Re: [PATCH v2 2/2] xen/balloon: Enforce various limits on target

[Stefano Stabellini]

On Tue, 30 Apr 2013, Daniel Kiper wrote:
> > > > > +/*
> > > > > + * Extra internal memory reserved by libxl.
> > > > > + * Check tools/libxl/libxl_memory.txt file in Xen source for more details.
> > > > > + */
> > > > > +#define LIBXL_MAXMEM_CONSTANT_PAGES	(1024 * 1024 / PAGE_SIZE)
> > > >
> > > > I think we need to find a way to achieve your aims which doesn't require
> > > > leaking internal implementation details of libxl into the guest kernels.
> > > > What happens if libxl decides to double this?
> > >
> > > I agree that this is not elegant solution. However, if we would like to
> > > be in line with docs/misc/libxl_memory.txt (this is correct path) this
> > > is a must.
> >
> > I'm not sure about this, that file describes the toolstacks view of the
> > memory in a system. That need not necessarily correspond with the
> > guest's ideas (although you would hope it would be a superset).
> >
> > Surely it is logically wrong to bake toolstack specific knowledge in the
> > guest? If someone can describe a meaningful semantic for this number in
> > a toolstack independent way then perhaps it would be appropriate to do
> > something with it. I've no idea, having looked at both the document and
> > the code, what this value actually is.
> 
> This was added by commit 9905ac2b90a3e7cecd9e7dfe21c252362e7080b2
> (libxenlight: implement libxl_set_memory_target). It was written
> by Keir and signed off by Stefano (both are CCed here). Guys,
> why did you added LIBXL_MAXMEM_CONSTANT? What does it mean?

libxl inherits the memory model from xapi.
LIBXL_MAXMEM_CONSTANT corresponds to "extra internal" in xapi, an amount
of memory that is not allocated to the domain but it is left as a slack
on top of the actual memory target to determine the Xen maxmem.
I believe it comes from empirical measurements and stress testing on the
platform.
The xapi guys, CC'ed, might have more insights on what exactly is.


I dislike having to pull this "hack" into Linux, but if it is actually
important to leave LIBXL_MAXMEM_CONSTANT unused, then it is worth doing.
I would add a big comment on top saying:

"libxl seems to think that we need to leave LIBXL_MAXMEM_CONSTANT
kilobytes unused, let's be gentle and do that."
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/