platform/kernel/linux-starfive.git
2 years agonet: dpaa: Use mac_dev variable in dpaa_netdev_init
Sean Anderson [Fri, 2 Sep 2022 21:57:33 +0000 (17:57 -0400)]
net: dpaa: Use mac_dev variable in dpaa_netdev_init

There are several references to mac_dev in dpaa_netdev_init. Make things a
bit more concise by adding a local variable for it.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fman: Change return type of disable to void
Sean Anderson [Fri, 2 Sep 2022 21:57:32 +0000 (17:57 -0400)]
net: fman: Change return type of disable to void

When disabling, there is nothing we can do about errors. In fact, the
only error which can occur is misuse of the API. Just warn in the mac
driver instead.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fman: Clean up error handling
Sean Anderson [Fri, 2 Sep 2022 21:57:31 +0000 (17:57 -0400)]
net: fman: Clean up error handling

This removes the _return label, since something like

err = -EFOO;
goto _return;

can be replaced by the briefer

return -EFOO;

Additionally, this skips going to _return_of_node_put when dev_node has
already been put (preventing a double put).

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fman: Specify type of mac_dev for exception_cb
Sean Anderson [Fri, 2 Sep 2022 21:57:30 +0000 (17:57 -0400)]
net: fman: Specify type of mac_dev for exception_cb

Instead of using a void pointer for mac_dev, specify its type
explicitly.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fman: Use mac_dev for some params
Sean Anderson [Fri, 2 Sep 2022 21:57:29 +0000 (17:57 -0400)]
net: fman: Use mac_dev for some params

Some params are already present in mac_dev. Use them directly instead of
passing them through params.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fman: Pass params directly to mac init
Sean Anderson [Fri, 2 Sep 2022 21:57:28 +0000 (17:57 -0400)]
net: fman: Pass params directly to mac init

Instead of having the mac init functions call back into the fman core to
get their params, just pass them directly to the init functions.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fman: Map the base address once
Sean Anderson [Fri, 2 Sep 2022 21:57:27 +0000 (17:57 -0400)]
net: fman: Map the base address once

We don't need to remap the base address from the resource twice (once in
mac_probe() and again in set_fman_mac_params()). We still need the
resource to get the end address, but we can use a single function call
to get both at once.

While we're at it, use platform_get_mem_or_io and devm_request_resource
to map the resource. I think this is the more "correct" way to do things
here, since we use the pdev resource, instead of creating a new one.
It's still a bit tricky, since we need to ensure that the resource is a
child of the fman region when it gets requested.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fman: Remove internal_phy_node from params
Sean Anderson [Fri, 2 Sep 2022 21:57:26 +0000 (17:57 -0400)]
net: fman: Remove internal_phy_node from params

This member was used to pass the phy node between mac_probe and the
mac-specific initialization function. But now that the phy node is
gotten in the initialization function, this parameter does not serve a
purpose. Remove it, and do the grabbing of the node/grabbing of the phy
in the same place.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fman: Inline several functions into initialization
Sean Anderson [Fri, 2 Sep 2022 21:57:25 +0000 (17:57 -0400)]
net: fman: Inline several functions into initialization

There are several small functions which were only necessary because the
initialization functions didn't have access to the mac private data. Now
that they do, just do things directly.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fman: Mark mac methods static
Sean Anderson [Fri, 2 Sep 2022 21:57:24 +0000 (17:57 -0400)]
net: fman: Mark mac methods static

These methods are no longer accessed outside of the driver file, so mark
them as static.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fman: Move initialization to mac-specific files
Sean Anderson [Fri, 2 Sep 2022 21:57:23 +0000 (17:57 -0400)]
net: fman: Move initialization to mac-specific files

This moves mac-specific initialization to mac-specific files. This will
make it easier to work with individual macs. It will also make it easier
to refactor the initialization to simplify the control flow. No
functional change intended.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Acked-by: Camelia Groza <camelia.groza@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agor8169: remove useless PCI region size check
Heiner Kallweit [Fri, 2 Sep 2022 21:16:52 +0000 (23:16 +0200)]
r8169: remove useless PCI region size check

Let's trust the hardware here and remove this useless check.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'lan937x-phy-link-interrupt'
David S. Miller [Mon, 5 Sep 2022 12:06:40 +0000 (13:06 +0100)]
Merge branch 'lan937x-phy-link-interrupt'

Arun Ramadoss says:

====================
net: dsa: microchip: lan937x: enable interrupt for internal phy link detection

This patch series enables the internal phy link detection for lan937x using the
interrupt method. lan937x acts as the interrupt controller for the internal
ports and phy, the irq_domain is registered for the individual ports and in
turn for the individual port interrupts.

RFC v3 -> Patch v1
- Removed the RFC v3 1/3 from the series - changing exit from reset
- Changed the variable name in ksz_port from irq to pirq
- Added the check for return value of irq_find_mapping during phy irq
  registeration.
- Moved the clearing of POR_READY_INT from girq_thread_fn to
  lan937x_reset_switch

RFC v2 -> v3
- Used the interrupt controller implementation of phy link

Changes in RFC v2
- fixed the compilation issue
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: microchip: lan937x: add interrupt support for port phy link
Arun Ramadoss [Fri, 2 Sep 2022 10:32:10 +0000 (16:02 +0530)]
net: dsa: microchip: lan937x: add interrupt support for port phy link

This patch enables the interrupts for internal phy link detection for
LAN937x. The interrupt enable bits are active low. There is global
interrupt mask for each port. And each port has the individual interrupt
mask for TAS. QCI, SGMII, PTP, PHY and ACL.
The first level of interrupt domain is registered for global port
interrupt and second level of interrupt domain for the individual port
interrupts. The phy interrupt is enabled in the lan937x_mdio_register
function. Interrupt from which port is raised will be detected based on
the interrupt host data.

Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: microchip: lan937x: clear the POR_READY_INT status bit
Arun Ramadoss [Fri, 2 Sep 2022 10:32:09 +0000 (16:02 +0530)]
net: dsa: microchip: lan937x: clear the POR_READY_INT status bit

In the lan937x_reset_switch(), it masks all the switch and port
registers. In the Global_Int_status register, POR ready bit is write 1
to clear bit and all other bits are read only. So, this patch clear the
por_ready_int status bit by writing 1.

Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: microchip: add reference to ksz_device inside the ksz_port
Arun Ramadoss [Fri, 2 Sep 2022 10:32:08 +0000 (16:02 +0530)]
net: dsa: microchip: add reference to ksz_device inside the ksz_port

struct ksz_port doesn't have reference to ksz_device as of now. In order
to find out from which port interrupt has triggered, we need to pass the
struct ksz_port as a host data. When the interrupt is triggered, we can
get the port from which interrupt triggered, but to identify it is phy
interrupt we have to read status register. The regmap structure for
accessing the device register is present in the ksz_device struct. To
access the ksz_device from the ksz_port, the reference is added to it
with port number as well.

Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'ipa-transaction-IDs'
David S. Miller [Mon, 5 Sep 2022 11:47:02 +0000 (12:47 +0100)]
Merge branch 'ipa-transaction-IDs'

Alex Elder says:

====================
net: ipa: start using transaction IDs

A previous group of patches added ID fields to track the state of
transactions:
  https://lore.kernel.org/netdev/20220831224017.377745-1-elder@linaro.org

This series starts using those IDs instead of the lists used
previously.  Most of this series involves reworking the function
that determines which transaction is the "last", which determines
when a channel has been quiesed.  The last patch is mainly used to
prove that the new index method of tracking transaction state is
equivalent to the previous use of lists.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: verify a few more IDs
Alex Elder [Fri, 2 Sep 2022 21:02:18 +0000 (16:02 -0500)]
net: ipa: verify a few more IDs

The completed transaction list is used in gsi_channel_trans_complete()
to return the next transaction in completed state.

Add some temporary checks to verify the transaction indicated by the
completed ID matches the one first in this list.

Similarly, we use the pending and completed transaction lists when
cancelling pending transactions in gsi_channel_trans_cancel_pending().

Add temporary checks there to verify the transactions indicated by
IDs match those tracked by these lists.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: further simplify gsi_channel_trans_last()
Alex Elder [Fri, 2 Sep 2022 21:02:17 +0000 (16:02 -0500)]
net: ipa: further simplify gsi_channel_trans_last()

Do a little more refactoring in gsi_channel_trans_last() to simplify
it further.  The resulting code should behave exactly as before.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: simplify gsi_channel_trans_last()
Alex Elder [Fri, 2 Sep 2022 21:02:16 +0000 (16:02 -0500)]
net: ipa: simplify gsi_channel_trans_last()

Using a little logic we can simplify gsi_channel_trans_last().

The first condition in that function looks like this:
    if (trans_info->allocated_id != trans_info->free_id)
And if that's false, we proceed to the next one:
    if (trans_info->committed_id != trans_info->allocated_id)

Failure of the first test implies:
    trans_info->allocated_id == trans_info->free_id
And therefore, the second one can be rewritten this way:
    if (trans_info->committed_id != trans_info->free_id)

Substituting free_id for allocated_id and committed_id can also be
done in the code blocks executed when these conditions yield true.
The net result is that all three blocks for TX endpoints can be
consolidated into just one.

The two blocks of code at the end of that function (used for both TX
and RX channels) can be similarly consolidated into a single block.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: use IDs exclusively for last transaction
Alex Elder [Fri, 2 Sep 2022 21:02:15 +0000 (16:02 -0500)]
net: ipa: use IDs exclusively for last transaction

Always use transaction IDs when finding the "last" transaction to
await when quiescing a channel.  This basically extends what was
done in the previous patch to all other transaction state IDs.

As a result we are no longer updating any transaction lists inside
gsi_channel_trans_last(), so there's no need to take the spinlock.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: use IDs for last allocated transaction
Alex Elder [Fri, 2 Sep 2022 21:02:14 +0000 (16:02 -0500)]
net: ipa: use IDs for last allocated transaction

Use the allocated and free transaction IDs to determine whether the
"last" transaction used for quiescing a channel is in allocated
state.  The last allocated transaction that has not been committed
(if any) immediately precedes the first free transaction in the
transaction array.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: rework last transaction determination
Alex Elder [Fri, 2 Sep 2022 21:02:13 +0000 (16:02 -0500)]
net: ipa: rework last transaction determination

When quiescing a channel, we find the "last" transaction, which is
the latest one to have been allocated.  (New transaction allocation
will have been prevented by the time this is called.)

Currently we do this by looking for the first non-empty transaction
list in each state, then return the last entry from that last.
Instead, determine the last entry in each list (if any) and return
that entry if found.

Temporarily (locally) introduce list_last_entry_or_null() as a
helper for this, mirroring list_first_entry_or_null().  This macro
definition will be removed by an upcoming patch.

Remove the temporary warnings added by the previous commit.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agor8169: use devm_clk_get_optional_enabled() to simplify the code
Heiner Kallweit [Fri, 2 Sep 2022 20:52:34 +0000 (22:52 +0200)]
r8169: use devm_clk_get_optional_enabled() to simplify the code

Now that we have devm_clk_get_optional_enabled(), we don't have to
open-code it.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agor8169: remove comment about apparently non-existing chip versions
Heiner Kallweit [Fri, 2 Sep 2022 20:21:57 +0000 (22:21 +0200)]
r8169: remove comment about apparently non-existing chip versions

It's not clear where these entries came from, and as I wrote in the
comment: Not even Realtek's r8101 driver knows these chip id's.
So remove the comment.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agor8169: merge handling of chip versions 12 and 17 (RTL8168B)
Heiner Kallweit [Fri, 2 Sep 2022 20:10:53 +0000 (22:10 +0200)]
r8169: merge handling of chip versions 12 and 17 (RTL8168B)

It's not clear why XID's 380 and 381..387 ever got different chip
version id's. VER_12 and VER_17 are handled exactly the same.
Therefore merge handling under the VER_17 umbrella.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'altera-tse-phylink'
David S. Miller [Mon, 5 Sep 2022 09:16:53 +0000 (10:16 +0100)]
Merge branch 'altera-tse-phylink'

Maxime Chevallier says:

====================
net: altera: tse: phylink conversion

This is V4 of a series converting the Altera TSE driver to phylink,
introducing a new PCS driver along the way.

The Altera TSE can be built with a SGMII/1000BaseX PCS, allowing to use
SFP ports with this MAC, which is the end goal of adding phylink support
and a proper PCS driver.

The PCS itself can either be mapped in the MAC's register space, in that
case, it's accessed through 32 bits registers, with the higher 16 bits
always 0. Alternatively, it can sit on its own register space, exposing
16 bits registers, some of which ressemble the standard PHY registers.

To tackle that rework, several things needs updating, starting by the DT
binding, since we add support for a new register range for the PCS.

Hence, the first patch of the series is a conversion to YAML of the
existing binding.

Then, patch 2 does a bit of simple cleanup to the TSE driver, using nice
reverse xmas tree definitions.

Patch 3 adds the actual PCS driver, as a standalone driver. Some future
series will then reuse that PCS driver from the dwmac-socfpga driver,
which implements support for this exact PCS too, allowing to share the
code nicely.

Patch 4 is then a phylink conversion of the altera_tse driver, to use
this new PCS driver.

Finally, patch 5 updates the newly converted DT binding to support the
pcs register range.

This series contains bits and pieces for this conversion, please tell me if
you want me to send it as individual patches.

