[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d17c270bcbf631736d52b762dad8a6758d7dc951.camel@linux.intel.com>
Date: Fri, 02 Dec 2022 13:51:10 -0800
From: Kristen Carlson Accardi <kristen@...ux.intel.com>
To: Dave Hansen <dave.hansen@...el.com>, jarkko@...nel.org,
dave.hansen@...ux.intel.com, tj@...nel.org,
linux-kernel@...r.kernel.org, linux-sgx@...r.kernel.org,
cgroups@...r.kernel.org, Thomas Gleixner <tglx@...utronix.de>,
Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
x86@...nel.org, "H. Peter Anvin" <hpa@...or.com>
Cc: zhiquan1.li@...el.com, Sean Christopherson <seanjc@...gle.com>
Subject: Re: [PATCH v2 04/18] x86/sgx: Use sgx_epc_lru_lists for existing
active page list
On Fri, 2022-12-02 at 13:43 -0800, Dave Hansen wrote:
> On 12/2/22 10:36, Kristen Carlson Accardi wrote:
> > + spin_lock(&sgx_global_lru.lock);
> > for (i = 0; i < SGX_NR_TO_SCAN; i++) {
> > - if (list_empty(&sgx_active_page_list))
> > + epc_page =
> > sgx_epc_pop_reclaimable(&sgx_global_lru);
> > + if (!epc_page)
> > break;
>
> One other nit about the structure of the patches: This introduced
> *both*
> reclaimable and unreclaimable list_heads. But, it has zero use for
> the
> unreclaimable ones during the refactoring here. I probably would
> have
> left out the 'unreclaimable' bits for now.
I know - and originally the addition of unreclaimable was added later,
but when I posted the RFC I felt there was some misunderstanding about
what this data structure was and how it would be used because the
addition of the unreclaimable bits came later. So I stuck both lists in
one so it'd be a better view of what the data structure would look
like.
>
> BTW, this is a nice sign:
>
> > arch/x86/kernel/cpu/sgx/main.c | 39 +++++++++++++++++-------------
> > ----
> > 1 file changed, 19 insertions(+), 20 deletions(-)
>
>
Powered by blists - more mailing lists