Re: [PATCH v4 1/3] x86/sgx: Use sgx_free_epc_page() in sgx_reclaim_pages()

[Jarkko Sakkinen <jarkko@kernel.org>]

On Mon, Mar 15, 2021 at 09:06:29PM +0200, Jarkko Sakkinen wrote:
> On Mon, Mar 15, 2021 at 08:32:13AM -0700, Dave Hansen wrote:
> > On 3/13/21 8:01 AM, Jarkko Sakkinen wrote:
> > > Replace the ad-hoc code with a sgx_free_epc_page(), in order to make sure
> > > that all the relevant checks and book keeping is done, while freeing a
> > > borrowed EPC page, and remove redundant code. EREMOVE inside
> > > sgx_free_epc_page() does not change the semantics, as EREMOVE to an
> > > uninitialize pages is a nop.
> > 
> >   ^ uninitialized
> > 
> > I know this is a short patch, but this changelog still falls a bit short
> > for me.
> > 
> > Why is this patch a part of _this_ series?  What *problem* does it
> > solve, related to this series?
> 
> I'm thinking of merging sgx_epc_section and sgx_numa_node. That's why I
> kept it as part of the series. 
> 
> Also, in any case it's better to clean up duplicate functionality. The
> code is essentially open coded implementation of sgx_free_epc_page()
> without EREMOVE.
> 
> > It would also be nice to remind me why the EREMOVE is redundant.  Why
> > didn't we need one before?  What put the page in the uninitialized
> > state?  Is EREMOVE guaranteed to do nothing?  How expensive is it?
> 
> EREMOVE gets removed by KVM series from sgx_free_epc_page() anyway.
> 
> Maybe should re-send this patch, or series, after KVM series is merged.
> Then there is no explaining with EREMOVE, as sgx_free_epc_page() won't
> contain it.

Anyway, forgot to put the end statement: I'm cool with dropping this but
I'll also send this right after KVM SGX series has landed as separate
patch, if I drop this now.

/Jarkko