[v5,00/10] DPP PKEX Changes

Message

James Prestwood Nov. 8, 2023, 5:21 p.m. UTC

v5:
 * Separate the DPP/PKEX interfaces using a new enum
   to distinguish which interfaces is using the dpp_sm.
   Similarly added a refcount to the dpp_sm object so when
   one interface goes away it won't destroy the SM for the
   other.
 * Removed unneeded checks on the identifier, and added
   some:
      - Assume utf8/null terminated when coming from DBus
        (just check length)
      - Validate utf8/length when coming from the peer.
 * Enforce any user-disabled bands when generating the
   default frequency list.
 * Return error on Stop() if not started. This is an API
   change, but I thought it made sense so its very clear
   when Stop() is called on the incorrect interface.
 * Included autotest changes.

James Prestwood (10):
  dpp-util: fix typo, 'REQUST'
  dpp: introduce dpp_interface type, prep for PKEX
  auto-t: fix testDPP after Stop() change
  doc: document Stop() correctly for both DPP interfaces
  dpp: initial version of PKEX enrollee support
  dpp: initial version of PKEX configurator support
  dpp: Add StartConfigurator, PKEX agent support
  auto-t: add utils for wpa_supplicant PKEX
  auto-t: add APIs for PKEX
  auto-t: add DPP PKEX tests

 autotests/testDPP/connection_test.py |    9 +-
 autotests/testDPP/hostapd.conf       |    2 +-
 autotests/testDPP/hw.conf            |    5 +-
 autotests/testDPP/pkex_test.py       |  209 ++++
 autotests/testDPP/ssidCCMP.psk       |   12 +
 autotests/util/iwd.py                |  152 ++-
 autotests/util/wpas.py               |   40 +-
 doc/device-provisioning-api.txt      |    8 +-
 src/dpp-util.h                       |    2 +-
 src/dpp.c                            | 1722 ++++++++++++++++++++++++--
 10 files changed, 2056 insertions(+), 105 deletions(-)
 create mode 100644 autotests/testDPP/pkex_test.py

Comments

Denis Kenzior Nov. 9, 2023, 4:35 p.m. UTC | #1

Hi James,

On 11/8/23 11:21, James Prestwood wrote:
> v5:
>   * Separate the DPP/PKEX interfaces using a new enum
>     to distinguish which interfaces is using the dpp_sm.
>     Similarly added a refcount to the dpp_sm object so when
>     one interface goes away it won't destroy the SM for the
>     other.
>   * Removed unneeded checks on the identifier, and added
>     some:
>        - Assume utf8/null terminated when coming from DBus
>          (just check length)
>        - Validate utf8/length when coming from the peer.
>   * Enforce any user-disabled bands when generating the
>     default frequency list.
>   * Return error on Stop() if not started. This is an API
>     change, but I thought it made sense so its very clear
>     when Stop() is called on the incorrect interface.
>   * Included autotest changes.
> 

All applied, thanks.

Regards,
-Denis