Dave Jiang <[email protected]> wrote:

> A thought occurred to me. For password update, would it make sense to do
> this instead:
> 1. get the existing key by: request_key("nvdimm:xxxxxxxx")
> 2. get the new key by: request_key("nvdimm.update:xxxxxxxx")
> 3. verify key with hardware
>       on success, copy new payload to existing key payload
> 4. invalidate "nvdimm.update" key
> 
> This way then we won't have to mess with needing the invalidated key to
> be garbage collected. Thoughts?

Can you tell me at what points you actually access the key?

David
_______________________________________________
Linux-nvdimm mailing list
[email protected]
https://lists.01.org/mailman/listinfo/linux-nvdimm

Reply via email to