0(ram_size / TARGET_PAGE_SIZE);
>> +}
>> +
>>   /*
>>    * The number of running CPUs. On s390 a shutdown is the state of all CPUs
>>    * being either stopped or disabled (for interrupts) waiting. We have to
>> @@ -176,7 +188,7 @@ void s390_init_ipl_dev(const char *kernel_filename,
>>       qdev_init_nofail(dev);
>>   }
>>
>> -void s390_init_cpus(const char *cpu_model, uint8_t *storage_keys)
>> +void s390_init_cpus(const char *cpu_model)
>>   {
>>       int i;
>>
>> @@ -196,7 +208,7 @@ void s390_init_cpus(const char *cpu_model, uint8_t *storage_keys)
>>           ipi_states[i] = cpu;
>>           cs->halted = 1;
>>           cpu->env.exception_index = EXCP_HLT;
>> -        cpu->env.storage_keys = storage_keys;
>> +        cpu->env.storage_keys = s390_get_storage_keys();
>
> Why not go this from the CPU initfn? Is there any ccw- vs. non-ccw
> difference? Thinking about -device s390-cpu here. I believe it's safe to
> assume that machine init and thus allocation has run before the CPU is
> instantiated - possibly assert that.
>
>

Patch 6/8 does exactly that.

-- 
-- Jason J. Herne (jjherne@linux.vnet.ibm.com)