On 6/16/26 15:49, Gregory Price wrote: > On Tue, Jun 16, 2026 at 02:30:34PM +0200, David Hildenbrand (Arm) wrote: >> >> Nothing too crazy here, however, the question is which of these values we >> actually want to guarantee that we will provide them with unchanged >> semantics in >> the future ... I guess anything we already expose to user space is alright >> (because it effectively already must remain mostly unchanged I assume). >> > > I suppose a risk of them going away entirely? I suppose that's > fixable but unlikely.
Yeah, the first bunch is all exported through /proc/vmstat AFAIKS. The other through /proc/pressure/ So I assume this is just fine. -- Cheers, David
