On some but not all devices using the RTL8221B 2.5GBit/s PHY the SerDes
setup sequence may hang under some circumstances (eg. <2500M link
partner present during boot).
RTL8221B-VB-CG 2.5Gbps PHY (C45) mdio-bus:01: rtl822xb_config_init failed: -110
Work-around the issue by performing a hardware reset and subsequent
retry of the SerDes setup, which seems to always succeed.
Doing this requires moving ALDPS setup to config_init (which is anyway
the better place for that) as it otherwise doesn't survive the reset.
Also disable listening on MDIO address 0 which may be used by other PHYs
despite being spec'ed as "broadcast address", as bus activity on address
0 may otherwise confuse the RealTek PHY for good reasons.
Tested-by: Luis Mita <luis@luismita.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit c87a767801)
Link: https://github.com/openwrt/openwrt/pull/17790
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Import patch "net: phy: realtek: mark existing MMDs as present"
When using Clause-45 mode to access RealTek RTL8221B 2.5G PHYs some
versions of the PHY fail to report the MMDs present on the PHY.
Mark MMDs PMAPMD, PCS and AN which are always existing according to
the datasheet as present to fix that.
Fixes: #16823, #17183, #17232
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Tested-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Tested-by: Juan Pedro Paredes Caballero <juanpedro.paredes@gmail.com>
(cherry picked from commit 730db6b893)
This patch has been accepted for linux v6.14 so we can move it from pending
to backport.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 2f2e21a52b)
Expose the temperature sensor built-into RTL822x 2.5G and 5G PHYs.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 3a189e3831)
1. Import pending patch to fix ramips/mt7621 64MB targets.
2. Do not enable CONFIG_PAGE_POOL_STATS by default.
Signed-off-by: Danila Romanov <pervokur@gmail.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 15887235c1)
These patches have been accepted, so we can move them from pending to
backported.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 7924acdd63)
Import patches to fix several issues with status reporting of RealTek
2.5G PHYs.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 41c164d32b)
This patch is needed on bmips since it fixes issues with GPIOs not being
properly configured due to gpio_request_enable not being called on bcm63xx
devices. Therefore we can now drop the bcm63268 gpio function patch.
Backported from f5b1d340be with the exception of
the realtek patch removal.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit f5b1d340be)
The broadcom PHY driver only has to depend upon PTP_1588_CLOCK_OPTIONAL
if NETWORK_PHY_TIMESTAMPING is enabled. The PTP functionality is stubbed
in this case.
Reflect this circumstance in the dependence condition. This allows to
build the driver as a built-in module even if PTP is built as a module.
This is required to include the broadcom PHY module regardless of the
built-setting of the PTP subsystem. On ath79 (and probably more)
targets with Broadcom PHY, Gigabit operation is currently broken as the
PHY driver is only built as a module in case all kernel-packages are
built. Due to this circumstance, affected devices fall back to using the
generic PHY driver.
Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit cbce32e30c29dc69907c6f4c0ab79dd5d9a8fb28)
commit eee3c695f3 ("linux-firmware: add offloading firmware for MT7988")
added mt7988_wo_{0,1}.bin in the 'mediatek/mt7988' directory while driver
currently expects the files in the 'mediatek' directory.
Import pending patch which changes the path in the driver header now
that the firmware has been added.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Fix conditions for handling offloaded packets
Fixes: #13430
Signed-off-by: Leon M. Busch-George <leon@georgemail.eu>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Backport newly introduced support for 'active-high' property and use
it to correctly implement polarity assignment for Aquantia PHY LEDs.
Previously the 'active-low' property was used to switch a LED PIN to
active-high ("drive VDD" in Aquantia-speak) mode.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Only add SerDes interface modes at lower speeds to host_interfaces.
Fixes: 5281033831 ("generic: net: phy: use all SerDes MAC interface modes")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
This solution that is needed for some routers to provide proper
LED activity when controlled directly by the MV88E6xxx switch,
has just been merged in upstream Linux.
Make this patch 901 as other backports from earlier kernels
are in patch 896 and this is the first free number after
that.
Patch offsets in pending patches are augmented as part of
the refresh.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Backport BLOCK OF support patch merged upstream and refresh pending
BLOCK patches.
This is a new way to declare partition table for BLOCK device (eMMC
currently supported) with the use of DTS.
Current pending patch are adapted to not cause regression with current
downstream implementation of a similar functionality.
Also enable the new OF_PARTITION config by default.
Link: https://github.com/openwrt/openwrt/pull/16663
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Instead of forcing 2.5G PHYs into rate-adapter mode which results higher
energy consumption, lack of support for half-duplex modes and typically
worse performance when linked at speeds less than 2.5G, use SGMII mode
which allows the MAC to follow the PHY speed.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Follow the advise of Russell King allows to greatly improve the driver
for RealTek's 1G and 2.5G Ethernet PHYs. The results are full/half
duplex as well as Gbit master/slave property being read from PHY
Specific Status Register (PHYSR), and fixes regarding link-partner
advertisement.
Fixes: #14504
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
This fixes some corner cases triggered by enabling fraglist GRO, where some
protocols may accidentally or intentionally linearize fraglist skbs.
Previously, these skbs became unusable and segmenting them led to crashes.
With this patch, they are properly handled by passing them to skb_segment
instead of skb_segment_list.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
of_property_read_u32 returns -EINVAL when property does not exist,
according to the documentation -ENOENT is not a valid return code.
So, instead of checking for -ENOENT check for -EINVAL as otherwise the
blamed commit breaks AQR probe since it will return -EINVAL during probe.
Fixes: cb2a11f49c ("generic: phy: aquantia: add pending patch to force MDI pair order")
Link: https://github.com/openwrt/openwrt/pull/16466
Signed-off-by: Robert Marko <robimarko@gmail.com>
Increase QDMA RESV_BUF from 2K to 3K for netsys v2 to match Mediatek SDK.
This helps reduce the possibility of Ethernet transmit timeouts.
Link: 19d8456c30
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The purpose of resetting the TX queue is to reset the
byte and packet count as well as to clear the software
flow control XOFF bit.
MediaTek developers pointed out that netdev_reset_queue would only
resets queue 0 of the network device.
Queues that are not reset may cause unexpected issues.
Packets may stop being sent after reset and "transmit timeout" log may
be displayed.
Import fix from MediaTek's SDK to resolve this issue.
Link: 5746a94456
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The 'active-low' property was not applied correctly and two fixes
are required to make inverted LEDs work as expected.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
