Hi,
On 10/26/20 4:39 PM, Limonciello, Mario wrote:
>> This was present in previous versions too, but I just noticed this are you
>> sure that using
>> .string.pointer is correct here? That seems wrong since the pointer gets
>> allocated by
>> the Linux ACPI core, so it is not under influence of the AML code?
>>
>> I think you want / need to use ".integer.value" here ?
>>
>> And maybe first do a type check, e.g.:
>>
>> if (obj->package.elements[CURRENT_VAL].type != ACPI_TYPE_INTEGER) {
>> ret = -EINVAL;
>> goto out;
>> }
>>
>> Adding this type check will also show if I'm right that you should use
>> .integer.value ...
>>
>> ret = snprintf(buf, PAGE_SIZE, "%lld\n", obj-
>
> We'll need to double check this, but I'm pretty sure the firmware outputs
> everything as a string.
Ok, in that case it should probably be printed as a "%s" though and not interpret
the buffer-address which the kernel allocated for storing the string as an
integer.
And it would still be good to do the type-check for this, but then checking
for a type of ACPI_TYPE_STRING.
Regards,
Hans
