The bootloader can leave the GPIO expander in a state which doesn't have
output drivers enabled so GPIOs will properly work for input but output
operations will have no effect.
To avoid disrupting the boot in case the bootloader left direction and
data registers in an inconsistent state (e.g. pulling SoC's reset to 0)
reconfigure everything as input.
Reviewed-by: Sander Vanheule <sander@svanheule.net>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
GPIOs > 31 require special handling. This patch fixes both the
initialisation and direction get/set operations.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Reviewed-by: Sander Vanheule <sander@svanheule.net>
A superflus ')' character has slipped into commit 91a52f22a1. Remove it
to fix build.
Fixes: 91a52f22a1 ("treewide: backport support for nvmem on non platform devices")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
In the current state, nvmem cells are only detected on platform device.
To quickly fix the problem, we register the affected problematic driver
with the of_platform but that is more an hack than a real solution.
Backport from net-next the required patch so that nvmem can work also
with non-platform devices and rework our current patch.
Drop the mediatek and dsa workaround and rework the ath10k patches.
Rework every driver that use the of_get_mac_address api.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
On reset, the PVID of all ports is set to 1; if this is reset to 0,
the special cases for VLAN 1 are no longer required.
port_vlan_add is called with vid=0 when the DSA port interfaces are
enabled with no VLAN; previously the VLAN was not configured in this
case, relying on VLAN 1 being present, but with the PVID set to 0,
configuring VLAN 0 as normal works as expected.
Signed-off-by: Thomas Nixon <tom@tomn.co.uk>
Fixes the buffer and packet length calculations for Ethernet TX on
the RTL8380 SoC when CRC calculation offload is enabled.
CRC-offload is always done by the SoC, but additional CRC
calculation was previously done also by the kernel.
It also fixes detection of the DSA tag for packets on RTL8390
SoCs for ports > 28.
v2 has correct whitespace
Signed-off-by: Birger Koblitz <mail@birger-koblitz.de>
This adds the CPU port to the unknown multicast flooding port mask,
which fixes the VLAN issues introduced by the multicast group patches
Tested-by: Russell Senior <russell@personaltelco.net> [Netgear GS108Tv3]
Signed-off-by: Birger Koblitz <mail@birger-koblitz.de>
Signed-off-by: Bjørn Mork <bjorn@mork.no> [whitespace fix]
Signed-off-by: Petr Štetiar <ynezz@true.cz> [unknwon typo fix]
Adds support for Layer 2 multicast by implementing the DSA port_mdb_*
callbacks. The Kernel bridge listens to IGMP/MLD messages trapped to
the CPU-port, and calls the Multicast Forwarding Database updates.
The updates manage the L2 forwarding entries and the multicast
port-maps.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
This sets up all VLANs with a default configuration on reset:
- forward based on VLAN-ID and not the FID/MSTI
- forward based on the inner VLAN-ID (not outer)
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
This adds SoC specific VLAN configuration routines, which
alsoe sets up the portmask table entries that are referred to
in the vlan profiles registers for unknown multicast flooding.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
This adds support for the MMD access registers the RTL-SoCs use to access clause 45
PHYs via mdio.
This new interface is used to add EEE-support for the RTL8226
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
Adds a hash-bucket size attribute for the different SoCs, in order to
accomodate the buckets with 8 entries of the L2-forwarding tables
on RTL93XX in contrast to only 4 on RTL83XX.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
This adds snooping support for IGMP and MLD on RTL8380/90/9300
by trapping IGMP and MLD packets to the CPU.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
Enables CRC calculation offloading on RTL8380/8390/9300.
Tested on Zyxel XGS1210-10 (RTL9302)/GS1900-48 (RTL8390)/GS1900-10HP (RTL8382)
On the Zyxel GS1900-10HP, an increase of 5% in iperf3 send throughput
and 11% in receive throughput is seen.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
From the validate docs in include/linux/phylink.h:
When state->interface is PHY_INTERFACE_MODE_NA, phylink expects the
MAC driver to return all supported link modes.
Tested-by: Birger Koblitz <mail@birger-koblitz.de>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Demote a number of debugging printk's to pr_debug to avoid log
nosie. Several of these functions are called as a result of
userspace activity. This can cause a lot of log noise when
userspace does periodic polling.
Most of this could probably be removed completely, but let's
keep it for now since these drivers are still in development.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Tested-by: Stijn Tintel <stijn@linux-ipv6.be>
Add a table API that has per accss register locking and uses
register description information to handle all table access
through a single set of api calls.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
Adds support for the RTL9300 and RTL9310 series of switches
with 10GBit per port and up to 56 ports.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
This adds support for the internal SerDes of the RTL9300 SoC
and for the RTL8218D and RTL8226B phys found in combination
with this SoC in switches.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
This fixes the usage of the RTL8231 GPIO extender chip
when used with the RTL839X SoCs. Specifically,
the PHY addresses may be different from 0.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
This adds support for identifying QoS information in packets
and use this and rate control information to submit to multiple
egress queues. The ethernet driver is also made to support
2 egress and up to 32 egress queues.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
this adds support for the SoC timer of the RTL9300 chips, it
provides 6 independent timer/counters, of which the first one
is used as a clocksource and the second one as event timer.
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
This adds support for the RTL8390 and RTL9300 SoCs
it also cleans up unnecessary definitions in mach-rtl83xx.h
and moves definitions relevant for irq routing to irq.h
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
This adds support to detect RTL930X based SoCs and the RTL9313 SoC.
Tested on Zyxel XGS1210-10 (RTL9302B SoC) and the
Zyxel XS1930-12 (RTL9313 SoC)
Signed-off-by: Birger Koblitz <git@birger-koblitz.de>
If _machine_hang is not defined on MIPS, the kernel will check if the
CPU can enter a more power efficient sleep mode. Since the realtek
platform supports mips32_r2, this should issue a WAIT instruction
instead of a trivial infinite loop.
Signed-off-by: Sander Vanheule <sander@svanheule.net>
Cppcheck shows here duplicated break.
Code `state->speed = SPEED_1000;` will be never executed because above
it there is break statement.
Almost identical statement is placed in another realtek driver
18a53d43d6/target/linux/realtek/files-5.4/drivers/net/dsa/rtl83xx/dsa.c (L286-L294)
Signed-off-by: Rafał Mikrut <mikrutrafal@protonmail.com>
Cppcheck shows self initialization error, which is an obvious bug.
Basing on logic of similar fragment below I assigned to this variable,
value `RTL838X_LED_GLB_CTRL` which I think is proper.
Signed-off-by: Rafał Mikrut <mikrutrafal@protonmail.com>
