Re: [Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[Ian Jackson]

Juergen Gross writes ("Re: [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series 
"C")"):
> Aah, so on ARM we have no dom0 support?

No, that is a mistake.

Ian.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[Juergen Gross]

On 25/04/18 15:54, Ian Jackson wrote:
> Juergen Gross writes ("Re: [PATCH for-4.11 0/2] SUPPORT.md matrix fixes 
> (series "C")"):
>> On 25/04/18 15:43, George Dunlap wrote:
>>> 2. Backport all renames / reorganizations to all supported versions
>>
>> +1
>>
>> As this will only be more specific it is a win. Again above example:
>> How would you read the 4.10 PVH support? Is dom0 supported? Its a guest
>> after all...
> 
> I think "guest" excludes dom0.  dom0 is a domain but not a guest.

Aah, so on ARM we have no dom0 support?


Juergen


___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[Ian Jackson]

Juergen Gross writes ("Re: [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series 
"C")"):
> On 25/04/18 15:43, George Dunlap wrote:
> > 2. Backport all renames / reorganizations to all supported versions
> 
> +1
> 
> As this will only be more specific it is a win. Again above example:
> How would you read the 4.10 PVH support? Is dom0 supported? Its a guest
> after all...

I think "guest" excludes dom0.  dom0 is a domain but not a guest.

Ian.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[Ian Jackson]

Ian Jackson writes ("Re: [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series 
"C")"):
> George Dunlap writes ("Re: [PATCH for-4.11 0/2] SUPPORT.md matrix fixes 
> (series "C")"):
> > 2. Backport all renames / reorganizations to all supported versions
...
> > I was initially opposed to #2, but I think the idea is growing on
> > me.  It does mean SUPPORT.md may end up being reorganized or renamed
> > in point releases, however.  It’s a bit hard for me to tell how
> > disruptive that would be.
...
> If someone wants to send a patch (on top of my backport series,
> please) that renames "PVH guest" to "PVH" in 4.10, and changes
> "Status:" to "Status, domU:", to match 4.11 that would be fine by me.

Oh, and, if we contemplate doing #2 occasionally then we should ask
people to not send mixed patches which do both
(i) renaming/reorganising features (ii) changing the support status.

Ian.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[Ian Jackson]

George Dunlap writes ("Re: [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series 
"C")"):
> Right, so there are four options:
> 
> 1. Never rename / reorganize SUPPORT.md categories

This is clearly unworkable.

> 3. Introduce some sort of “mapping” of options so that the table
> generator can correctly construct rows

I think this would be quite annoying.  It would have to be maintained
separately, probably in the branch for the next version.  It would
also make the table generator more complicated and it's quite bad
enough already.

> 2. Backport all renames / reorganizations to all supported versions
> 4. Tolerate duplicate rows for renamed / reorganized features

So it has to be one of these.  IMO either of these is fine.

> I was initially opposed to #2, but I think the idea is growing on
> me.  It does mean SUPPORT.md may end up being reorganized or renamed
> in point releases, however.  It’s a bit hard for me to tell how
> disruptive that would be.

We don't have to make this decision the same way for every feature.

If someone wants to send a patch (on top of my backport series,
please) that renames "PVH guest" to "PVH" in 4.10, and changes
"Status:" to "Status, domU:", to match 4.11 that would be fine by me.

Ian.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[Juergen Gross]

On 25/04/18 15:43, George Dunlap wrote:
> 
> 
>> On Apr 25, 2018, at 2:32 PM, Juergen Gross  wrote:
>>
>> On 25/04/18 15:21, Ian Jackson wrote:
>>> Juergen Gross writes ("Re: [PATCH for-4.11 0/2] SUPPORT.md matrix fixes 
>>> (series "C")"):
 Not related to these patches, but:

 SUPPORT.md of 4.10 seems to have some entries different to 4.11. Do we
 want to change those? This might result in a more readable table.

 e.g.:

 4.10: ### x86/PVH guest
  Status: Supported

 4.11: ### x86/PVH
  Status, domU: Supported
  Status, dom0: Experimental
>>>
>>> Indeed.  I noticed this when I was backporting my reformatting.
>>> I considered changing this but I think TBH that this slight deviation
>>> in naming is going to occur occasionally.
>>
>> The resulting table is rather hard to read, don't you think?
>>
>> Especially the supported guest types are difficult to compare between
>> 4.10 and 4.11.
> 
> Right, so there are four options:
> 
> 1. Never rename / reorganize SUPPORT.md categories

As we can see in the example above this won't work very well.

> 2. Backport all renames / reorganizations to all supported versions

+1

As this will only be more specific it is a win. Again above example:
How would you read the 4.10 PVH support? Is dom0 supported? Its a guest
after all...

> 3. Introduce some sort of “mapping” of options so that the table generator 
> can correctly construct rows

Seems to be rather complex, e.g. in above example

> 4. Tolerate duplicate rows for renamed / reorganized features

This might grow rather ugly results after some more versions.


Juergen


___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[George Dunlap]

> On Apr 25, 2018, at 2:32 PM, Juergen Gross  wrote:
> 
> On 25/04/18 15:21, Ian Jackson wrote:
>> Juergen Gross writes ("Re: [PATCH for-4.11 0/2] SUPPORT.md matrix fixes 
>> (series "C")"):
>>> Not related to these patches, but:
>>> 
>>> SUPPORT.md of 4.10 seems to have some entries different to 4.11. Do we
>>> want to change those? This might result in a more readable table.
>>> 
>>> e.g.:
>>> 
>>> 4.10: ### x86/PVH guest
>>>  Status: Supported
>>> 
>>> 4.11: ### x86/PVH
>>>  Status, domU: Supported
>>>  Status, dom0: Experimental
>> 
>> Indeed.  I noticed this when I was backporting my reformatting.
>> I considered changing this but I think TBH that this slight deviation
>> in naming is going to occur occasionally.
> 
> The resulting table is rather hard to read, don't you think?
> 
> Especially the supported guest types are difficult to compare between
> 4.10 and 4.11.

