The sja1105 driver will gain support for the next-gen SJA1110 switch,
which is very similar except for the fact it has more than 5 ports.

So we need to replace the hardcoded SJA1105_NUM_PORTS in this driver
with ds->num_ports. This patch is as mechanical as possible (save for
the fact that ds->num_ports is not an integer constant expression).
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The current internal sja1105 driver API is optimized for retrieving many
statistics counters at once. But the switch does not do atomic snapshotting
for them anyway.

In case we start reporting the hardware port counters through
ndo_get_stats64 as well, not just ethtool, it would be good to be able
to read individual port counters and not all of them.

Additionally, since Arnd Bergmann's commit ae1804de ("dsa: sja1105:
dynamically allocate stats structure"), sja1105_get_ethtool_stats
allocates memory dynamically, since struct sja1105_port_status was
deemed to consume too much stack memory. That is not ideal.
The large structure is only needed because of the burst read.
If we read statistics one by one, we can consume less memory, and
we can avoid dynamic allocation.

Additionally, latency-sensitive interfaces such as PTP operations (for
phc2sys) might suffer if the SPI mutex is being held for too long, which
happens in the case of SPI burst reads. By reading counters one by one,
we give a chance for higher priority processes to preempt and take the
SPI bus mutex for accessing the PTP clock.
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The queue levels are not counters, but instead they represent the
occupancy of the MAC TX queues. Having these in ethtool port counters is
not helpful, so remove them.
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The static config of the sja1105 switch is a long stream of bytes which
is programmed to the hardware in chunks (portions with the chip select
continuously asserted) of max 256 bytes each. Each chunk is a
spi_message composed of 2 spi_transfers: the buffer with the data and a
preceding buffer with the SPI access header.

Only that certain SPI controllers, such as the spi-sc18is602 I2C-to-SPI
bridge, cannot keep the chip select asserted for that long.
The spi_max_transfer_size() and spi_max_message_size() functions are how
the controller can impose its hardware limitations upon the SPI
peripheral driver.

For the sja1105 driver to work with these controllers, both buffers must
be smaller than the transfer limit, and their sum must be smaller than
the message limit.

Regression-tested on a switch connected to a controller with no
limitations (spi-fsl-dspi) as well as with one with caps for both
max_transfer_size and max_message_size (spi-sc18is602).
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The sja1105 driver has been described by Mark Brown as "not using the
[ SPI ] API at all idiomatically" due to the use of cs_change:
https://patchwork.kernel.org/project/netdevbpf/patch/20210520135031.2969183-1-olteanv@gmail.com/

According to include/linux/spi/spi.h, the chip select is supposed to be
asserted for the entire length of a SPI message, as long as cs_change is
false for all member transfers. The cs_change flag changes the following:

(i) When a non-final SPI transfer has cs_change = true, the chip select
    should temporarily deassert and then reassert starting with the next
    transfer.
(ii) When a final SPI transfer has cs_change = true, the chip select
     should remain asserted until the following SPI message.

The sja1105 driver only uses cs_change for its first property, to form a
single SPI message whose layout can be seen below:

                                             this is an entire, single spi_message
           _______________________________________________________________________________________________
          /                                                                                               \
          +-------------+---------------+-------------+---------------+ ... +-------------+---------------+
          | hdr_xfer[0] | chunk_xfer[0] | hdr_xfer[1] | chunk_xfer[1] |     | hdr_xfer[n] | chunk_xfer[n] |
          +-------------+---------------+-------------+---------------+ ... +-------------+---------------+
cs_change      false          true           false           true                false          false

           ____________________________  _____________________________       _____________________________
CS line __/                            \/                             \ ... /                             \__

The fact of the matter is that spi_max_message_size() has an ambiguous
meaning if any non-final transfer has cs_change = true.

If the SPI master has a limitation in that it cannot keep the chip
select asserted for more than, say, 200 bytes (like the spi-sc18is602),
the normal thing for it to do is to implement .max_transfer_size and
.max_message_size, and limit both to 200: in the "worst case" where
cs_change is always false, then the controller can, indeed, not send
messages larger than 200 bytes.

