linux-stable.git/net/mac802154, branch v6.3

Merge tag 'ieee802154-for-net-next-2023-02-20' of git://git.kernel.org/pub/scm/linux/kernel/git/sschmidt/wpan-next

2023-02-21T00:40:52+00:00

Stefan Schmidt says:

====================
pull-request: ieee802154-next 2023-02-20

Miquel Raynal build upon his earlier work and introduced two new
features into the ieee802154 stack. Beaconing to announce existing
PAN's and passive scanning to discover the beacons and associated
PAN's. The matching changes to the userspace configuration tool
have been posted as well and will be released together with the
kernel release.

Arnd Bergmann and Dmitry Torokhov worked on converting the
at86rf230 and cc2520 drivers away from the unused platform_data
usage and towards the new gpiod API. (I had to add a revert as
Dmitry found a regression on an already pushed tree on my side).

Changes since v1 (pull request 2023-02-02)
- Netlink API extack and NLA_POLICY* usage as suggested by Jakub
- Removed always true condition found by kernel test robot
- Simplify device removal with running background job for scanning
- Fix problems with beacon sending in some cases by using the MLME
  tx path

* tag 'ieee802154-for-net-next-2023-02-20' of git://git.kernel.org/pub/scm/linux/kernel/git/sschmidt/wpan-next:
  ieee802154: Drop device trackers
  mac802154: Fix an always true condition
  mac802154: Send beacons using the MLME Tx path
  ieee802154: Change error code on monitor scan netlink request
  ieee802154: Convert scan error messages to extack
  ieee802154: Use netlink policies when relevant on scan parameters
  ieee802154: at86rf230: switch to using gpiod API
  ieee802154: at86rf230: drop support for platform data
  Revert "at86rf230: convert to gpio descriptors"
  cc2520: move to gpio descriptors
  mac802154: Avoid superfluous endianness handling
  at86rf230: convert to gpio descriptors
  mac802154: Handle basic beaconing
  ieee802154: Add support for user beaconing requests
  mac802154: Handle passive scanning
  mac802154: Add MLME Tx locked helpers
  mac802154: Prepare forcing specific symbol duration
  ieee802154: Introduce a helper to validate a channel
  ieee802154: Define a beacon frame header
  ieee802154: Add support for user scanning requests
====================

Link: https://lore.kernel.org/r/20230220213749.386451-1-stefan@datenfreihafen.org
Signed-off-by: Jakub Kicinski

mac802154: Fix an always true condition

2023-02-18T15:47:26+00:00

At this stage we simply do not care about the delayed work value,
because active scan is not yet supported, so we can blindly queue
another work once a beacon has been sent.

It fixes a smatch warning:
    mac802154_beacon_worker() warn: always true condition
    '(local->beacon_interval >= 0) => (0-u32max >= 0)'

Fixes: 3accf4762734 ("mac802154: Handle basic beaconing")
Reported-by: kernel test robot 
Signed-off-by: Miquel Raynal 
Link: https://lore.kernel.org/r/20230214135035.1202471-6-miquel.raynal@bootlin.com
Signed-off-by: Stefan Schmidt

mac802154: Send beacons using the MLME Tx path

2023-02-18T15:44:53+00:00

Using ieee802154_subif_start_xmit() to bypass the net queue when
sending beacons is broken because it does not acquire the
HARD_TX_LOCK(), hence not preventing datagram buffers to be smashed by
beacons upon contention situation. Using the mlme_tx helper is not the
best fit either but at least it is not buggy and has little-to-no
performance hit. More details are given in the comment explaining this
choice in the code.

Fixes: 3accf4762734 ("mac802154: Handle basic beaconing")
Reported-by: Alexander Aring 
Signed-off-by: Miquel Raynal 
Link: https://lore.kernel.org/r/20230214135035.1202471-5-miquel.raynal@bootlin.com
Signed-off-by: Stefan Schmidt

mac802154: Avoid superfluous endianness handling

2023-01-30T16:16:28+00:00

When compiling scan.c with C=1, Sparse complains with:

   sparse:     expected unsigned short [usertype] val
   sparse:     got restricted __le16 [usertype] pan_id
   sparse: sparse: cast from restricted __le16

   sparse:     expected unsigned long long [usertype] val
   sparse:     got restricted __le64 [usertype] extended_addr
   sparse: sparse: cast from restricted __le64

The tool is right, both pan_id and extended_addr already are rightfully
defined as being __le16 and __le64 on both sides of the operations and
do not require extra endianness handling.