Right, so there are four options:

1. Never rename / reorganize SUPPORT.md categories
2. Backport all renames / reorganizations to all supported versions
3. Introduce some sort of “mapping” of options so that the table generator can 
correctly construct rows
4. Tolerate duplicate rows for renamed / reorganized features

I was initially opposed to #2, but I think the idea is growing on me.  It does 
mean SUPPORT.md may end up being reorganized or renamed in point releases, 
however.  It’s a bit hard for me to tell how disruptive that would be.

 -George
___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[Juergen Gross]

On 25/04/18 15:21, Ian Jackson wrote:
> Juergen Gross writes ("Re: [PATCH for-4.11 0/2] SUPPORT.md matrix fixes 
> (series "C")"):
>> Not related to these patches, but:
>>
>> SUPPORT.md of 4.10 seems to have some entries different to 4.11. Do we
>> want to change those? This might result in a more readable table.
>>
>> e.g.:
>>
>> 4.10: ### x86/PVH guest
>>   Status: Supported
>>
>> 4.11: ### x86/PVH
>>   Status, domU: Supported
>>   Status, dom0: Experimental
> 
> Indeed.  I noticed this when I was backporting my reformatting.
> I considered changing this but I think TBH that this slight deviation
> in naming is going to occur occasionally.

The resulting table is rather hard to read, don't you think?

Especially the supported guest types are difficult to compare between
4.10 and 4.11.


Juergen

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[Ian Jackson]

Juergen Gross writes ("Re: [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series 
"C")"):
> Not related to these patches, but:
> 
> SUPPORT.md of 4.10 seems to have some entries different to 4.11. Do we
> want to change those? This might result in a more readable table.
> 
> e.g.:
> 
> 4.10: ### x86/PVH guest
>   Status: Supported
> 
> 4.11: ### x86/PVH
>   Status, domU: Supported
>   Status, dom0: Experimental

Indeed.  I noticed this when I was backporting my reformatting.
I considered changing this but I think TBH that this slight deviation
in naming is going to occur occasionally.

There's a paragraph about it in the intro and everything.

> 4.10: ### x86/Emulated graphics (QEMU):
>   Status, stgvga: Supported <- note the typo in "stgvga"
> 4.11: ### x86/Emulated graphics (QEMU):
>   Status, stdvga: Supported

This should be fixed.  I'll send a followup patch.

Ian.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[Juergen Gross]

On 25/04/18 15:04, Ian Jackson wrote:
> I was just testing the results of my backports of the SUPPORT.md
> series to 4.10 - just sent, under the Subject line:
> [PATCH for-4.10 0/9] SUPPORT.md backports to support matrix generation
> 
> I found a bug in the support matrix generator.  Sadly the bug was not
> evident without the backported changes to SUPPORT.md.  These two
> patches fix this bug.
> 
> Example output can be found here:
>   https://xenbits.xen.org/people/iwj/2018/support-matrix-example-C/t.html
> (The hyperlink references for staging are to the live xenbits version;
> the references for 4.10 are to the example SUPPORT.html output file
> from the 4.10 backport series, as discussed above.)
> 
> For my reference, this was made as follows:
>   docs/support-matrix-generate -D HEAD 
> https://xenbits.xen.org/docs/unstable-staging/SUPPORT.html 
> refs/heads/wip.support-stmt-NN-2 SUPPORT.html 2>&1 >docs/html/t.html |less
>   rsync -LvP docs/html/{t,SUPPORT}.html 
> xenbits:public_html/2018/support-matrix-example-C
> 
> Thanks,
> Ian.
> 

For the series:

Release-acked-by: Juergen Gross 


Not related to these patches, but:

SUPPORT.md of 4.10 seems to have some entries different to 4.11. Do we
want to change those? This might result in a more readable table.

e.g.:

4.10: ### x86/PVH guest
  Status: Supported

4.11: ### x86/PVH
  Status, domU: Supported
  Status, dom0: Experimental

4.10: ### x86/Emulated graphics (QEMU):
  Status, stgvga: Supported <- note the typo in "stgvga"

4.11: ### x86/Emulated graphics (QEMU):
  Status, stdvga: Supported


Juergen

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

[Xen-devel] [PATCH for-4.11 0/2] SUPPORT.md matrix fixes (series "C")

[Ian Jackson]

I was just testing the results of my backports of the SUPPORT.md
series to 4.10 - just sent, under the Subject line:
[PATCH for-4.10 0/9] SUPPORT.md backports to support matrix generation

I found a bug in the support matrix generator.  Sadly the bug was not
evident without the backported changes to SUPPORT.md.  These two
patches fix this bug.

Example output can be found here:
  https://xenbits.xen.org/people/iwj/2018/support-matrix-example-C/t.html
(The hyperlink references for staging are to the live xenbits version;
the references for 4.10 are to the example SUPPORT.html output file
from the 4.10 backport series, as discussed above.)

For my reference, this was made as follows:
  docs/support-matrix-generate -D HEAD 
https://xenbits.xen.org/docs/unstable-staging/SUPPORT.html 
refs/heads/wip.support-stmt-NN-2 SUPPORT.html 2>&1 >docs/html/t.html |less
  rsync -LvP docs/html/{t,SUPPORT}.html 
xenbits:public_html/2018/support-matrix-example-C

Thanks,
Ian.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel