On Thu, May 22, 2014 at 7:50 AM, Rob Scott <[email protected]>wrote:
> > What is the advantage? What is wrong with the standard key8? The (E)CSA > resident blocks do not gain from the PPT entry and the server blocks are > not available to the client code. > > A lot of the system services invoked by software that I have written > require PKM 0-7. > > I try to avoid key-switching as much as possible - having a software > product with a E(CSA) anchor block in key 0-7 and then executing in key8 > means that every time you want to change something in the anchor you have > to switch keys. Some products approach this issue by running in Key0 all > the time - personally, I think the risk of unintentional storage overlays > make this an uncomfortable solution. > Why not use MVCDK to update the non-key8 storage? > > If your server program runs in, for example, key2 - then : > > (o) You can keep key-switching to a minimum > (o) The anchor block in (E)CSA can be in the same key2 > (o) The server programs private storage is in key2 > (o) The chances of your server code overlaying storage is drastically > reduced. > (o) Other minor advantages when looking at dumps and diagnostics is that > storage owned by your server code is much easier to spot and QA for storage > leak analysis. > I especially agree with the last two. And I'll add one more. If the data might contain "confidential" information, such as passwords or "keys", then you can make your (E)CSA fetch protected to decrease the likelihood of them being disclosed. > > > Rob Scott > Lead Developer > Rocket Software > -- There is nothing more pleasant than traveling and meeting new people! Genghis Khan Maranatha! <>< John McKown
