Re: [PATCH 1/3] KVM: X86: Add "nopvspin" parameter to disable PV spinlocks

From: Zhenzhong Duan
Date: Thu Oct 03 2019 - 05:53:13 EST



On 2019/10/3 0:47, Sean Christopherson wrote:
On Tue, Oct 01, 2019 at 05:47:00PM +0800, Zhenzhong Duan wrote:
On 2019/10/1 16:39, Vitaly Kuznetsov wrote:
Zhenzhong Duan<zhenzhong.duan@xxxxxxxxxx> writes:

On 2019/9/30 23:41, Vitaly Kuznetsov wrote:
Zhenzhong Duan<zhenzhong.duan@xxxxxxxxxx> writes:

There are cases where a guest tries to switch spinlocks to bare metal
behavior (e.g. by setting "xen_nopvspin" on XEN platform and
"hv_nopvspin" on HYPER_V).

That feature is missed on KVM, add a new parameter "nopvspin" to disable
PV spinlocks for KVM guest.

This new parameter is also intended to replace "xen_nopvspin" and
"hv_nopvspin" in the future.
Any reason to not do it right now? We will probably need to have compat
code to support xen_nopvspin/hv_nopvspin too but emit a 'is deprecated'
warning.
Sorry the description isn't clear, I'll fix it.

I did the compat work in the other two patches.
[PATCH 2/3] xen: Mark "xen_nopvspin" parameter obsolete and map it to "nopvspin"
[PATCH 3/3] x86/hyperv: Mark "hv_nopvspin" parameter obsolete and map it to "nopvspin"

For some reason I got CCed only on the first one and moreover,
The three patches have different maintainers/reviewers by get_maintainer.pl, I added
"Cc: maintainers/reviewers" to each patch then git-sendemail picked them automaticly.
I meaned to not disturb maintainers with the field they aren't in charge of. It looks
I'm wrong.

So what's the correct way dealing with this? Should I send the whole patchset to all
the maintainers/reviewers related to all the patches?
There's no one right answer to that question, folks have different
preferences. My general rule of thumb is to cc everyone on all patches
unless the series is obnoxiously large *and* isolated to a specific part
of the kernel. The idea being that people are more likely to be annoyed
if they can't find all patches in a relatively small series (this case)
than they are about receiving a mail or two that they don't care about.

At a minimum I would cc everyone involved on the cover letter, and cc the
relevant mailing lists on all patches. Sending everyone the cover letter
provides people a quick overview of the patches they didn't receive, as
well as a starting point if they want to find those patches. Cc'ing the
mailing list(s) can make it even easier to find the patches. The cover
letter is also a good place to explain why you didn't cc everyone on all
patches (or vice versa).

Also, the cover letter should have the shortlog and overall diffstats.
'git format-patch --cover-letter' will do the work for you.

Thanks for your detailed reply, I's clear to me what to do now.

Zhenzhong