Re: [PATCH V4 14/31] x86/sgx: Support VA page allocation without reclaiming

[Haitao Huang]

On Thu, 14 Apr 2022 11:30:34 -0500, Reinette Chatre  
<reinette.chatre@xxxxxxxxx> wrote:

> Hi Jarkko,
>
> On 4/14/2022 4:18 AM, Jarkko Sakkinen wrote:
> > On Wed, 2022-04-13 at 14:10 -0700, Reinette Chatre wrote:
> > > struct sgx_encl should be protected with the mutex
> > > sgx_encl->lock. One exception is sgx_encl->page_cnt that
> > > is incremented (in sgx_encl_grow()) when an enclave page
> > > is added to the enclave. The reason the mutex is not held
> > > is to allow the reclaimer to be called directly if there are
> > > no EPC pages (in support of a new VA page) available at the time.
> > >
> > > Incrementing sgx_encl->page_cnt without sgc_encl->lock held
> > > is currently (before SGX2) safe from concurrent updates because
> > > all paths in which sgx_encl_grow() is called occur before
> > > enclave initialization and are protected with an atomic
> > > operation on SGX_ENCL_IOCTL.
> > >
> > > SGX2 includes support for dynamically adding pages after
> > > enclave initialization where the protection of SGX_ENCL_IOCTL
> > > is not available.
> > >
> > > Make direct reclaim of EPC pages optional when new VA pages
> > > are added to the enclave. Essentially the existing "reclaim"
> > > flag used when regular EPC pages are added to an enclave
> > > becomes available to the caller when used to allocate VA pages
> > > instead of always being "true".
> > >
> > > When adding pages without invoking the reclaimer it is possible
> > > to do so with sgx_encl->lock held, gaining its protection against
> > > concurrent updates to sgx_encl->page_cnt after enclave
> > > initialization.
> > >
> > > No functional change.
> > >
> > > Reported-by: Haitao Huang <haitao.huang@xxxxxxxxx>
> > > Tested-by: Haitao Huang <haitao.huang@xxxxxxxxx>
> > > Signed-off-by: Reinette Chatre <reinette.chatre@xxxxxxxxx>
> >
> > Reviewed-by: Jarkko Sakkinen <jarkko@xxxxxxxxxx>
>
> Thank you.
>
> > Nit: I don't think tested-by is in the right patch here. Maybe
> > Haitao's tested-by should be moved into patch that actually adds
> > support for EAUG? Not something I would NAK this patch, just
> > wondering...
>
> Yes, that is a good point. While this is the bulk of the fix where
> the new API is introduced, the test is only applicable when this API
> is used and that is in "x86/sgx: Support adding of pages to an
> initialized enclave". I will move the "Tested-by" to that patch.

You can also add my Tested-by for patches adding the new IOCTLs.
Our team and I have tested EAUG on #PF,  modifying types and permissions  
with Intel SGX SDK/PSW.
Thanks
Haitao