V4 Changes:
 - Add missing MODULE_* macros to the TSE PCS driver

V3 Changes:
 - YAML binding conversion changes and PCS addition changes thanks to
   Krzysztof's reviews

V2 Changes :
 - Fixed the binding after the YAML conversion
 - Added a pcs_validate() callback
 - Introduced a comment to justify a soft reset for the PCS
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodt-bindings: net: altera: tse: add an optional pcs register range
Maxime Chevallier [Fri, 2 Sep 2022 08:32:05 +0000 (10:32 +0200)]
dt-bindings: net: altera: tse: add an optional pcs register range

Some implementations of the TSE have their PCS as an external bloc,
exposed at its own register range. Document this, and add a new example
showing a case using the pcs and the new phylink conversion to connect
an sfp port to a TSE mac.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: altera: tse: convert to phylink
Maxime Chevallier [Fri, 2 Sep 2022 08:32:04 +0000 (10:32 +0200)]
net: altera: tse: convert to phylink

Convert the Altera Triple Speed Ethernet Controller to phylink.
This controller supports MII, GMII and RGMII with its MAC, and
SGMII + 1000BaseX through a small embedded PCS.

The PCS itself has a register set very similar to what is found in a
typical 802.3 ethernet PHY, but this register set memory-mapped instead
of lying on an mdio bus.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: pcs: add new PCS driver for altera TSE PCS
Maxime Chevallier [Fri, 2 Sep 2022 08:32:03 +0000 (10:32 +0200)]
net: pcs: add new PCS driver for altera TSE PCS

The Altera Triple Speed Ethernet has a SGMII/1000BaseC PCS that can be
integrated in several ways. It can either be part of the TSE MAC's
address space, accessed through 32 bits accesses on the mapped mdio
device 0, or through a dedicated 16 bits register set.

This driver allows using the TSE PCS outside of altera TSE's driver,
since it can be used standalone by other MACs.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: altera: tse: cosmetic change to use reverse xmas tree ordering
Maxime Chevallier [Fri, 2 Sep 2022 08:32:02 +0000 (10:32 +0200)]
net: altera: tse: cosmetic change to use reverse xmas tree ordering

Make the driver code cleaner through a strictly cosmetic change, using
he reverse xmas tree variable declaration ordering.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodt-bindings: net: Convert Altera TSE bindings to yaml
Maxime Chevallier [Fri, 2 Sep 2022 08:32:01 +0000 (10:32 +0200)]
dt-bindings: net: Convert Altera TSE bindings to yaml

Convert the bindings for the Altera Triple-Speed Ethernet to yaml.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge tag 'wireless-next-2022-09-03' of git://git.kernel.org/pub/scm/linux/kernel...
David S. Miller [Sun, 4 Sep 2022 10:24:34 +0000 (11:24 +0100)]
Merge tag 'wireless-next-2022-09-03' of git://git./linux/kernel/git/wireless/wireless-next

Johannes Berg says:

====================
drivers
 - rtw89: large update across the map, e.g. coex, pci(e), etc.
 - ath9k: uninit memory read fix
 - ath10k: small peer map fix and a WCN3990 device fix
 - wfx: underflow

stack
 - the "change MAC address while IFF_UP" change from James
   we discussed
 - more MLO work, including a set of fixes for the previous
   code, now that we have more code we can exercise it more
 - prevent some features with MLO that aren't ready yet
   (AP_VLAN and 4-address connections)
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agowifi: mac80211_hwsim: fix multi-channel handling in netlink RX
Johannes Berg [Fri, 2 Sep 2022 14:12:49 +0000 (16:12 +0200)]
wifi: mac80211_hwsim: fix multi-channel handling in netlink RX

In netlink RX, now that we can actually have multiple
channel contexts for MLO, things don't work well as we
only keep a single pointer, and then on link switching
we might NULL it, and hit the return if the channel is
NULL.

However, we already use mac80211_hwsim_tx_iter() which
deals with all this, so remove the test and adjust the
remaining code a bit.

This then means we no longer use the chanctx pointer,
so remove it as well.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: call drv_sta_state() under sdata_lock() in reconfig
Johannes Berg [Fri, 2 Sep 2022 14:12:51 +0000 (16:12 +0200)]
wifi: mac80211: call drv_sta_state() under sdata_lock() in reconfig

Currently, other paths calling drv_sta_state() hold the mutex
and therefore drivers can assume that, and look at links with
that protection. Fix that for the reconfig path as well; to
do it more easily use ieee80211_reconfig_stations() for the
AP/AP_VLAN station reconfig as well.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: add MLD address to assoc BSS entries
Johannes Berg [Fri, 2 Sep 2022 14:12:50 +0000 (16:12 +0200)]
wifi: nl80211: add MLD address to assoc BSS entries

Add an MLD address attribute to BSS entries that the interface
is currently associated with to help userspace figure out what's
going on.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: refactor QoS settings code
Johannes Berg [Fri, 2 Sep 2022 14:12:46 +0000 (16:12 +0200)]
wifi: mac80211: mlme: refactor QoS settings code

Refactor the code to apply QoS settings to the driver so
we can call it on link switch.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211_hwsim: warn on invalid link address
Johannes Berg [Fri, 2 Sep 2022 14:12:39 +0000 (16:12 +0200)]
wifi: mac80211_hwsim: warn on invalid link address

Catch the bugs fixed in mac80211 by the previous commits
and warn if an invalid address is added (or removed).

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: fix double SW scan stop
Johannes Berg [Fri, 2 Sep 2022 14:12:55 +0000 (16:12 +0200)]
wifi: mac80211: fix double SW scan stop

When we stop a not-yet-started scan, we erroneously call
into the driver, causing a sequence of sw_scan_start()
followed by sw_scan_complete() twice. This will cause a
warning in hwsim with next in line commit that validates
the address passed to wmediumd/virtio. Fix this by doing
the calls only if we were actually scanning.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: mlme: assign link address correctly
Johannes Berg [Fri, 2 Sep 2022 14:12:38 +0000 (16:12 +0200)]
wifi: mac80211: mlme: assign link address correctly

Right now, we assign the link address only after we add
the link to the driver, which is quite obviously wrong.
It happens to work in many cases because it gets updated
immediately, and then link_conf updates may update it,
but it's clearly not really right.

Set the link address during ieee80211_mgd_setup_link()
so it's set before telling the driver about the link.

Fixes: 81151ce462e5 ("wifi: mac80211: support MLO authentication/association with one link")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: move link code to a new file
Johannes Berg [Fri, 2 Sep 2022 14:12:37 +0000 (16:12 +0200)]
wifi: mac80211: move link code to a new file

We probably should've done that originally, we already have
about 300 lines of code there, and will add more. Move all
the link code we wrote to a new file.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211_hwsim: refactor RX a bit
Johannes Berg [Fri, 2 Sep 2022 14:12:36 +0000 (16:12 +0200)]
wifi: mac80211_hwsim: refactor RX a bit

Refactor some common RX functionality between the netlink
and non-netlink paths, adding the special hwsim TLV (if
compiled) also in the netlink path.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211_hwsim: check STA magic in change_sta_links
Johannes Berg [Fri, 2 Sep 2022 14:12:35 +0000 (16:12 +0200)]
wifi: mac80211_hwsim: check STA magic in change_sta_links

Just as an additional check that mac80211 isn't doing
anything strange, add a check of the STA magic (which
gets assigned when the station is added, and cleared
when the station is removed).