But the fact that the SPI controller's max_message_size does not
necessarily mean that we cannot send messages larger than that.
Notably, if the SPI master special-cases the transfers with cs_change
and treats every chip select toggling as an entirely new transaction,
then a SPI message can easily exceed that limit. So there is a
temptation to ignore the controller's reported max_message_size when
using cs_change = true in non-final transfers.

But that can lead to false conclusions. As Mark points out, the SPI
controller might have a different kind of limitation with the max
message size, that has nothing at all to do with how long it can keep
the chip select asserted.
For example, that might be the case if the device is able to offload the
chip select changes to the hardware as part of the data stream, and it
packs the entire stream of commands+data (corresponding to a SPI
message) into a single DMA transfer that is itself limited in size.

So the only thing we can do is avoid ambiguity by not using cs_change at
all. Instead of sending a single spi_message, we now send multiple SPI
messages as follows:

                  spi_message 0                 spi_message 1                       spi_message n
           ____________________________   ___________________________        _____________________________
          /                            \ /                           \      /                             \
          +-------------+---------------+-------------+---------------+ ... +-------------+---------------+
          | hdr_xfer[0] | chunk_xfer[0] | hdr_xfer[1] | chunk_xfer[1] |     | hdr_xfer[n] | chunk_xfer[n] |
          +-------------+---------------+-------------+---------------+ ... +-------------+---------------+
cs_change      false          true           false           true                false          false

           ____________________________  _____________________________       _____________________________
CS line __/                            \/                             \ ... /                             \__

which is clearer because the max_message_size limit is now easier to
enforce. What is transmitted on the wire stays, of course, the same.

Additionally, because we send no more than 2 transfers at a time, we now
avoid dynamic memory allocation too, which might be seen as an
improvement by some.
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The chip can configure unicast flooding, broadcast flooding and learning.
Learning is per port, while flooding is per {ingress, egress} port pair
and we need to configure the same value for all possible ingress ports
towards the requested one.

While multicast flooding is not officially supported, we can hack it by
using a feature of the second generation (P/Q/R/S) devices, which is that
FDB entries are maskable, and multicast addresses always have an odd
first octet. So by putting a match-all for 00:01:00:00:00:00 addr and
00:01:00:00:00:00 mask at the end of the FDB, we make sure that it is
always checked last, and does not take precedence in front of any other
MDB. So it behaves effectively as an unknown multicast entry.

For the first generation switches, this feature is not available, so
unknown multicast will always be treated the same as unknown unicast.
So the only thing we can do is request the user to offload the settings
for these 2 flags in tandem, i.e.

ip link set swp2 type bridge_slave flood off
Error: sja1105: This chip cannot configure multicast flooding independently of unicast.
ip link set swp2 type bridge_slave flood off mcast_flood off
ip link set swp2 type bridge_slave mcast_flood on
Error: sja1105: This chip cannot configure multicast flooding independently of unicast.
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As explained in Documentation/networking/dsa/sja1105.rst, this switch
has a static config held in the driver's memory and re-uploaded from
time to time into the device (after any major change).

The format of this static config is in fact described in UM10944.pdf and
it contains all the switch's settings (it also contains device ID, table
CRCs, etc, just like in the manual). So it is a useful and universal
devlink region to expose to user space, for debugging purposes.
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since struct sja1105_private only holds a const pointer to one of these
structures based on device tree compatible string, the structures
themselves can be made const.

Also add an empty line between each structure definition, to appease
checkpatch.
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

SJA1105, being AVB/TSN switches, provide hardware assist for the
Credit-Based Shaper as described in the IEEE 8021Q-2018 document.

First generation has 10 shapers, freely assignable to any of the 4
external ports and 8 traffic classes, and second generation has 16
shapers.

The Credit-Based Shaper tables are accessed through the dynamic
reconfiguration interface, so we have to restore them manually after a
switch reset. The tables are backed up by the static config only on
P/Q/R/S, and we don't want to add custom code only for that family,
since the procedure that is in place now works for both.

