On 10/15/19 1:09 AM, Michal Hocko wrote:
On Wed 09-10-19 17:45:09, Kirill A. Shutemov wrote:
THPs on deferred split queue got split by shrinker if memory pressure
comes.
In absence of memory pressure, there is no bound on how long the
deferred split queue can be. In extreme cases, deferred
On Wed 09-10-19 17:45:09, Kirill A. Shutemov wrote:
> THPs on deferred split queue got split by shrinker if memory pressure
> comes.
>
> In absence of memory pressure, there is no bound on how long the
> deferred split queue can be. In extreme cases, deferred queue can grow
> to tens of
THPs on deferred split queue got split by shrinker if memory pressure
comes.
In absence of memory pressure, there is no bound on how long the
deferred split queue can be. In extreme cases, deferred queue can grow
to tens of gigabytes.
It is suboptimal: even without memory pressure we can find
3 matches
Mail list logo
