Re: [PATCH 0/7] Factor out common build rules and helpers in tools/tests

[Edwin Torok <edwin.torok@xxxxxxxxxx>]

> On 27 Feb 2026, at 18:00, Edwin Torok <edwin.torok@xxxxxxxxxx> wrote:
> 
> 
>> On 23 Feb 2026, at 10:42, Jan Beulich <jbeulich@xxxxxxxx> wrote:
>> 
>> On 23.02.2026 11:14, Edwin Török wrote:
>>> There are a lot of duplicate rules and code in tools/tests.
>>> To simplify writing new tests move common build rules into a
>>> `tools/tests/Rules.mk`, and helper macros/functions into 
>>> `common/{tests,guests}.{c,h}`.
>>> 
>>> ….
>> 
>> Just fyi that there is also
>> https://lists.xen.org/archives/html/xen-devel/2026-02/msg00751.html.
>> Can the two of you maybe work together to have a single resulting approach?
>> 
>> Jan
> 
> There is definitely a lot in common between the two series, and I’m talking 
> with Denis.
> The Makefile cleanups that both of us have done can probably go in first as 
> part of a shared series,
> I’ll try to setup a combined branch to see how that’d look like.
> 


I have a combined branch here:
https://gitlab.com/xen-project/people/edwintorok/xen/-/compare/staging...private%2Fedvint%2Fmigration-tests5?from_project_id=2336572

TODO:
* rebase the vPCI Makefile patch on top of this 
* there is review feedback for the previous series on the list

Today is my last actual working day at Citrix, and likely next week onward this 
email address will stop working too.
I intended to finish my pending patch series, unfortunately I ran out of time 
(I managed to get the ~70+ XAPI+Xen branches down to ~18, but cycling through 
each series took more time than available).
I hope that someone else (e.g. from XenServer) will be able to pick up where I 
left off and complete these.

I intended to send out the rest of the branches I have as patches, but git 
send-email or the internal SMTP server stopped working.

Here are some git repositories instead:
* this series: 
https://gitlab.com/xen-project/people/edwintorok/xen/-/compare/staging...private%2Fedvint%2Fmigration-tests5?from_project_id=2336572

* the migration foreign copy optimizations as a Proof of Concept (~2.5x faster 
localhost migrations). Don’t use these as is, should be done properly by 
introducing a separate foreign copy operation, instead of piggy-backing on top 
of grant-copy as I’ve done here (piggy-backing was useful for measuring the 
performance optimisation potential, since a lot of the glue code between 
hypervisor and user space was already there): 
https://gitlab.com/xen-project/people/edwintorok/xen/-/compare/staging...private%2Fedvint%2Ftests2?from_project_id=2336572

* combined XenServer patch queue + PMU stacktrace (flame graph) + migration 
foreign copy optimizations 
https://gitlab.com/xen-project/people/edwintorok/xen/-/compare/staging...guilt%2Fpatches?from_project_id=2336572
* foreign copy Proof of Concept for the kernel (this is a kernel tree in a Xen 
git repo):  
https://gitlab.com/xen-project/people/edwintorok/xen/-/commit/ea13155dce8f2f1cfd1a2c54a3f8158be8beed62

I’m not leaving the Xen ecosystem entirely, in my new role I’ll be working on 
the Unikraft and MirageOS subprojects, instead of the XAPI subproject of Xen
(it’ll likely mean a lot less interaction with the hypervisor itself)

Thank you all for all your patience and advice when reviewing patches!


Best regards,
—Edwin