atm_qos struct should be the same both for user and kernel spaces. Via
the __SO_ENCODE() macro it is used to define the SO_ATMQOS socket IOC.
During the VRX518 support introduction, the atm_trafprm sturct nested
into the atm_qos stucture was update with newer fields that are
referenced by the ATM TC layer of the VRX518 TC driver. These new fields
are intended to communicate information for extra traffic classes
supported by the driver. But we are still using vanilla kernel headers
to build the toolchain. Due to the atm.h header incoherency br2684ctl
from linux-atm tools is incapable to configure the ATM bridge netdev:
br2684ctl: Interface "dsl0" created sucessfully
br2684ctl: Communicating over ATM 0.1.2, encapsulation: LLC
br2684ctl: setsockopt SO_ATMQOS 22 <-- EINVAL errno
br2684ctl: Fatal: failed to connect on socket; File descriptor in bad state
There are two options to fix this incoherency. (a) update the header
file in the toolchain to build linux-atm against updated atm_trafprm and
atm_qos structures, or (b) revert atm_trafprm changes.
Since there are no actual users of the extra ATM QoS traffic classes,
just drop these extra traffic classes from vrx518_tc ATM TC layer and
drop the kernel patch updating atm.h.
Besides fixing the compatibility with linux-atm tools, removing the
kernel patch should simplify kernel updates removing unneeded burden of
maintenance.
Run tested with FRITZ!Box 7530 with disabled extra traffic classes and
then removed them entirely before the submission.
CC: John Crispin <john@phrozen.org>
Fixes: cfd42a0098 ("ipq40xx: add Intel/Lantiq ATM hacks")
Suggested-by: Andre Heider <a.heider@gmail.com>
Reported-and-tested-by: nebibigon93@yandex.ru
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Link: https://patchwork.ozlabs.org/project/openwrt/patch/20250122222654.21833-4-ryazanov.s.a@gmail.com/
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Some VRX518 modems fail to initialize properly with the error message
"dc_ep_clk_on failed". As a result, the DSL data path doesn't work.
This hack, which is based on code from the FRITZ!Box 7530 GPL archive,
fixes the issue. It changes the PCIe vendor/device ID to values matching
a Lantiq SoC. It also appears to emulate a Lantiq CPU ID register for
connected PCIe devices, by remapping the matching address area to a
specially crafted buffer using the address translation unit.
A dedicated compatible is created to activate this in
the device tree, so this shouldn't affect any devices other than
FRITZ!Box 7530/7520.
Original investigation was done in 59f5212517 which used the "avm,host_magic" property to enabled the patch.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Florian Maurer <f.maurer@outlook.de>
Link: https://github.com/openwrt/openwrt/pull/17622
Signed-off-by: Robert Marko <robimarko@gmail.com>
Avoids having to call mutex_destroy in remove.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16755
Signed-off-by: Robert Marko <robimarko@gmail.com>
Easier to just use devm_platform_ioremap_resource.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16701
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Notify via dev_info when a random MAC address is set.
of_get_ethdev_address can return -EPROBE_DEFER for NVMEM devices,
return this up, so that ipqess can defer as well.
Also move this MAC assignment from _init into _probe, so that this defer
can happen earlier.
Before change, with MAC address allocated from a built-in NVMEM layout
driver (mikrotik,routerboot-nvmem) with extra of_get_ethdev_address
result and random MAC printfs:
[ 1.197571] ipqess-edma c080000.ethernet: ipqess_init of_get_ethdev_address ret:-517, of_node: /soc/ethernet@c080000
[ 1.197690] ipqess-edma c080000.ethernet: generated random MAC address 22:e7:36:e0:e4:a3
[ 1.614444] ipqess-edma c080000.ethernet eth0: entered promiscuous mode
[ 5.055929] ipqess-edma c080000.ethernet eth0: configuring for fixed/internal link mode
after change:
[ 1.173776] ipqess-edma c080000.ethernet: of_get_ethdev_address: -517
[ 1.614607] ipqess-edma c080000.ethernet eth0: entered promiscuous mode
[ 5.246105] ipqess-edma c080000.ethernet eth0: configuring for fixed/internal link mode
[ 5.260754] ipqess-edma c080000.ethernet eth0: Link is Up - 1Gbps/Full - flow control rx/tx
with extra __func__ printf to show _probe and _init:
[ 1.173685] ipqess-edma c080000.ethernet: ipqess_axi_probe
[ 1.173784] ipqess-edma c080000.ethernet: of_get_ethdev_address: -517
[ 1.280347] ipqess-edma c080000.ethernet: ipqess_axi_probe
[ 1.304844] ipqess-edma c080000.ethernet: ipqess_init
[ 1.614664] ipqess-edma c080000.ethernet eth0: entered promiscuous mode
Signed-off-by: John Thomson <git@johnthomson.fastmail.com.au>
Link: https://github.com/openwrt/openwrt/pull/15831
Signed-off-by: Robert Marko <robimarko@gmail.com>
Since the introduction of out-of-band tagging, writing the outbound tag
had been completely broken: First, in place of a port mask containing
the port number, just the port number itself was set in the register
value. Just after that, the full port mask 0x3e (all 5 external ports)
was set unconditionally.
This remained unnoticed because the switch would then use the FDB to
decide where to send unicast packets; broadcast and multicast packets
were however sent to every port.
Fix the port tag computation and only use the full port mask as a
fallback for non-DSA mode, as it was done in the older driver patches
used on Linux 5.15.
Fixes: cd9c7211241e ("ipq40xx: 6.1: use latest DSA and ethernet patches")
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Signed-off-by: Robert Marko <robimarko@gmail.com>
Currently, in case that PSGMII calibration fails it will panic the kernel
which is not ideal and is preventing any debugging to be done.
So, since PGMII calibration failing only means that wired networking wont
work lets convet the panic() call to dev_error.
Link: https://github.com/openwrt/openwrt/pull/15542
Signed-off-by: Robert Marko <robimarko@gmail.com>
AVM FRITZ!Repeater 1200 does not use QCA807x PHY at all and thus it
disables all of the individual PHY nodes, however this is not enough
anymore since the conversion to PHY package.
Now its now enough to disable the PHY-s in the package alone, but the PHY
package node itself must also be disabled.
Fixes: 1b931c33a28e ("ipq40xx: adapt to new Upstream QCA807x PHY driver")
Fixes: #15355
Link: https://github.com/openwrt/openwrt/pull/15365
Signed-off-by: Robert Marko <robimarko@gmail.com>
CI tells us that we need to, so lets refresh them.
Fixes: 4265fe9348c0 ("ipq40xx: refresh patches")
Link: https://github.com/openwrt/openwrt/pull/15310
Signed-off-by: Robert Marko <robimarko@gmail.com>
Add pending patch fixing mtdcore with MTD OTP with a fragile detection
if Nand supports OTP. Patch has been sent upstream and will be backported
to stable kernel if accepted.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Rework kernel patches for new kernel. Mainly adaptation for patch
related to DTS, OOB Tagger and SDHCI patch.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
This is an automatically generated commit.
When doing `git bisect`, consider `git bisect --skip`.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