Tested with the following commands:

data_rate_kbps=67000
port_transmit_rate_kbps=1000000
idleslope=$data_rate_kbps
sendslope=$(($idleslope - $port_transmit_rate_kbps))
locredit=$((-0x80000000))
hicredit=$((0x7fffffff))
tc qdisc add dev swp2 root handle 1: mqprio hw 0 num_tc 8 \
        map 0 1 2 3 4 5 6 7 \
        queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7
tc qdisc replace dev swp2 parent 1:1 cbs \
        idleslope $idleslope \
        sendslope $sendslope \
        hicredit $hicredit \
        locredit $locredit \
        offload 1
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In VLAN-unaware mode, sja1105 uses VLAN tags with a custom TPID of
0xdadb. While in the yet-to-be introduced best_effort_vlan_filtering
mode, it needs to work with normal VLAN TPID values.

A complication arises when we must transmit a VLAN-tagged packet to the
switch when it's in VLAN-aware mode. We need to construct a packet with
2 VLAN tags, and the switch will use the outer header for routing and
pop it on egress. But sadly, here the 2 hardware generations don't
behave the same:

- E/T switches won't pop an ETH_P_8021AD tag on egress, it seems
  (packets will remain double-tagged).
- P/Q/R/S switches will drop a packet with 2 ETH_P_8021Q tags (it looks
  like it tries to prevent VLAN hopping).

But looks like the reverse is also true:

- E/T switches have no problem popping the outer tag from packets with
  2 ETH_P_8021Q tags.
- P/Q/R/S will have no problem popping a single tag even if that is
  ETH_P_8021AD.

So it is clear that if we want the hardware to work with dsa_8021q
tagging in VLAN-aware mode, we need to send different TPIDs depending on
revision. Keep that information in priv->info->qinq_tpid.

The per-port tagger structure will hold an xmit_tpid value that depends
not only upon the qinq_tpid, but also upon the VLAN awareness state
itself (in case we must transmit using 0xdadb).
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: NFlorian Fainelli <f.fainelli@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Restrict the TTEthernet hardware support on this switch to operate as
closely as possible to IEEE 802.1Qci as possible. This means that it can
perform PTP-time-based ingress admission control on streams identified
by {DMAC, VID, PCP}, which is useful when trying to ensure the
determinism of traffic scheduled via IEEE 802.1Qbv.

The oddity comes from the fact that in hardware (and in TTEthernet at
large), virtual links always need a full-blown action, including not
only the type of policing, but also the list of destination ports. So in
practice, a single tc-gate action will result in all packets getting
dropped. Additional actions (either "trap" or "redirect") need to be
specified in the same filter rule such that the conforming packets are
actually forwarded somewhere.

Apart from the VL Lookup, Policing and Forwarding tables which need to
be programmed for each flow (virtual link), the Schedule engine also
needs to be told to open/close the admission gates for each individual
virtual link. A fairly accurate (and detailed) description of how that
works is already present in sja1105_tas.c, since it is already used to
trigger the egress gates for the tc-taprio offload (IEEE 802.1Qbv). Key
point here, we remember that the schedule engine supports 8
"subschedules" (execution threads that iterate through the global
schedule in parallel, and that no 2 hardware threads must execute a
schedule entry at the same time). For tc-taprio, each egress port used
one of these 8 subschedules, leaving a total of 4 subschedules unused.
In principle we could have allocated 1 subschedule for the tc-gate
offload of each ingress port, but actually the schedules of all virtual
links installed on each ingress port would have needed to be merged
together, before they could have been programmed to hardware. So
simplify our life and just merge the entire tc-gate configuration, for
all virtual links on all ingress ports, into a single subschedule. Be
sure to check that against the usual hardware scheduling conflicts, and
program it to hardware alongside any tc-taprio subschedule that may be
present.

