On Fri Apr 5, 2024 at 6:07 AM EEST, Huang, Kai wrote:
> On Thu, 2024-04-04 at 12:05 -0500, Haitao Huang wrote:
> > > > -static inline int sgx_cgroup_try_charge(struct sgx_cgroup *sgx_cg)
> > > > +static inline int sgx_cgroup_try_charge(struct sgx_cgroup *sgx_cg,
> > > > enum sgx_reclaim r)
> > >
On Thu, 2024-04-04 at 12:05 -0500, Haitao Huang wrote:
> > > -static inline int sgx_cgroup_try_charge(struct sgx_cgroup *sgx_cg)
> > > +static inline int sgx_cgroup_try_charge(struct sgx_cgroup *sgx_cg,
> > > enum sgx_reclaim r)
> >
> > Is the @r here intentional for shorter typing?
> >
>
>
On Thu, 2024-04-04 at 12:05 -0500, Haitao Huang wrote:
> > Please also mention why "leaving asynchronous reclamation to later
> > patch(es)" is
> > fine. E.g., it won't break anything I suppose.
> >
>
> Right. Pages are still in the global list at the moment and only global
> reclaiming is
Hi Kai,
Thanks for your suggestions. I'll adopt most of it as it.
Minor details below.
On Wed, 03 Apr 2024 08:08:28 -0500, Huang, Kai wrote:
On Wed, 2024-03-27 at 17:22 -0700, Haitao Huang wrote:
From: Kristen Carlson Accardi
When a cgroup usage reaches its limit, and it is to be charged,
On Wed, 2024-03-27 at 17:22 -0700, Haitao Huang wrote:
> From: Kristen Carlson Accardi
>
> When a cgroup usage reaches its limit, and it is to be charged, i.e.,
> sgx_cgroup_try_charge() called for new allocations, the cgroup needs to
> reclaim pages from its LRU or LRUs of its descendants to
5 matches
Mail list logo
