On Fri, Dec 14, 2018 at 10:29:04AM -0800, Matthew Wilcox wrote:
> On Fri, Dec 14, 2018 at 10:07:20AM -0800, Roman Gushchin wrote:
> > Vmalloc() is getting more and more used these days (kernel stacks,
> > bpf and percpu allocator are new top users), and the total %
> > of memory consumed by vmalloc() can be pretty significant
> > and changes dynamically.
> >
> > /proc/meminfo is the best place to display this information:
> > its top goal is to show top consumers of the memory.
> >
> > Since the VmallocUsed field in /proc/meminfo is not in use
> > for quite a long time (it has been defined to 0 by the
> > commit a5ad88ce8c7f ("mm: get rid of 'vmalloc_info' from
> > /proc/meminfo")), let's reuse it for showing the actual
> > physical memory consumption of vmalloc().
>
> Do you see significant contention on nr_vmalloc_pages? Also, if it's
> just an atomic_long_t, is it worth having an accessor for it? And if
> it is worth having an accessor for it, then it can be static.
Not really, so I decided that per-cpu counter is an overkill
right now; but we can easily switch over once we'll notice any contention.
Will add static.
>
> Also, I seem to be missing 3/4.
>
Hm, https://lkml.org/lkml/2018/12/14/1048 ?
Thanks!