The following scenarios were tested:

1. Quantitative testing:

   tc qdisc add dev swp2 clsact
   tc filter add dev swp2 ingress flower skip_sw \
           dst_mac 42:be:24:9b:76:20 \
           action gate index 1 base-time 0 \
           sched-entry OPEN 1200 -1 -1 \
           sched-entry CLOSE 1200 -1 -1 \
           action trap

   ping 192.168.1.2 -f
   PING 192.168.1.2 (192.168.1.2) 56(84) bytes of data.
   .............................
   --- 192.168.1.2 ping statistics ---
   948 packets transmitted, 467 received, 50.7384% packet loss, time 9671ms

2. Qualitative testing (with a phase-aligned schedule - the clocks are
   synchronized by ptp4l, not shown here):

   Receiver (sja1105):

   tc qdisc add dev swp2 clsact
   now=$(phc_ctl /dev/ptp1 get | awk '/clock time is/ {print $5}') && \
           sec=$(echo $now | awk -F. '{print $1}') && \
           base_time="$(((sec + 2) * 1000000000))" && \
           echo "base time ${base_time}"
   tc filter add dev swp2 ingress flower skip_sw \
           dst_mac 42:be:24:9b:76:20 \
           action gate base-time ${base_time} \
           sched-entry OPEN  60000 -1 -1 \
           sched-entry CLOSE 40000 -1 -1 \
           action trap

   Sender (enetc):
   now=$(phc_ctl /dev/ptp0 get | awk '/clock time is/ {print $5}') && \
           sec=$(echo $now | awk -F. '{print $1}') && \
           base_time="$(((sec + 2) * 1000000000))" && \
           echo "base time ${base_time}"
   tc qdisc add dev eno0 parent root taprio \
           num_tc 8 \
           map 0 1 2 3 4 5 6 7 \
           queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7 \
           base-time ${base_time} \
           sched-entry S 01  50000 \
           sched-entry S 00  50000 \
           flags 2

   ping -A 192.168.1.1
   PING 192.168.1.1 (192.168.1.1): 56 data bytes
   ...
   ^C
   --- 192.168.1.1 ping statistics ---
   1425 packets transmitted, 1424 packets received, 0% packet loss
   round-trip min/avg/max = 0.322/0.361/0.990 ms

   And just for comparison, with the tc-taprio schedule deleted:

   ping -A 192.168.1.1
   PING 192.168.1.1 (192.168.1.1): 56 data bytes
   ...
   ^C
   --- 192.168.1.1 ping statistics ---
   33 packets transmitted, 19 packets received, 42% packet loss
   round-trip min/avg/max = 0.336/0.464/0.597 ms
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Some boards do not have the RX_ER MII signal connected. Normally in such
situation, those pins would be grounded, but then again, some boards
left it electrically floating.

When sending traffic to those switch ports, one can see that the
N_SOFERR statistics counter is incrementing once per each packet. The
user manual states for this counter that it may count the number of
frames "that have the MII error input being asserted prior to or
up to the SOF delimiter byte". So the switch MAC is sampling an
electrically floating signal, and preventing proper traffic reception
because of that.

As a workaround, enable the internal weak pull-downs on the input pads
for the MII control signals. This way, a floating signal would be
internally tied to ground.

The logic levels of signals which _are_ externally driven should not be
bothered by this 40-50 KOhm internal resistor. So it is not an issue to
enable the internal pull-down unconditionally, irrespective of PHY
interface type (MII, RMII, RGMII, SGMII) and of board layout.
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It looks like the P/Q/R/S series supports some more counters,
generically named "Ethernet statistics counter", which we were not
printing. Add them.
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The SJA1105 switch family has a PTP_CLK pin which emits a signal with
fixed 50% duty cycle, but variable frequency and programmable start time.

On the second generation (P/Q/R/S) switches, this pin supports even more
functionality. The use case described by the hardware documents talks
about synchronization via oneshot pulses: given 2 sja1105 switches,
arbitrarily designated as a master and a slave, the master emits a
single pulse on PTP_CLK, while the slave is configured to timestamp this
pulse received on its PTP_CLK pin (which must obviously be configured as
input). The difference between the timestamps then exactly becomes the
slave offset to the master.