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: remove unused arg to ieee80211_chandef_eht_oper
Johannes Berg [Fri, 2 Sep 2022 14:12:34 +0000 (16:12 +0200)]
wifi: mac80211: remove unused arg to ieee80211_chandef_eht_oper

We don't need the sdata argument, and it doesn't make any
sense for a direct conversion from one value to another,
so just remove the argument

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211_hwsim: remove multicast workaround
Johannes Berg [Fri, 2 Sep 2022 14:12:33 +0000 (16:12 +0200)]
wifi: mac80211_hwsim: remove multicast workaround

Now that we have proper multicast TX in mac80211, there's
no longer a need to fake something here.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: remove redundant err variable
Jinpeng Cui [Mon, 29 Aug 2022 11:29:53 +0000 (11:29 +0000)]
wifi: nl80211: remove redundant err variable

Return value from rdev_set_mcast_rate() directly instead of
taking this in another redundant variable.

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Jinpeng Cui <cui.jinpeng2@zte.com.cn>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: Support POWERED_ADDR_CHANGE feature
James Prestwood [Fri, 26 Aug 2022 17:00:32 +0000 (10:00 -0700)]
wifi: mac80211: Support POWERED_ADDR_CHANGE feature

Adds support in mac80211 for NL80211_EXT_FEATURE_POWERED_ADDR_CHANGE.
The motivation behind this functionality is to fix limitations of
address randomization on frequencies which are disallowed in world
roaming.

The way things work now, if a client wants to randomize their address
per-connection it must power down the device, change the MAC, and
power back up. Here lies a problem since powering down the device
may result in frequencies being disabled (until the regdom is set).
If the desired BSS is on one such frequency the client is unable to
connect once the phy is powered again.

For mac80211 based devices changing the MAC while powered is possible
but currently disallowed (-EBUSY). This patch adds some logic to
allow a MAC change while powered by removing the interface, changing
the MAC, and adding it again. mac80211 will advertise support for
this feature so userspace can determine the best course of action e.g.
disallow address randomization on certain frequencies if not
supported.

There are certain limitations put on this which simplify the logic:
 - No active connection
 - No offchannel work, including scanning.

Signed-off-by: James Prestwood <prestwoj@gmail.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: nl80211: Add POWERED_ADDR_CHANGE feature
James Prestwood [Fri, 26 Aug 2022 17:00:31 +0000 (10:00 -0700)]
wifi: nl80211: Add POWERED_ADDR_CHANGE feature

Add a new extended feature bit signifying that the wireless hardware
supports changing the MAC address while the underlying net_device is
powered. Note that this has a different meaning from
IFF_LIVE_ADDR_CHANGE as additional restrictions might be imposed by
the hardware, such as:

 - No connection is active on this interface, carrier is off
 - No scan is in progress
 - No offchannel operations are in progress

Signed-off-by: James Prestwood <prestwoj@gmail.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: prevent 4-addr use on MLDs
Johannes Berg [Fri, 2 Sep 2022 14:12:58 +0000 (16:12 +0200)]
wifi: mac80211: prevent 4-addr use on MLDs

We haven't tried this yet, and it's not very likely to
work well right now, so for now disable 4-addr use on
interfaces that are MLDs.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Link: https://lore.kernel.org/r/20220902161143.f2e4cc2efaa1.I5924e8fb44a2d098b676f5711b36bbc1b1bd68e2@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agowifi: mac80211: prevent VLANs on MLDs
Johannes Berg [Fri, 2 Sep 2022 14:12:59 +0000 (16:12 +0200)]
wifi: mac80211: prevent VLANs on MLDs

Do not allow VLANs to be added to AP interfaces that are
MLDs, this isn't going to work because the link structs
aren't propagated to the VLAN interfaces yet.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Link: https://lore.kernel.org/r/20220902161144.8c88531146e9.If2ef9a3b138d4f16ed2fda91c852da156bdf5e4d@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2 years agoMerge branch 'net_sched-redundant-resource-cleanups'
David S. Miller [Sat, 3 Sep 2022 09:40:40 +0000 (10:40 +0100)]
Merge branch 'net_sched-redundant-resource-cleanups'

Zhengchao Shao says:

====================
net: sched: remove redundant resource cleanup when init() fails

qdisc_create() calls .init() to initialize qdisc. If the initialization
fails, qdisc_create() will call .destroy() to release resources.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: sched: htb: remove redundant resource cleanup in htb_init()
Zhengchao Shao [Fri, 2 Sep 2022 08:34:30 +0000 (16:34 +0800)]
net: sched: htb: remove redundant resource cleanup in htb_init()

If htb_init() fails, qdisc_create() invokes htb_destroy() to clear
resources. Therefore, remove redundant resource cleanup in htb_init().

Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: sched: fq_codel: remove redundant resource cleanup in fq_codel_init()
Zhengchao Shao [Fri, 2 Sep 2022 08:34:29 +0000 (16:34 +0800)]
net: sched: fq_codel: remove redundant resource cleanup in fq_codel_init()

If fq_codel_init() fails, qdisc_create() invokes fq_codel_destroy() to
clear resources. Therefore, remove redundant resource cleanup in
fq_codel_init().

Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fec: add stop mode support for imx8 platform
Wei Fang [Fri, 2 Sep 2022 02:30:01 +0000 (10:30 +0800)]
net: fec: add stop mode support for imx8 platform

The current driver support stop mode by calling machine api.
The patch add dts support to set GPR register for stop request.

imx8mq enter stop/exit stop mode by setting GPR bit, which can
be accessed by A core.
imx8qm enter stop/exit stop mode by calling IMX_SC ipc APIs that
communicate with M core ipc service, and the M core set the related
GPR bit at last.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agor8152: Add MAC passthrough support for Lenovo Travel Hub
André Apitzsch [Thu, 1 Sep 2022 17:00:13 +0000 (19:00 +0200)]
r8152: Add MAC passthrough support for Lenovo Travel Hub

The Lenovo USB-C Travel Hub supports MAC passthrough.

Signed-off-by: André Apitzsch <git@apitzsch.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet/ipv4: Use __DECLARE_FLEX_ARRAY() helper
Gustavo A. R. Silva [Wed, 31 Aug 2022 19:12:42 +0000 (14:12 -0500)]
net/ipv4: Use __DECLARE_FLEX_ARRAY() helper

We now have a cleaner way to keep compatibility with user-space
(a.k.a. not breaking it) when we need to keep in place a one-element
array (for its use in user-space) together with a flexible-array
member (for its use in kernel-space) without making it hard to read
at the source level. This is through the use of the new
__DECLARE_FLEX_ARRAY() helper macro.

The size and memory layout of the structure is preserved after the
changes. See below.

