On Thu, Oct 17, 2019 at 09:10:54PM +0300, Jarkko Sakkinen wrote: > On Wed, Oct 16, 2019 at 11:37:33AM -0700, Sean Christopherson wrote: > > As requested, the merge of the EPC and EADD bug fix series. The headliners > > are two critical bug fixes for a memory leak in sgx_encl_destroy() and a > > livelock due to the EPC page free count getting corrupted. > > BTW, are you still planning to send multi-page EADD? I think it should > be done. It is obviously superior design when combined with my small > twist where mrmask is calculated inside the kernel. > > Not only does it provide multi-page EADD but you could always do one > ioctl per area. With mrmask in parameters you end up needing two if the > user space is designed in a way that measurement is not a multiple of a > page. > > If mrmask would make sense in the parameters, then a single-page EADD > would be cool but I think we want to render it out completely. My earlier comments about doing first a single page EADD were based on API where you provide mrmask. /Jarkko
