Re: [Xen-devel] Ping: [PATCH v2 RESEND] x86/time: correctly honor late clearing of TSC related feature flags

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

On 15/12/16 09:49, Jan Beulich wrote:
>>>> On 06.12.16 at 11:51, <JBeulich@xxxxxxxx> wrote:
>> As such clearing of flags may have an impact on the selected rendezvous
>> function, handle such in a central place.
>>
>> But don't allow such feature flags to be cleared during CPU hotplug:
>> Platform and local system times may have diverged significantly by
>> then, potentially causing noticably (even if only temporary) strange
>> behavior. As we're anyway expecting only sufficiently similar CPUs to
>> appear during hotplug, this shouldn't be introducing new limitations.
>>
>> Reported-by: Joao Martins <joao.m.martins@xxxxxxxxxx>
>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
>> Tested-by: Dario Faggioli <dario.faggioli@xxxxxxxxxx>
>> Tested-by: Joao Martins <joao.m.martins@xxxxxxxxxx>
>> ---
>> The resend is mainly to get the discussion going again on what the
>> alternatives are, if this patch is not acceptable.
> Even if you don't agree with the patch, can we at least revive
> the discussion of what alternatives there are?

Sorry - it slipped through the cracks.  I have no issue with the
principle of the patch.

The only problem I have, which we didn't sort out last time, is the
initialisation of rendezvous_fn

It is still my opinion that under no circumstance is it ok for
clear_tsc_cap() to modify time_calibration_rendezvous_fn from
time_calibration_tsc_rendezvous to time_calibration_std_rendezvous,
which can in principle happen because rendezvous_fn doesn't get
initialised from the current time_calibration_rendezvous_fn.

~Andrew

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel