linux-dev/net/mac802154, branch master

mac802154: Fix LQI recording

2022-10-24T09:07:39Z

Back in 2014, the LQI was saved in the skb control buffer (skb->cb, or mac_cb(skb)) without any actual reset of this area prior to its use. As part of a useful rework of the use of this region, 32edc40ae65c ("ieee802154: change _cb handling slightly") introduced mac_cb_init() to basically memset the cb field to 0. In particular, this new function got called at the beginning of mac802154_parse_frame_start(), right before the location where the buffer got actually filled. What went through unnoticed however, is the fact that the very first helper called by device drivers in the receive path already used this area to save the LQI value for later extraction. Resetting the cb field "so late" led to systematically zeroing the LQI. If we consider the reset of the cb field needed, we can make it as soon as we get an skb from a device driver, right before storing the LQI, as is the very first time we need to write something there. Cc: stable@vger.kernel.org Fixes: 32edc40ae65c ("ieee802154: change _cb handling slightly") Signed-off-by: Miquel Raynal Acked-by: Alexander Aring Link: https://lore.kernel.org/r/20221020142535.1038885-1-miquel.raynal@bootlin.com Signed-off-by: Stefan Schmidt

net: mac802154: Fix a condition in the receive path

2022-08-29T09:10:22Z

Upon reception, a packet must be categorized, either it's destination is the host, or it is another host. A packet with no destination addressing fields may be valid in two situations: - the packet has no source field: only ACKs are built like that, we consider the host as the destination. - the packet has a valid source field: it is directed to the PAN coordinator, as for know we don't have this information we consider we are not the PAN coordinator. There was likely a copy/paste error made during a previous cleanup because the if clause is now containing exactly the same condition as in the switch case, which can never be true. In the past the destination address was used in the switch and the source address was used in the if, which matches what the spec says. Cc: stable@vger.kernel.org Fixes: ae531b9475f6 ("ieee802154: use ieee802154_addr instead of *_sa variants") Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/r/20220826142954.254853-1-miquel.raynal@bootlin.com Signed-off-by: Stefan Schmidt

net: mac802154: Fix symbol durations

2022-04-30T18:29:47Z

There are two major issues in the logic calculating the symbol durations based on the page/channel: - The page number is used in place of the channel value. - The BIT() macro is missing because we want to check the channel value against a bitmask. Fix these two errors and apologize loudly for this mistake. Signed-off-by: Miquel Raynal Acked-by: Alexander Aring Link: https://lore.kernel.org/r/20220428164140.251965-1-miquel.raynal@bootlin.com Signed-off-by: Stefan Schmidt

net: mac802154: Create an error helper for asynchronous offloading errors

2022-04-25T18:51:12Z

A few drivers do the full transmit operation asynchronously, which means that a bus error that happens when forwarding the packet to the transmitter or a timeout happening when offloading the request to the transmitter will not be reported immediately. The solution in this case is to call this new helper to free the necessary resources, restart the queue and always return the same generic TRAC error code: IEEE802154_SYSTEM_ERROR. Suggested-by: Alexander Aring Signed-off-by: Miquel Raynal Acked-by: Alexander Aring Link: https://lore.kernel.org/r/20220407100903.1695973-6-miquel.raynal@bootlin.com Signed-off-by: Stefan Schmidt

net: mac802154: Create an offloaded transmission error helper

2022-04-25T18:51:12Z

So far there is only a helper for successful transmissions, which led device drivers to implement their own handling in case of error. Unfortunately, we really need all the drivers to give the hand back to the core once they are done in order to be able to build a proper synchronous API. So let's create a _xmit_error() helper and take this opportunity to fill the new device-global field storing Tx statuses. Suggested-by: Alexander Aring Signed-off-by: Miquel Raynal Acked-by: Alexander Aring Link: https://lore.kernel.org/r/20220407100903.1695973-5-miquel.raynal@bootlin.com Signed-off-by: Stefan Schmidt

net: mac802154: Save a global error code on transmissions

2022-04-25T18:51:12Z

So far no error is returned from a failing transmission. However it might sometimes be useful, and particularly easy to use during sync transfers (for certain MLME commands). Let's create an internal variable for that, global to the device. Right now only success are registered, which is rather useless, but soon we will have more situations filling this field. Signed-off-by: Miquel Raynal Acked-by: Alexander Aring Link: https://lore.kernel.org/r/20220407100903.1695973-4-miquel.raynal@bootlin.com Signed-off-by: Stefan Schmidt

net: mac802154: Set durations automatically

2022-02-10T14:41:58Z

As depicted in the IEEE 802.15.4 specification, modulation/bands are tight to a number of page/channels so we can for most of them derive the durations automatically. The two locations that must call this new helper to set the variou symbol durations are: - when manually requesting a channel change though the netlink interface - at PHY creation, once the device driver has set the default page/channel If an information is missing, the symbol duration is not touched, a debug message is eventually printed. This keeps the compatibility with the unconverted drivers for which it was too complicated for me to find their precise information. If they initially provided a symbol duration, it would be kept. If they don't, the symbol duration value is left untouched. Once the symbol duration derived, the lifs and sifs durations are updated as well. Signed-off-by: Miquel Raynal Acked-by: Alexander Aring Link: https://lore.kernel.org/r/20220201180629.93410-4-miquel.raynal@bootlin.com Signed-off-by: Stefan Schmidt

net: mac802154: Convert the symbol duration into nanoseconds

2022-02-10T14:41:58Z

Tdsym is often given in the spec as pretty small numbers in microseconds and hence was reflected in the code as symbol_duration and was stored as a u8. Actually, for UWB PHYs, the symbol duration is given in nanoseconds and are as precise as picoseconds. In order to handle better these PHYs, change the type of symbol_duration to u32 and store this value in nanoseconds. All the users of this variable are updated in a mechanical way. Signed-off-by: Miquel Raynal Acked-by: Alexander Aring Link: https://lore.kernel.org/r/20220201180629.93410-3-miquel.raynal@bootlin.com Signed-off-by: Stefan Schmidt

mac802154: use dev_addr_set() - manual

2021-10-20T13:27:40Z

Commit 406f42fa0d3c ("net-next: When a bond have a massive amount of VLANs...") introduced a rbtree for faster Ethernet address look up. To maintain netdev->dev_addr in this tree we need to make all the writes to it got through appropriate helpers. Signed-off-by: Jakub Kicinski Signed-off-by: David S. Miller

mac802154: use dev_addr_set()

2021-10-20T13:27:40Z