Before changes:

$ pahole -C ip_msfilter net/ipv4/igmp.o
struct ip_msfilter {
union {
struct {
__be32     imsf_multiaddr_aux;   /*     0     4 */
__be32     imsf_interface_aux;   /*     4     4 */
__u32      imsf_fmode_aux;       /*     8     4 */
__u32      imsf_numsrc_aux;      /*    12     4 */
__be32     imsf_slist[1];        /*    16     4 */
};                                       /*     0    20 */
struct {
__be32     imsf_multiaddr;       /*     0     4 */
__be32     imsf_interface;       /*     4     4 */
__u32      imsf_fmode;           /*     8     4 */
__u32      imsf_numsrc;          /*    12     4 */
__be32     imsf_slist_flex[0];   /*    16     0 */
};                                       /*     0    16 */
};                                               /*     0    20 */

/* size: 20, cachelines: 1, members: 1 */
/* last cacheline: 20 bytes */
};

After changes:

$ pahole -C ip_msfilter net/ipv4/igmp.o
struct ip_msfilter {
__be32                     imsf_multiaddr;       /*     0     4 */
__be32                     imsf_interface;       /*     4     4 */
__u32                      imsf_fmode;           /*     8     4 */
__u32                      imsf_numsrc;          /*    12     4 */
union {
__be32             imsf_slist[1];        /*    16     4 */
struct {
struct {
} __empty_imsf_slist_flex;       /*    16     0 */
__be32     imsf_slist_flex[0];   /*    16     0 */
};                                       /*    16     0 */
};                                               /*    16     4 */

/* size: 20, cachelines: 1, members: 5 */
/* last cacheline: 20 bytes */
};

In the past, we had to duplicate the whole original structure within
a union, and update the names of all the members. Now, we just need to
declare the flexible-array member to be used in kernel-space through
the __DECLARE_FLEX_ARRAY() helper together with the one-element array,
within a union. This makes the source code more clean and easier to read.

Link: https://github.com/KSPP/linux/issues/193
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet/sched: cls_api: remove redundant 0 check in tcf_qevent_init()
Zhengchao Shao [Thu, 1 Sep 2022 01:16:17 +0000 (09:16 +0800)]
net/sched: cls_api: remove redundant 0 check in tcf_qevent_init()

tcf_qevent_parse_block_index() never returns a zero block_index.
Therefore, it is unnecessary to check the value of block_index in
tcf_qevent_init().

Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Link: https://lore.kernel.org/r/20220901011617.14105-1-shaozhengchao@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: lantiq_etop: Fix return type for implementation of ndo_start_xmit
GUO Zihua [Fri, 2 Sep 2022 08:15:21 +0000 (16:15 +0800)]
net: lantiq_etop: Fix return type for implementation of ndo_start_xmit

Since Linux now supports CFI, it will be a good idea to fix mismatched
return type for implementation of hooks. Otherwise this might get
cought out by CFI and cause a panic.

ltq_etop_tx() would return either NETDEV_TX_BUSY or NETDEV_TX_OK, so
change the return type to netdev_tx_t directly.

Signed-off-by: GUO Zihua <guozihua@huawei.com>
Link: https://lore.kernel.org/r/20220902081521.59867-1-guozihua@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: sunplus: Fix return type for implementation of ndo_start_xmit
GUO Zihua [Fri, 2 Sep 2022 08:15:50 +0000 (16:15 +0800)]
net: sunplus: Fix return type for implementation of ndo_start_xmit

Since Linux now supports CFI, it will be a good idea to fix mismatched
return type for implementation of hooks. Otherwise this might get
cought out by CFI and cause a panic.

spl2sw_ethernet_start_xmit() would return either NETDEV_TX_BUSY or
NETDEV_TX_OK, so change the return type to netdev_tx_t directly.

Signed-off-by: GUO Zihua <guozihua@huawei.com>
Link: https://lore.kernel.org/r/20220902081550.60095-1-guozihua@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: xscale: Fix return type for implementation of ndo_start_xmit
GUO Zihua [Fri, 2 Sep 2022 08:16:12 +0000 (16:16 +0800)]
net: xscale: Fix return type for implementation of ndo_start_xmit

Since Linux now supports CFI, it will be a good idea to fix mismatched
return type for implementation of hooks. Otherwise this might get
cought out by CFI and cause a panic.

eth_xmit() would return either NETDEV_TX_BUSY or NETDEV_TX_OK, so
change the return type to netdev_tx_t directly.

Signed-off-by: GUO Zihua <guozihua@huawei.com>
Link: https://lore.kernel.org/r/20220902081612.60405-1-guozihua@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: broadcom: Fix return type for implementation of
GUO Zihua [Fri, 2 Sep 2022 07:54:07 +0000 (15:54 +0800)]
net: broadcom: Fix return type for implementation of

Since Linux now supports CFI, it will be a good idea to fix mismatched
return type for implementation of hooks. Otherwise this might get
cought out by CFI and cause a panic.

bcm4908_enet_start_xmit() would return either NETDEV_TX_BUSY or
NETDEV_TX_OK, so change the return type to netdev_tx_t directly.

Signed-off-by: GUO Zihua <guozihua@huawei.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20220902075407.52358-1-guozihua@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: ieee802154: Fix compilation error when CONFIG_IEEE802154_NL802154_EXPERIMENTAL...
Gal Pressman [Fri, 2 Sep 2022 03:06:20 +0000 (20:06 -0700)]
net: ieee802154: Fix compilation error when CONFIG_IEEE802154_NL802154_EXPERIMENTAL is disabled

When CONFIG_IEEE802154_NL802154_EXPERIMENTAL is disabled,
NL802154_CMD_DEL_SEC_LEVEL is undefined and results in a compilation
error:
net/ieee802154/nl802154.c:2503:19: error: 'NL802154_CMD_DEL_SEC_LEVEL' undeclared here (not in a function); did you mean 'NL802154_CMD_SET_CCA_ED_LEVEL'?
 2503 |  .resv_start_op = NL802154_CMD_DEL_SEC_LEVEL + 1,
      |                   ^~~~~~~~~~~~~~~~~~~~~~~~~~
      |                   NL802154_CMD_SET_CCA_ED_LEVEL

Unhide the experimental commands, having them defined in an enum
makes no difference.

Fixes: 9c5d03d36251 ("genetlink: start to validate reserved header bytes")
Signed-off-by: Gal Pressman <gal@nvidia.com>
Acked-by: Stefan Schmidt <stefan@datenfreihafen.org>
Tested-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Link: https://lore.kernel.org/r/20220902030620.2737091-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: remove netif_tx_napi_add()
Jakub Kicinski [Thu, 1 Sep 2022 00:00:58 +0000 (17:00 -0700)]
net: remove netif_tx_napi_add()

All callers are now gone.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: bql: add more documentation
Eric Dumazet [Wed, 31 Aug 2022 18:44:27 +0000 (18:44 +0000)]
net: bql: add more documentation

Add some documentation for netdev_tx_sent_queue() and
netdev_tx_completed_queue()

Stating that netdev_tx_completed_queue() must be called once
per TX completion round is apparently not obvious for everybody.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'net-ipa-transaction-state-IDs'
David S. Miller [Fri, 2 Sep 2022 11:08:44 +0000 (12:08 +0100)]
Merge branch 'net-ipa-transaction-state-IDs'

Alex Elder says:

====================
net: ipa: use IDs to track transaction state

This series is the first of three groups of changes that simplify
the way the IPA driver tracks the state of its transactions.

