Re: [Qemu-ppc] [Qemu-devel] [QEMU RFC PATCH v3 4/6] Migration: migrate QTAILQ

[Dr. David Alan Gilbert]

* Paolo Bonzini (address@hidden) wrote:
> 
> 
> On 07/06/2016 18:34, Dr. David Alan Gilbert wrote:
> > * Paolo Bonzini (address@hidden) wrote:
> >>
> >>
> >> On 07/06/2016 16:43, Dr. David Alan Gilbert wrote:
> >>>   b) I think you should really try and split it into two parts:
> >>>     1) A VMSTATE_ARRAY_CUSTOM (?) - so it's an array of elements but 
> >>> we've got a special
> >>>        way of allocating/counting/reading the elements
> >>>     2) A version of that which is for a QTAILQ.
> >>>        It's important that whatever ends up on the migration stream 
> >>> doesn't reflect
> >>>        that it happens to be implemented as a QTAILQ; so if you decide to 
> >>> change
> >>>        it later the migration compatibility doesn't break.
> >>
> >> (Just to clarify before Jianjun wakes up) a VMSTATE_ARRAY is just a
> >> sequence of elements.  The count, if needed as in a VARRAY, is stored
> >> earlier and separately.  Currently lists (including this QTAILQ) are
> >> usually represented in the migration stream as a sequence of elements
> >> preceded by "1" and terminated by a "0".  Would you like to change it to
> >> a count + sequence as well?
> >>
> >> This would prevent using the new QTAILQ support for other existing lists
> >> such as virtio-blk and scsi's request lists.
> > 
> > That depends if it's something you're trying to keep migration compatibility
> > with;  if you're trying to keep format compaibility then sure keep it as is.
> > If you're not trying to keep compatibility, then why can't virtio-blk or
> > scsi request lists also use a count rather than the markers?
> 
> We're trying to keep compatibility, and I think it's among the last bits
> that are resisting conversion to vmstate.  Which explains my interest in
> Jianjun's patches. :)

OK, fine - if you're trying to keep format compatibility then I agree.
Although I'm not entirely sure that things like virtio-blk, scsi and everything 
else
are consistent in their structure.

Dave

> 
> Paolo
> 
> >>>   c) Use new trace_ names for get_qtailq rather than misusing 
> >>> trace_vmstate_load_state*
> >>>   d) Add a trace_ for put_qtailq as well - that way when it goes horribly 
> >>> wrong we can
> >>>      turn the tracing on on both sides :-)
> >>>   e) Is there anyway to make QTAILQ_RAW_INSERT_TAIL any more readable?
> >>>      I don't think I understand why you can't use the standard QTAILQ 
> >>> macros.
> >>
> >> Because they assume a particular type. The "raw" version need to work on
> >> void*.
> > 
> > Ah OK.
> > 
> > Dave
> > 
> >>
> >> Thanks,
> >>
> >> Paolo
> >>
> >>>   f) You might need to fix up Amit's scripts/vmstate-static-checker.py
> > --
> > Dr. David Alan Gilbert / address@hidden / Manchester, UK
> > 
--
Dr. David Alan Gilbert / address@hidden / Manchester, UK