[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [RFC 1/3] acpi: cpuhp: fix 'Command data' description is spec
From: |
Laszlo Ersek |
Subject: |
Re: [RFC 1/3] acpi: cpuhp: fix 'Command data' description is spec |
Date: |
Tue, 22 Oct 2019 19:37:05 +0200 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1 |
(I've been dropped from the address list, not sure why)
On 10/22/19 19:17, Christophe de Dinechin wrote:
>
> Laszlo Ersek writes:
>
>> On 10/10/19 15:31, Laszlo Ersek wrote:
>>> 2nd round:
>>>
>>> On 10/09/19 15:22, Igor Mammedov wrote:
>>>> QEMU returns 0, in case of erro or invalid value in 'Command field',
>>>> make spec match reality, i.e.
>>>
>>> AHA! so this is exactly where you meant to list the particular cases
>>> when "command data" reads as 0:
>>> - CPU >= max_cpus selected,
>>> - CPU with pending events asked for, but none found
>>>
>>>> Also fix returned value description in case 'Command field' == 0x0,
>>>> it's in not PXM but CPU selector value with pending event
>>>>
>>>> Signed-off-by: Igor Mammedov <address@hidden>
>>>> ---
>>>> docs/specs/acpi_cpu_hotplug.txt | 5 +++--
>>>> 1 file changed, 3 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/docs/specs/acpi_cpu_hotplug.txt
>>>> b/docs/specs/acpi_cpu_hotplug.txt
>>>> index ee219c8358..ac5903b2b1 100644
>>>> --- a/docs/specs/acpi_cpu_hotplug.txt
>>>> +++ b/docs/specs/acpi_cpu_hotplug.txt
>>>> @@ -44,9 +44,10 @@ read access:
>>>> 3-7: reserved and should be ignored by OSPM
>>>> [0x5-0x7] reserved
>>>> [0x8] Command data: (DWORD access)
>>>> - in case of error or unsupported command reads is 0xFFFFFFFF
>>>> + in case of error or unsupported command reads is 0x0
>>>> current 'Command field' value:
>>>> - 0: returns PXM value corresponding to device
>>>> + 0: returns CPU selector value corresponding to a CPU with
>>>> + pending event.
>>>>
>>>> write access:
>>>> offset:
>>>>
>>>
>>> How about:
>>>
>>> [0x8] Command data: (DWORD access, little endian)
>>> If the "CPU selector" value last stored by the guest refers to
>>> an impossible CPU, then 0.
>>> Otherwise, if the "Command field" value last stored by the
>>> guest differs from 0, then 0.
>>> Otherwise, if there is at least one CPU with a pending event,
>>> then that CPU has been selected; the command data register
>>> returns that selector.
>>> Otherwise, 0.
>>
>> Hmmm not exactly. Let me try again.
>>
>> [0x8] Command data: (DWORD access, little endian)
>> If the "CPU selector" value last stored by the guest refers to
>> an impossible CPU, then 0.
>> Otherwise, if the "Command field" value last stored by the
>> guest differs from 0, then 0.
>> Otherwise, the currently selected CPU.
>
> How about phrasing it to put the more general case first, e.g.
>
> [0x8] Command data: (DWORD access, little endian)
>
> The currently selected CPU, unless:
> - The "CPU selector" value refers to an impossible CPU
> - The "Command field" value last stored by the guest differs
> from 0
> In these last two cases, the value is 0.
I prefer my proposal because it translates to source code more directly
(a ladder of "if / else" pairs, or similar).
(I know that some programming languages support "unless"; I could never
wrap my brain around that idea! :) )
Thanks
Laszlo
[RFC 2/3] acpi: cpuhp: add typical usecases into spec, Igor Mammedov, 2019/10/09
[RFC 3/3] acpi: cpuhp: add CPHP_GET_CPU_ID_CMD command, Igor Mammedov, 2019/10/09