The only trouble with the above is that the hardware is very much tied
into this use case only, and not very generic beyond that:
 - When emitting a oneshot pulse, instead of being told when to emit it,
   the switch just does it "now" and tells you later what time it was,
   via the PTPSYNCTS register. [ Incidentally, this is the same register
   that the slave uses to collect the ext_ts timestamp from, too. ]
 - On the sync slave, there is no interrupt mechanism on reception of a
   new extts, and no FIFO to buffer them, because in the foreseen use
   case, software is in control of both the master and the slave pins,
   so it "knows" when there's something to collect.

These 2 problems mean that:
 - We don't support (at least yet) the quirky oneshot mode exposed by
   the hardware, just normal periodic output.
 - We abuse the hardware a little bit when we expose generic extts.
   Because there's no interrupt mechanism, we need to poll at double the
   frequency we expect to receive a pulse. Currently that means a
   non-configurable "twice a second".
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Acked-by: NRichard Cochran <richardcochran@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

SJA1105 switches R and S have one SerDes port with an 802.3z
quasi-compatible PCS, hardwired on port 4. The other ports are still
MII/RMII/RGMII. The PCS performs rate adaptation to lower link speeds;
the MAC on this port is hardwired at gigabit. Only full duplex is
supported.

The SGMII port can be configured as part of the static config tables, as
well as through a dedicated SPI address region for its pseudo-clause-22
registers. However it looks like the static configuration is not
able to change some out-of-reset values (like the value of MII_BMCR), so
at the end of the day, having code for it is utterly pointless. We are
just going to use the pseudo-C22 interface.

Because the PCS gets reset when the switch resets, we have to add even
more restoration logic to sja1105_static_config_reload, otherwise the
SGMII port breaks after operations such as enabling PTP timestamping
which require a switch reset.

>From PHYLINK perspective, the switch supports *only* SGMII (it doesn't
support 1000Base-X). It also doesn't expose access to the raw config
word for in-band AN in registers MII_ADV/MII_LPA.
It is able to work in the following modes:
 - Forced speed
 - SGMII in-band AN slave (speed received from PHY)
 - SGMII in-band AN master (acting as a PHY)

The latter mode is not supported by this patch. It is even unclear to me
how that would be described. There is some code for it left in the
patch, but 'an_master' is always passed as false.
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: NRussell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We don't really need 10k species of reset. Remove everything except cold
reset which is what is actually used. Too bad the hardware designers
couldn't agree to use the same bit field for rev 1 and rev 2, so the
(*reset_cmd) function pointer is there to stay.

However let's simplify the prototype and give it a struct dsa_switch (we
want to avoid forward-declarations of structures, in this case struct
sja1105_private, wherever we can).
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Tested using the following bash script and the tc from iproute2-next:

	#!/bin/bash

	set -e -u -o pipefail

	NSEC_PER_SEC="1000000000"

	gatemask() {
		local tc_list="$1"
		local mask=0

		for tc in ${tc_list}; do
			mask=$((${mask} | (1 << ${tc})))
		done

		printf "%02x" ${mask}
	}

	if ! systemctl is-active --quiet ptp4l; then
		echo "Please start the ptp4l service"
		exit
	fi

	now=$(phc_ctl /dev/ptp1 get | gawk '/clock time is/ { print $5; }')
	# Phase-align the base time to the start of the next second.
	sec=$(echo "${now}" | gawk -F. '{ print $1; }')
	base_time="$(((${sec} + 1) * ${NSEC_PER_SEC}))"

	tc qdisc add dev swp5 parent root handle 100 taprio \
		num_tc 8 \
		map 0 1 2 3 5 6 7 \
		queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7 \
		base-time ${base_time} \
		sched-entry S $(gatemask 7) 100000 \
		sched-entry S $(gatemask "0 1 2 3 4 5 6") 400000 \
		clockid CLOCK_TAI flags 2

The "state machine" is a workqueue invoked after each manipulation
command on the PTP clock (reset, adjust time, set time, adjust
frequency) which checks over the state of the time-aware scheduler.
So it is not monitored periodically, only in reaction to a PTP command
typically triggered from a userspace daemon (linuxptp). Otherwise there
is no reason for things to go wrong.