Each GSI channel has a fixed number of transactions allocated at
initialization time.  The number allocated matches the number of
TREs in the transfer ring associated with the channel.  This is
because the transfer ring limits the number of transfers that can
ever be underway, and in the worst case, each transaction represents
a single TRE.

Transactions go through various states during their lifetime.
Currently a set of lists keeps track of which transactions are in
each state.  Initially, all transactions are free.  An allocated
transaction is placed on the allocated list.  Once an allocated
transaction is committed, it is moved from the allocated to the
committed list.  When a committed transaction is sent to hardware
(via a doorbell) it is moved to the pending list.  When hardware
signals that some work has completed, transactions are moved to the
completed list.  Finally, when a completed transaction is polled
it's moved to the polled list before being removed when it becomes
free.

Changing a transaction's state thus normally involves manipulating
two lists, and to prevent corruption a spinlock is held while the
lists are updated.

Transactions move through their states in a well-defined sequence
though, and they do so strictly in order.  So transaction 0 is
always allocated before transaction 1; transaction 0 is always
committed before transaction 1; and so on, through completion,
polling, and becoming free.  Because of this, it's sufficient to
just keep track of which transaction is the first in each state.
The rest of the transactions in a given state can be derived from
the first transaction in an "adjacent" state.  As a result, we can
track the state of all transactions with a set of indexes, and can
update these without the need for a spinlock.

This first group of patches just defines the set of indexes that
will be used for this new way of tracking transaction state.  Two
more groups of patches will follow.  I've broken the 17 patches into
these three groups to facilitate review.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: track polled transactions with an ID
Alex Elder [Wed, 31 Aug 2022 22:40:17 +0000 (17:40 -0500)]
net: ipa: track polled transactions with an ID

Add a transaction ID to track the first element in the transaction
array that has been polled.  Advance the ID when we are releasing a
transaction.

Temporarily add warnings that verify that the first polled
transaction tracked by the ID matches the first element on the
polled list, both when polling and freeing.

Remove the temporary warnings added by the previous commit.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: track completed transactions with an ID
Alex Elder [Wed, 31 Aug 2022 22:40:16 +0000 (17:40 -0500)]
net: ipa: track completed transactions with an ID

Add a transaction ID field to track the first element in the
transaction array that has completed but has not yet been polled.

Advance the ID when we are processing a transaction in the NAPI
polling loop (where completed transactions become polled).

Temporarily add warnings that verify that the first completed
transaction tracked by the ID matches the first element on the
completed list, both when pending and completing.

Remove the temporary warnings added by the previous commit.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: track pending transactions with an ID
Alex Elder [Wed, 31 Aug 2022 22:40:15 +0000 (17:40 -0500)]
net: ipa: track pending transactions with an ID

Add a transaction ID field to track the first element in the
transaction array that is pending (sent to hardware) but not yet
complete.  Advance the ID when a completion event for a channel
indicates that transactions have completed.

Temporarily add warnings that verify that the first pending
transaction tracked by the ID matches the first element on the
pending list, both when pending and completing, as well as when
resetting the channel.

Remove the temporary warnings added by the previous commit.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: track committed transactions with an ID
Alex Elder [Wed, 31 Aug 2022 22:40:14 +0000 (17:40 -0500)]
net: ipa: track committed transactions with an ID

Add a transaction ID field to track the first element in a channel's
transaction array that has been committed, but not yet passed to the
hardware.  Advance the ID when the hardware is notified via doorbell
that TREs from a transaction are ready for consumption.

Temporarily add warnings that verify that the first committed
transaction tracked by the ID matches the first element on the
committed list, both when committing and pending (at doorbell).

Remove the temporary warnings added by the previous commit.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: track allocated transactions with an ID
Alex Elder [Wed, 31 Aug 2022 22:40:13 +0000 (17:40 -0500)]
net: ipa: track allocated transactions with an ID

Transactions for a channel are now managed in an array, with a free
transaction ID indicating which is the next one free.

Add another transaction ID field to track the first element in the
array that has been allocated.  Advance it when a transaction is
committed (because that is when that transaction leaves allocated
state).

Temporarily add warnings that verify that the first allocated
transaction tracked by the ID matches the first element on the
allocated list, both when allocating and committing a transaction.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ipa: use an array for transactions
Alex Elder [Wed, 31 Aug 2022 22:40:12 +0000 (17:40 -0500)]
net: ipa: use an array for transactions

Transactions are always allocated one at a time.  The maximum number
of them we could ever need occurs if each TRE is assigned to a
transaction.  So a channel requires no more transactions than the
number of TREs in its transfer ring.  That number is known to be a
power-of-2 less than 65536.

The transaction pool abstraction is used for other things, but for
transactions we can use a simple array of transaction structures,
and use a free index to indicate which entry in the array is the
next one free for allocation.

By having the number of elements in the array be a power-of-2, we
can use an ever-incrementing 16-bit free index, and use it modulo
the array size.  Distinguish a "trans_id" (whose value can exceed
the number of entries in the transaction array) from a "trans_index"
(which is less than the number of entries).

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'lan966x-make-reset-optional'
David S. Miller [Fri, 2 Sep 2022 10:37:27 +0000 (11:37 +0100)]
Merge branch 'lan966x-make-reset-optional'

Michael Walle says:

====================
net: lan966x: make reset optional

This is the remaining part of the reset rework on the LAN966x targetting
the netdev tree.

The former series can be found at:
https://lore.kernel.org/lkml/20220826115607.1148489-1-michael@walle.cc/
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: lan966x: make reset optional
Michael Walle [Wed, 31 Aug 2022 11:18:55 +0000 (13:18 +0200)]
net: lan966x: make reset optional

There is no dedicated reset for just the switch core. The reset which
is used up until now, is more of a global reset, resetting almost the
whole SoC and cause spurious errors by doing so. Make it possible to
handle the reset elsewhere and make the reset optional.

Signed-off-by: Michael Walle <michael@walle.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodt-bindings: net: sparx5: don't require a reset line
Michael Walle [Wed, 31 Aug 2022 11:18:54 +0000 (13:18 +0200)]
dt-bindings: net: sparx5: don't require a reset line

Make the reset line optional. It turns out, there is no dedicated reset
for the switch. Instead, the reset which was used up until now, was kind
of a global reset. This is now handled elsewhere, thus don't require a
reset.

Signed-off-by: Michael Walle <michael@walle.cc>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agowifi: move from strlcpy with unused retval to strscpy
Wolfram Sang [Tue, 30 Aug 2022 20:14:53 +0000 (22:14 +0200)]
wifi: move from strlcpy with unused retval to strscpy

Follow the advice of the below link and prefer 'strscpy' in this
subsystem. Conversion is 1:1 because the return value is not used.
Generated by a coccinelle script.