Fixes: 3accf4762734 ("mac802154: Handle basic beaconing")
Reported-by: kernel test robot 
Signed-off-by: Miquel Raynal 
Link: https://lore.kernel.org/r/20230130154306.114265-1-miquel.raynal@bootlin.com
Signed-off-by: Stefan Schmidt

mac802154: Handle basic beaconing

2023-01-28T12:55:10+00:00

Implement the core hooks in order to provide the softMAC layer support
for sending beacons. Coordinators may be requested to send beacons in a
beacon enabled PAN in order for the other devices around to self
discover the available PANs automatically.

Changing the channels is prohibited while a beacon operation is
ongoing.

The implementation uses a workqueue triggered at a certain interval
depending on the symbol duration for the current channel and the
interval order provided.

Sending beacons in response to a BEACON_REQ frame (ie. answering active
scans) is not yet supported.

This initial patchset has no security support (llsec).

Co-developed-by: David Girault 
Signed-off-by: David Girault 
Signed-off-by: Miquel Raynal 
Acked-by: Alexander Aring 
Link: https://lore.kernel.org/r/20230125102923.135465-3-miquel.raynal@bootlin.com
Signed-off-by: Stefan Schmidt

mac802154: Handle passive scanning

2023-01-03T18:48:43+00:00

Implement the core hooks in order to provide the softMAC layer support
for passive scans. Scans are requested by the user and can be aborted.

Changing channels manually is prohibited during scans.

The implementation uses a workqueue triggered at a certain interval
depending on the symbol duration for the current channel and the
duration order provided. More advanced drivers with internal scheduling
capabilities might require additional care but there is none mainline
yet.

Received beacons during a passive scan are processed in a work queue and
their result forwarded to the upper layer.

Active scanning is not supported yet.

Co-developed-by: David Girault 
Signed-off-by: David Girault 
Signed-off-by: Miquel Raynal 
Acked-by: Alexander Aring 
Link: https://lore.kernel.org/r/20230103165644.432209-7-miquel.raynal@bootlin.com
Signed-off-by: Stefan Schmidt

mac802154: Add MLME Tx locked helpers

2023-01-03T18:45:44+00:00

These have the exact same behavior as before, except they expect the
rtnl to be already taken (and will complain otherwise). This allows
performing MLME transmissions from different contexts.

Signed-off-by: Miquel Raynal 
Acked-by: Alexander Aring 
Link: https://lore.kernel.org/r/20230103165644.432209-6-miquel.raynal@bootlin.com
Signed-off-by: Stefan Schmidt

mac802154: Prepare forcing specific symbol duration

2023-01-03T18:41:16+00:00

The scan logic will bypass the whole ->set_channel() logic from the top
by calling the driver hook to just switch between channels when
required.

We can no longer rely on the "current" page/channel settings to set the
right symbol duration. Let's add these as new parameters to allow
providing the page/channel couple that we want.

There is no functional change.

Signed-off-by: Miquel Raynal 
Acked-by: Alexander Aring 
Link: https://lore.kernel.org/r/20230103165644.432209-5-miquel.raynal@bootlin.com
Signed-off-by: Stefan Schmidt

mac802154: Fix possible double free upon parsing error

2022-12-19T10:38:12+00:00

Commit 4d1c7d87030b ("mac802154: Move an skb free within the rx path")
tried to simplify error handling within the receive path by moving the
kfree_skb() call at the very end of the top-level function but missed
one kfree_skb() called upon frame parsing error. Prevent this possible
double free from happening.

Fixes: 4d1c7d87030b ("mac802154: Move an skb free within the rx path")
Reported-by: Dan Carpenter 
Signed-off-by: Miquel Raynal 
Link: https://lore.kernel.org/r/20221216235742.646134-1-miquel.raynal@bootlin.com
Signed-off-by: Stefan Schmidt

Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

2022-12-09T02:19:59+00:00

No conflicts.

Signed-off-by: Jakub Kicinski