Now that the timecounter/cyclecounter has been replaced with hardware
operations on the PTP clock, the TAS Kconfig now depends upon PTP and
the standalone clocksource operating mode has been removed.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The PTPSTRTSCH and PTPSTOPSCH bits are actually readable and indicate
whether the time-aware scheduler is running or not. We will be using
that for monitoring the scheduler in the next patch, so refactor the PTP
command API in order to allow that.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The PTP time of the switch is not preserved when uploading a new static
configuration. Work around this hardware oddity by reading its PTP time
before a static config upload, and restoring it afterwards.

Static config changes are expected to occur at runtime even in scenarios
directly related to PTP, i.e. the Time-Aware Scheduler of the switch is
programmed in this way.

Perhaps the larger implication of this patch is that the PTP .gettimex64
and .settime functions need to be exposed to sja1105_main.c, where the
PTP lock needs to be held during this entire process. So their core
implementation needs to move to some common functions which get exposed
in sja1105_ptp.h.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Through the PTP_SYS_OFFSET_EXTENDED ioctl, it is possible for userspace
applications (i.e. phc2sys) to compensate for the delays incurred while
reading the PHC's time.

The task itself of taking the software timestamp is delegated to the SPI
subsystem, through the newly introduced API in struct spi_transfer. The
goal is to cross-timestamp I/O operations on the switch's PTP clock with
values in the local system clock (CLOCK_REALTIME). For that we need to
understand a bit of the hardware internals.

The 'read PTP time' message is a 12 byte structure, first 4 bytes of
which represent the SPI header, and the last 8 bytes represent the
64-bit PTP time. The switch itself starts processing the command
immediately after receiving the last bit of the address, i.e. at the
middle of byte 3 (last byte of header). The PTP time is shadowed to a
buffer register in the switch, and retrieved atomically during the
subsequent SPI frames.

A similar thing goes on for the 'write PTP time' message, although in
that case the switch waits until the 64-bit PTP time becomes fully
available before taking any action. So the byte that needs to be
software-timestamped is byte 11 (last) of the transfer.

The patch creates a common (and local) sja1105_xfer implementation for
the SPI I/O, and offers 3 front-ends:

- sja1105_xfer_u32 and sja1105_xfer_u64: these are capable of optionally
  requesting a PTP timestamp

- sja1105_xfer_buf: this is for large transfers (e.g. the static config
  buffer) and other misc data, and there is no point in giving
  timestamping capabilities to this.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Adjusting the hardware clock (PTPCLKVAL, PTPCLKADD, PTPCLKRATE) is a
requirement for the auxiliary PTP functionality of the switch
(TTEthernet, PPS input, PPS output).

Therefore we need to switch to using these registers to keep a
synchronized time in hardware, instead of the timecounter/cyclecounter
implementation, which is reliant on the free-running PTPTSCLK.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This reworks the SPI transfer implementation to make use of more of the
SPI core features. The main benefit is to avoid the memcpy in
sja1105_xfer_buf().

The memcpy was only needed because the function was transferring a
single buffer at a time. So it needed to copy the caller-provided buffer
at buf + 4, to store the SPI message header in the "headroom" area.

But the SPI core supports scatter-gather messages, comprised of multiple
transfers. We can actually use those to break apart every SPI message
into 2 transfers: one for the header and one for the actual payload.

To keep the behavior the same regarding the chip select signal, it is
necessary to tell the SPI core to de-assert the chip select after each
chunk. This was not needed before, because each spi_message contained
only 1 single transfer.

The meaning of the per-transfer cs_change=1 is:

- If the transfer is the last one of the message, keep CS asserted
- Otherwise, deassert CS

We need to deassert CS in the "otherwise" case, which was implicit
before.

Avoiding the memcpy creates yet another opportunity. The device can't
process more than 256 bytes of SPI payload at a time, so the
sja1105_xfer_long_buf() function used to exist, to split the larger
caller buffer into chunks.

