Re: [PATCH] docs: fix documentation to notice credit2 is the default

[Jan Beulich <jbeulich@xxxxxxxx>]

On 10.11.2020 14:13, Jürgen Groß wrote:
> On 10.11.20 14:07, Andrew Cooper wrote:
>> On 10/11/2020 12:49, Roger Pau Monné wrote:
>>> On Tue, Nov 10, 2020 at 12:31:14PM +0100, Jürgen Groß wrote:
>>>> On 10.11.20 12:21, Roger Pau Monne wrote:
>>>>> Fix the command line document to account for credit2 now being the
>>>>> default scheduler.
>>>>>
>>>>> Fixes: dafd936dddbd ('Make credit2 the default scheduler')
>>>>> Signed-off-by: Roger Pau Monné <roger.pau@xxxxxxxxxx>
>>>>> ---
>>>>>    docs/misc/xen-command-line.pandoc | 2 +-
>>>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/docs/misc/xen-command-line.pandoc 
>>>>> b/docs/misc/xen-command-line.pandoc
>>>>> index 4ae9391fcd..789aead148 100644
>>>>> --- a/docs/misc/xen-command-line.pandoc
>>>>> +++ b/docs/misc/xen-command-line.pandoc
>>>>> @@ -1876,7 +1876,7 @@ with read and write permissions.
>>>>>    ### sched
>>>>>    > `= credit | credit2 | arinc653 | rtds | null`
>>>>> -> Default: `sched=credit`
>>>>> +> Default: `sched=credit2`
>>>>>    Choose the default scheduler.
>>>>>
>>>> Tried that before:
>>>>
>>>> https://lists.xen.org/archives/html/xen-devel/2019-01/msg01097.html
>>>>
>>>> And Andrew didn't like it...
>>> One way or another we need to get this fixed in the document. Listing
>>> credit as the still the default is wrong.
>>
>> I agree that what is there is wrong, but so is saying credit2.
>>
>> This documentation is for users, because develops know exactly how they
>> configured their schedulers, and won't actually need to refer to it.
>>
>> As a consequence, it depends heavily on what a specific
>> distro/downstream chose, config-wise.
>>
>>> I think there are several places in xen-command-line.pandoc that just
>>> contain the default values set in Kconfig, so IMO if we want to
>>> change this it should be done wholesale rather than picking on every
>>> default value change. Opinions?
>>
>> xen-command-line.pandoc wholly predates Kconfig.  We're only in this
>> mess because previous patches haven't been appropriately reviewed.
>>
>> Lets fix it up to be correct, but lets not delay fixing this to look for
>> potential other cases.
> 
> The ultimate fix would be to generate this document according to
> Kconfig settings. :-D

Except that's not suitable for putting up as a web page for
everyone to use as "cannoical" reference. Every distro could
do so, sure.

Jan