Link: https://lore.kernel.org/r/CAHk-=wgfRnXz0W3D37d01q3JFkr_i_uTL=V6A6G1oUZcprmknw@mail.gmail.com/
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220830201457.7984-2-wsa+renesas@sang-engineering.com
2 years agowifi: wilc1000: remove redundant ret variable
Jinpeng Cui [Tue, 30 Aug 2022 10:55:05 +0000 (10:55 +0000)]
wifi: wilc1000: remove redundant ret variable

Return value from cfg80211_rx_mgmt() directly instead of
taking this in another redundant variable.

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Jinpeng Cui <cui.jinpeng2@zte.com.cn>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220830105505.287564-1-cui.jinpeng2@zte.com.cn
2 years agowifi: rtw88: add missing destroy_workqueue() on error path in rtw_core_init()
Yang Yingliang [Fri, 26 Aug 2022 02:38:17 +0000 (10:38 +0800)]
wifi: rtw88: add missing destroy_workqueue() on error path in rtw_core_init()

Add the missing destroy_workqueue() before return from rtw_core_init()
in error path.

Fixes: fe101716c7c9 ("rtw88: replace tx tasklet with work queue")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Reviewed-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220826023817.3908255-1-yangyingliang@huawei.com
2 years agowifi: wfx: prevent underflow in wfx_send_pds()
Dan Carpenter [Fri, 19 Aug 2022 05:23:43 +0000 (08:23 +0300)]
wifi: wfx: prevent underflow in wfx_send_pds()

This does a "chunk_len - 4" subtraction later when it calls:

ret = wfx_hif_configuration(wdev, buf + 4, chunk_len - 4);

so check for "chunk_len" is less than 4.

Fixes: dcbecb497908 ("staging: wfx: allow new PDS format")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/Yv8eX7Xv2ubUOvW7@kili
2 years agowifi: rtl8xxxu: tighten bounds checking in rtl8xxxu_read_efuse()
Dan Carpenter [Fri, 19 Aug 2022 05:22:32 +0000 (08:22 +0300)]
wifi: rtl8xxxu: tighten bounds checking in rtl8xxxu_read_efuse()

There some bounds checking to ensure that "map_addr" is not out of
bounds before the start of the loop.  But the checking needs to be
done as we iterate through the loop because "map_addr" gets larger as
we iterate.

Fixes: 26f1fad29ad9 ("New driver: rtl8xxxu (mac80211)")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Jes Sorensen <Jes.Sorensen@gmail.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/Yv8eGLdBslLAk3Ct@kili
2 years agowifi: rtw89: declare to support beamformee above bandwidth 80MHz
Ping-Ke Shih [Fri, 26 Aug 2022 06:10:11 +0000 (14:10 +0800)]
wifi: rtw89: declare to support beamformee above bandwidth 80MHz

Declare this to tell AP we can support beamformee over bandwidth 160M,
and then yield better performance in field.

Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220826061011.9037-3-pkshih@realtek.com
2 years agowifi: rtw89: correct polling address of address CAM
Ping-Ke Shih [Fri, 26 Aug 2022 06:10:10 +0000 (14:10 +0800)]
wifi: rtw89: correct polling address of address CAM

Writing address to kick hardware to initialize address CAM, and then poll
ready bit to determine completed. Old wrong code poll wrong register
address, so it can lead error and fail to bring up interface.

Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220826061011.9037-2-pkshih@realtek.com
2 years agowifi: rtw89: no HTC field if TX rate might fallback to legacy
Ping-Ke Shih [Fri, 26 Aug 2022 06:10:09 +0000 (14:10 +0800)]
wifi: rtw89: no HTC field if TX rate might fallback to legacy

Packets containing HTC field with legacy rate could be dropped by AP. If
TX rate of report is lower than MCS2, hardware might fall back rate to
legacy. Therefore, add a checking rule to avoid HTC field in this
situation.

Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220826061011.9037-1-pkshih@realtek.com
2 years agowifi: rtw89: pci: correct TX resource checking in low power mode
Ping-Ke Shih [Wed, 24 Aug 2022 06:33:12 +0000 (14:33 +0800)]
wifi: rtw89: pci: correct TX resource checking in low power mode

Number of TX resource must be minimum of TX_BD and TX_WD. Only considering
TX_BD could drop TX packets pulled from mac80211 if TX_WD is unavailable.

Fixes: 52edbb9fb78a ("rtw89: ps: access TX/RX rings via another registers in low power mode")
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220824063312.15784-2-pkshih@realtek.com
2 years agowifi: rtw89: pci: fix interrupt stuck after leaving low power mode
Ping-Ke Shih [Wed, 24 Aug 2022 06:33:11 +0000 (14:33 +0800)]
wifi: rtw89: pci: fix interrupt stuck after leaving low power mode

We turn off interrupt in ISR, and re-enable interrupt in threadfn or
napi_poll according to the mode it stays. If we are turning off interrupt,
rtwpci->running flag is unset and interrupt handler stop processing even
if it was called, so disallow to re-enable interrupt in this situation.
Or, wifi chip doesn't trigger interrupt events anymore because interrupt
status (ISR) isn't clear by interrupt handler anymore.

Fixes: c83dcd0508e2 ("rtw89: pci: add a separate interrupt handler for low power mode")
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220824063312.15784-1-pkshih@realtek.com
2 years agowifi: rtw89: enlarge the CFO tracking boundary
Cheng-Chieh Hsieh [Wed, 24 Aug 2022 06:14:25 +0000 (14:14 +0800)]
wifi: rtw89: enlarge the CFO tracking boundary

The calibration value of XTAL offset may be too large in some wifi
modules, that the CFO tracking mechanism under the existing tracking
boundary can not adjust the CFO to the tolerable range. So we enlarge it.

Signed-off-by: Cheng-Chieh Hsieh <cj.hsieh@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220824061425.13764-1-pkshih@realtek.com
2 years agowifi: rtw89: pci: correct suspend/resume setting for variant chips
Chin-Yen Lee [Fri, 19 Aug 2022 06:48:11 +0000 (14:48 +0800)]
wifi: rtw89: pci: correct suspend/resume setting for variant chips

We find that suspend/resume tests cause 8852CE lost, because some pci
registers are changed for 8852CE. So, correct them accordingly.

Signed-off-by: Chin-Yen Lee <timlee@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220819064811.37700-6-pkshih@realtek.com
2 years agowifi: rtw89: pci: enable CLK_REQ, ASPM, L1 and L1ss for 8852c
Chin-Yen Lee [Fri, 19 Aug 2022 06:48:10 +0000 (14:48 +0800)]
wifi: rtw89: pci: enable CLK_REQ, ASPM, L1 and L1ss for 8852c

8852CE controls CLKREQ, ASPM L1, L1ss via wifi registers
instead, so change them accordingly.

Signed-off-by: Chin-Yen Lee <timlee@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220819064811.37700-5-pkshih@realtek.com
2 years agowifi: rtw89: pci: fix PCI PHY auto adaption by using software restore
Chia-Yuan Li [Fri, 19 Aug 2022 06:48:09 +0000 (14:48 +0800)]
wifi: rtw89: pci: fix PCI PHY auto adaption by using software restore

There is chance that PCI PHY auto adaption fail. When first time boot up,
software restore the right adaption value and close PHY auto adaption
mechanism.