But these chunks couldn't be used as scatter/gather buffers for
spi_message until now, because of that memcpy (we would have needed more
memory for each chunk). So we can now remove the sja1105_xfer_long_buf()
function and have a single implementation for long and short buffers.

Another benefit is lower usage of stack memory. Previously we had to
store 2 SPI buffers for each chunk. Due to the elimination of the
memcpy, we can now send pointers to the actual chunks from the
caller-supplied buffer to the SPI core.

Since the patch merges two functions into a rewritten implementation,
the function prototype was also changed, mainly for cosmetic consistency
with the structures used within it.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This is a cosmetic patch that reduces some boilerplate in the SPI
interaction of the driver.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The new rule (as already started for sja1105_tas.h) is for functions of
optional driver components (ones which may be disabled via Kconfig - PTP
and TAS) to take struct dsa_switch *ds instead of struct sja1105_private
*priv as first argument.

This is so that forward-declarations of struct sja1105_private can be
avoided.

So make sja1105_ptp.h the second user of this rule.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix sparse warnings:

drivers/net/dsa/sja1105/sja1105_spi.c:159:5: warning: symbol 'sja1105_xfer_long_buf' was not declared. Should it be static?
Reported-by: NHulk Robot <hulkci@huawei.com>
Signed-off-by: Nzhengbin <zhengbin13@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The most commonly called function in the driver is long due for a
rename. The "packed" word is redundant (it doesn't make sense to
transfer an unpacked structure, since that is in CPU endianness yadda
yadda), and the "spi" word is also redundant since argument 2 of the
function is SPI_READ or SPI_WRITE.

As for the sja1105_spi_send_long_packed_buf function, it is only being
used from sja1105_spi.c, so remove its global prototype.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Having a function that takes a variable number of unpacked bytes which
it generically calls an "int" is confusing and makes auditing patches
next to impossible.

We only use spi_send_int with the int sizes of 32 and 64 bits. So just
make the spi_send_int function less generic and replace it with the
appropriate two explicit functions, which can now type-check the int
pointer type.

Note that there is still a small weirdness in the u32 function, which
has to convert it to a u64 temporary. This is because of how the packing
API works at the moment, but the weirdness is at least hidden from
callers of sja1105_xfer_u32 now.
Suggested-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In sja1105_static_config_upload, in two cases memory is leaked: when
static_config_buf_prepare_for_upload fails and when sja1105_inhibit_tx
fails. In both cases config_buf should be released.

Fixes: 8aa9ebcc ("net: dsa: Introduce driver for NXP SJA1105 5-port L2 switch")
Fixes: 1a4c6940 ("net: dsa: sja1105: Prevent PHY jabbering during switch reset")
Signed-off-by: NNavid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As Arnd Bergmann pointed out in commit 78fe8a28 ("net: dsa: sja1105:
fix ptp link error"), there is no point in having PTP support as a
separate loadable kernel module.

So remove the exported symbols and make sja1105.ko contain PTP support
or not based on CONFIG_NET_DSA_SJA1105_PTP.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Acked-by: NWillem de Bruijn <willemb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As per the DT phy-mode specification, RGMII delays are applied by the
MAC when there is no PHY present on the link.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Reviewed-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The pad_mii_tx registers point to the same memory region but were
unused. So convert to using these for RGMII I/O cell configuration, as
they bear a shorter name.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This will be used to stop egress traffic in .phylink_mac_link_up.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Reviewed-by: NAndrew Lunn <andrew@lunn.ch>
Reviewed-by: NFlorian Fainelli <f.fainelli@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

On TX, timestamping is performed synchronously from the
port_deferred_xmit worker thread.
In management routes, the switch is requested to take egress timestamps
(again partial), which are reconstructed and appended to a clone of the
skb that was just sent.  The cloning is done by DSA and we retrieve the
pointer from the structure that DSA keeps in skb->cb.
Then these clones are enqueued to the socket's error queue for
application-level processing.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The design of this PHC driver is influenced by the switch's behavior
w.r.t. timestamping.  It exposes two PTP counters, one free-running
(PTPTSCLK) and the other offset- and frequency-corrected in hardware
through PTPCLKVAL, PTPCLKADD and PTPCLKRATE.  The MACs can sample either
of these for frame timestamps.

