RE: [PATCH v10 16/24] x86/resctrl: Add interface to the assign counter

[Luck, Tony]

> >>>>>> It is right thing to continue assignment if one of the domain is out of
> >>>>>> counters. In that case how about we save the error(say error_domain) and
> >>>>>> continue. And finally return success if both ret and error_domain are zeros.
> >>>>>>
> >>>>>>       return ret ?  ret : error_domain:
> >>>>>
> >>>>> If there are many domains, then you might have 3 succeed and 5 fail.
> >>>>>
> >>>>> I think the best you can do is return success if everything succeeded
> >>>>> and an error if any failed.
> >>>>
> >>>> Yes. The above check should take care of this case.
> >>>>
> >>>
> >>> If I understand correctly "error_domain" can capture the ID of
> >>> a single failing domain. If there are multiple failing domains like
> >>> in Tony's example then "error_domain" will not be accurate and thus
> >>> can never be trusted. Instead of a single check of a failure user
> >>> space is then forced to parse the more complex "mbm_assign_control"
> >>> file to learn what succeeded and failed.
> >>>
> >>> Would it not be simpler to process sequentially and then fail on
> >>> first error encountered with detailed error message? With that
> >>> user space can determine exactly which portion of request
> >>> succeeded and which portion failed.
> >>
> >> One more option is to print the error for each failure and continue. And finally return error.

There's limited space allocated for use by last_cmd_*() messages:

static char last_cmd_status_buf[512];

        seq_buf_init(&last_cmd_status, last_cmd_status_buf,
                     sizeof(last_cmd_status_buf));

If you keep parsing and trying to apply changes from user input you will
quickly hit that limit.


> >>
> >> "Group mon1, domain:1 Out of MBM counters"
> >>
> >> We have the error information as well as the convenience of assignment on domains where counters are available when user is working with "*"(all domains).
> >
> > This may be possible. Please keep in mind that any errors have to be
> > easily consumed in an automated way to support the user space tools
> > that interact with resctrl. I do not think we have thus far focused
> > on the "last_cmd_status" buffer as part of the user space ABI so this opens
> > up more considerations.
> >
> > At this time the error handling of "all domains" does not seem to be
> > consistent and obvious to user space. From what I can tell the
> > implementation continues on to the next domain if one domain is out
> > of counters but it exits immediately if a counter cannot be configured
> > on a particular domain.
>
> Yes. We can handle both the errors in the same way.

I think it is simplest to make the "same way" be "fail on first error".

>
> >
> >>
> >> Note: I will be out of office starting next week Until Jan 10.
> >
> > Thank you for letting me know. I am currently reviewing this series
> > and will post feedback by tomorrow.
>
> Sure. Thanks. I will try to get to some of it at least. The review
> comments which needs investigation may have to wait. Lets see.

-Tony