Re: [PATCH v30 08/20] x86/sgx: Add functions to allocate and free EPC pages

[Jarkko Sakkinen]

On Fri, May 29, 2020 at 11:13:09AM +0300, Jarkko Sakkinen wrote:
> On Thu, May 28, 2020 at 08:37:16PM -0700, Sean Christopherson wrote:
> > On Fri, May 29, 2020 at 06:28:16AM +0300, Jarkko Sakkinen wrote:
> > > On Thu, May 28, 2020 at 12:59:17PM -0700, Sean Christopherson wrote:
> > > > On Thu, May 28, 2020 at 10:07:18PM +0300, Jarkko Sakkinen wrote:
> > > > >  * sgx_grab_page() - Grab a free EPC page
> > > > >  * @owner:	the owner of the EPC page
> > > > >  * @reclaim:	reclaim pages if necessary
> > > > >  *
> > > > >  * Iterate through EPC sections and borrow a free EPC page to the caller. When a
> > > > >  * page is no longer needed it must be released with sgx_free_page(). If
> > > > >  * @reclaim is set to true, directly reclaim pages when we are out of pages. No
> > > > >  * mm's can be locked when @reclaim is set to true.
> > > > >  *
> > > > >  * Finally, wake up ksgxswapd when the number of pages goes below the watermark
> > > > >  * before returning back to the caller.
> > > > >  *
> > > > >  * Return:
> > > > >  *   an EPC page,
> > > > >  *   -errno on error
> > > > >  */
> > > > > 
> > > > > I also rewrote the kdoc.
> > > > > 
> > > > > I do agree that sgx_try_grab_page() should be renamed as __sgx_grab_page().
> > > > 
> > > > FWIW, I really, really dislike "grab".  The nomenclature for normal memory
> > > > and pages uses "alloc" when taking a page off a free list, and "grab" when
> > > > elevating the refcount.  I don't understand the motivation for diverging
> > > > from that.  SGX is weird enough as is, using names that don't align with
> > > > exist norms will only serve to further obfuscate the code.
> > > 
> > > OK, what would be a better name then? The semantics are not standard
> > > memory allocation semantics in the first place. And kdoc in v30 speaks
> > > about grabbing.
> > 
> > In what way are they not standard allocation semantics?  sgx_alloc_page()
> > is an API to allocate (EPC) memory on-demand, sgx_free_page() is its partner
> > to free that memory when it is no longer needed.  There are many different
> > ways to manage and allocate memory, but the basic premise is the same for
> > all and no different than what we're doing.
> 
> I end up to (ignoring unchanged names):
> 
> - __sgx_alloc_epc_page_from_section() (static)
> - __sgx_alloc_epc_page()
> - sgx_alloc_epc_page()
> - sgx_setup_epc_section()

OK, also sgx_free_epc_page (not sgx_free_page() as before).

/Jarkko