However, the user manual warns that taking timestamps based on the
corrected clock is less than useful, as the switch can deliver corrupted
timestamps in a variety of circumstances.

Therefore, this PHC uses the free-running PTPTSCLK together with a
timecounter/cyclecounter structure that translates it into a software
time domain.  Thus, the settime/adjtime and adjfine callbacks are
hardware no-ops.

The timestamps (introduced in a further patch) will also be translated
to the correct time domain before being handed over to the userspace PTP
stack.

The introduction of a second set of PHC operations that operate on the
hardware PTPCLKVAL/PTPCLKADD/PTPCLKRATE in the future is somewhat
unavoidable, as the TTEthernet core uses the corrected PTP time domain.
However, the free-running counter + timecounter structure combination
will suffice for now, as the resulting timestamps yield a sub-50 ns
synchronization offset in steady state using linuxptp.

For this patch, in absence of frame timestamping, the operations of the
switch PHC were tested by syncing it to the system time as a local slave
clock with:

phc2sys -s CLOCK_REALTIME -c swp2 -O 0 -m -S 0.01
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

These are needed for the situation where the switch driver and the PTP
driver are both built as modules.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The DSA callbacks were written with the E/T (first generation) in mind,
which is quite different.

For P/Q/R/S completely new implementations need to be provided, which
are held as function pointers in the priv->info structure.  We are
taking a slightly roundabout way for this (a function from
sja1105_main.c reads a structure defined in sja1105_spi.c that
points to a function defined in sja1105_main.c), but it is what it is.

The FDB dump callback works for both families, hence no function pointer
for that.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The while-loop exit condition check is not correct; the
loop should continue if the returns from the function calls are
negative or the CRC status returns are invalid.  Currently it
is ignoring the returns from the function calls.  Fix this by
removing the status return checks and only break from the loop
at the very end when we know that all the success condtions have
been met.

Kudos to Dan Carpenter for describing the correct fix and
Vladimir Oltean for noting the change to the check on the number
of retries.

Addresses-Coverity: ("Uninitialized scalar variable")
Fixes: 8aa9ebcc ("net: dsa: Introduce driver for NXP SJA1105 5-port L2 switch")
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Tested-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

drivers/net/dsa/sja1105/sja1105_spi.c:486:21: warning: symbol 'sja1105et_regs' was not declared. Should it be static?
drivers/net/dsa/sja1105/sja1105_spi.c:511:21: warning: symbol 'sja1105pqrs_regs' was not declared. Should it be static?

Fixes: 8aa9ebcc ("net: dsa: Introduce driver for NXP SJA1105 5-port L2 switch")
Reported-by: NHulk Robot <hulkci@huawei.com>
Signed-off-by: NWang Hai <wanghai26@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Resetting the switch at runtime is currently done while changing the
vlan_filtering setting (due to the required TPID change).

But reset is asynchronous with packet egress, and the switch core will
not wait for egress to finish before carrying on with the reset
operation.

As a result, a connected PHY such as the BCM5464 would see an
unterminated Ethernet frame and start to jabber (repeat the last seen
Ethernet symbols - jabber is by definition an oversized Ethernet frame
with bad FCS). This behavior is strange in itself, but it also causes
the MACs of some link partners (such as the FRDM-LS1012A) to completely
lock up.

So as a remedy for this situation, when switch reset is required, simply
inhibit Tx on all ports, and wait for the necessary time for the
eventual one frame left in the egress queue (not even the Tx inhibit
command is instantaneous) to be flushed.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Reviewed-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

At this moment the following is supported:
* Link state management through phylib
* Autonomous L2 forwarding managed through iproute2 bridge commands.

IP termination must be done currently through the master netdevice,
since the switch is unmanaged at this point and using
DSA_TAG_PROTO_NONE.
Signed-off-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NGeorg Waibel <georg.waibel@sensor-technik.de>
Acked-by: NFlorian Fainelli <f.fainelli@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>