Signed-off-by: Chia-Yuan Li <leo.li@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220819064811.37700-4-pkshih@realtek.com
2 years agowifi: rtw89: 8852c: set TBTT shift configuration
Chia-Yuan Li [Fri, 19 Aug 2022 06:48:08 +0000 (14:48 +0800)]
wifi: rtw89: 8852c: set TBTT shift configuration

It is found that 8852ce loses some beacon after
enabling deep ps mode. We set TBTT shift to wake up
firmware early to open RF/BB for receiving beacon in time.

Signed-off-by: Chia-Yuan Li <leo.li@realtek.com>
Signed-off-by: Po-Hao Huang <phhuang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220819064811.37700-3-pkshih@realtek.com
2 years agowifi: rtw89: add retry to change power_mode state
Chin-Yen Lee [Fri, 19 Aug 2022 06:48:07 +0000 (14:48 +0800)]
wifi: rtw89: add retry to change power_mode state

When starting to send heavy traffic in low power mode,
driver will call multiple tx wake notify to wake firmware
within a short time. In this situation, firmware may miss
power mode change request from driver and leads to status
error. So we change driver to call power_mode_change at most
three times to make sure firmware could get the request.

Signed-off-by: Chin-Yen Lee <timlee@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220819064811.37700-2-pkshih@realtek.com
2 years agowifi: rtw89: correct BA CAM allocation
Ping-Ke Shih [Tue, 16 Aug 2022 01:32:47 +0000 (09:32 +0800)]
wifi: rtw89: correct BA CAM allocation

BA CAM entries are global resource of hardware, so move the bitmap and
instances to rtw89_cam_info, and then use link list from rtw89_sta to
these instances.

To check the allocation, add ba_cam to debugfs:

  map:
   mac_id:    01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
   addr_cam:  01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
   bssid_cam: 01 00 00 00 00 00 00 00
   sec_cam:   00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
   ba_cam:    03 00 00 00 00 00 00 00
  VIF [0] 94:08:53:8e:ef:21
   bssid_cam_idx=0
   addr_cam_idx=0
   -> bssid_cam_idx=0
   sec_cam_bitmap=00 00 00 00 00 00 00 00
  STA [0] 38:78:62:8b:cb:c6
   addr_cam_idx=0
   -> bssid_cam_idx=0
   sec_cam_bitmap=00 00 00 00 00 00 00 00
   ba_cam tid[6]=0, tid[1]=1

Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220816013247.6243-4-pkshih@realtek.com
2 years agowifi: rtw89: 8852c: initialize and correct BA CAM content
Ping-Ke Shih [Tue, 16 Aug 2022 01:32:46 +0000 (09:32 +0800)]
wifi: rtw89: 8852c: initialize and correct BA CAM content

The bacam_v1 must do additional initialization, and H2C content of BA CAM
is also different. So, correct them accordingly.

Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220816013247.6243-3-pkshih@realtek.com
2 years agowifi: rtw89: 8852c: declare correct BA CAM number
Ping-Ke Shih [Tue, 16 Aug 2022 01:32:45 +0000 (09:32 +0800)]
wifi: rtw89: 8852c: declare correct BA CAM number

8852A has 2 BA CAM entries, but 8852C has 8 entries. Add a field to
discriminate their differences.

Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220816013247.6243-2-pkshih@realtek.com
2 years agowifi: rtw89: 8852c: update TX power tables to R49
Zong-Zhe Yang [Fri, 12 Aug 2022 09:31:16 +0000 (17:31 +0800)]
wifi: rtw89: 8852c: update TX power tables to R49

TX power byrate:
doesn't change

TX power limit:
configure values for KCC and UK
refine a bit configuration vales

TX power limit_ru:
configure values for KCC and UK
refine a bit configuration values
change 6GHz to follow indoor setting
configure 6GHz values for ETSI

TX power shape:
change with TX power limit/limit_ru

Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220812093116.56791-4-pkshih@realtek.com
2 years agowifi: rtw89: 8852c: update RF radio A/B parameters to R49
Ping-Ke Shih [Fri, 12 Aug 2022 09:31:15 +0000 (17:31 +0800)]
wifi: rtw89: 8852c: update RF radio A/B parameters to R49

Update 8852c radio A/B parameters from internal HALRF_027_00_071 R49.

Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220812093116.56791-3-pkshih@realtek.com
2 years agowifi: rtw89: TX power limit/limit_ru consider negative
Zong-Zhe Yang [Fri, 12 Aug 2022 09:31:14 +0000 (17:31 +0800)]
wifi: rtw89: TX power limit/limit_ru consider negative

Some chips' RF TX power limit/limit_ru tables start to configure some
negative values. Fix the setting logic to prevent negative values from
polluting fields of others.

Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220812093116.56791-2-pkshih@realtek.com
2 years agortw89: coex: Update Wi-Fi driver/firmware TDMA cycle report for RTL8852c
Ching-Te Ku [Mon, 25 Jul 2022 02:35:09 +0000 (10:35 +0800)]
rtw89: coex: Update Wi-Fi driver/firmware TDMA cycle report for RTL8852c

Because RTL8852c firmware handshake use different structure definition
with RTL8852a, so it's necessary to update a version for RTL8852c.

Signed-off-by: Ching-Te Ku <ku920601@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220725023509.43114-10-pkshih@realtek.com
2 years agortw89: coex: Add v1 Wi-Fi SCC coexistence policy
Ching-Te Ku [Mon, 25 Jul 2022 02:35:08 +0000 (10:35 +0800)]
rtw89: coex: Add v1 Wi-Fi SCC coexistence policy

Because the later firmware had patched some new feature, it can control
the Wi-Fi/BT slots more efficiently. This patch enhance it for better
Wi-Fi SCC mode performance.

Signed-off-by: Ching-Te Ku <ku920601@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220725023509.43114-9-pkshih@realtek.com
2 years agortw89: coex: Move _set_policy to chip_ops
Ching-Te Ku [Mon, 25 Jul 2022 02:35:07 +0000 (10:35 +0800)]
rtw89: coex: Move _set_policy to chip_ops

Due to the difference of Wi-Fi firmware supported feature, RTL8852C
need to defined more policy to enable the features.
(Ex: DBCC, Wi-Fi multi-role, TDMA instant and so on)

Signed-off-by: Ching-Te Ku <ku920601@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220725023509.43114-8-pkshih@realtek.com
2 years agortw89: coex: update WL role info v1 for RTL8852C branch using
Ching-Te Ku [Mon, 25 Jul 2022 02:35:06 +0000 (10:35 +0800)]
rtw89: coex: update WL role info v1 for RTL8852C branch using

The H2C format and support feature are different. The newer Wi-Fi firmware
and driver branch need to handshake more information like DBCC or P2P
connection info.

Signed-off-by: Ching-Te Ku <ku920601@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220725023509.43114-7-pkshih@realtek.com