This allows the user to specify a larger tx ring buffer size via ethtool.
Having symmetrical ring buffer sizes increases throughput on high bandwidth
(1 gbps tested) network connections.
The default value is not changed so the same behaviour is saved.
Signed-off-by: Robert Meijer <robert.s.meijer@gmail.com>
[ improve title, commit description and wrap to 80 columns ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
napi_build_skb() reuses NAPI skbuff_head cache in order to save some
cycles on freeing/allocating skbuff_heads on every new Rx or completed
Tx.
Use napi_consume_skb() to feed the cache with skbuff_heads of completed
Tx so it's never empty.
Signed-off-by: Sieng Piaw Liew <liew.s.piaw@gmail.com>
[ fixed commit title ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
The phy-mask property is read by the ag71xx-mdio driver to set the
mii_bus's phy_mask field. On OF platforms, the devicetree is expected to
provide all present ethernet phy-s however, so the phy_mask field is
later set to all-ones. Having a devicetree override is of no use then,
so let's drop it.
Cc: David Bauer <mail@david-bauer.net>
Cc: John Crispin <john@phrozen.org>
Signed-off-by: Sander Vanheule <sander@svanheule.net>
Acked-by: David Bauer <mail@david-bauer.net>
On MikroTik RB91x board series a reset key shares SoC gpio
line #15 with NAND ALE and NAND IO7. So we need a custom
gpio driver to manage this non-trivial connection schema.
Also rb91x-nand needs to have an ability to disable a polling
of the key while it works with NAND.
While we've been integrating rb91x-key into a firmware, we've
figured out that:
* In the gpio-latch driver we need to add a "cansleep" suffix to
several gpiolib calls,
* When gpio-latch and rb91x-nand fail to get a gpio and an error
is -EPROBE_DEFER, they shouldn't report about this, since this
actually is not an error and occurs when the gpio-latch probe
function is called before the rb91x-key probe.
We fix these related things here too.
Signed-off-by: Denis Kalashnikov <denis281089@gmail.com>
Reviewed-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Tested-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
ag71xx_probe is registering ag71xx_interrupt as handler for the gmac0/gmac1
interrupts. The handler is trying to use napi_schedule to handle the
processing of packets. But the netif_napi_add for this device is
called a lot later in ag71xx_probe.
It can therefore happen that a still running gmac0/gmac1 is triggering the
interrupt handler with a bit from AG71XX_INT_POLL set in
AG71XX_REG_INT_STATUS. The handler will then call napi_schedule and the
napi code will crash the system because the ag->napi is not yet
initialized:
libphy: Fixed MDIO Bus: probed
CPU 0 Unable to handle kernel paging request at virtual address 00000000, epc == 00000000, ra == 81373408
Oops[#1]:
CPU: 0 PID: 1 Comm: swapper Not tainted 5.4.152 #0
$ 0 : 00000000 00000001 00000000 8280bf28
$ 4 : 82a98cb0 00000000 81620000 00200140
$ 8 : 00000000 00000000 74657272 7570743a
$12 : 0000005b 8280bdb9 ffffffff ffffffff
$16 : 00000001 82a98cb0 00000000 8280bf27
$20 : 8280bf28 81620000 ffff8b00 8280bf30
$24 : 00000000 8125af9c
$28 : 82828000 8280bed8 81610000 81373408
Hi : 00005fff
Lo : 2e48f657
epc : 00000000 0x0
ra : 81373408 __napi_poll+0x3c/0x11c
Status: 1100dc03 KERNEL EXL IE
Cause : 00800008 (ExcCode 02)
BadVA : 00000000
PrId : 00019750 (MIPS 74Kc)
Modules linked in:
Process swapper (pid: 1, threadinfo=(ptrval), task=(ptrval), tls=00000000)
Stack : ffff8afb ffff8afa 81620000 00200140 00000000 82a98cb0 00000008 0000012c
81625620 81373684 ffffffff ffffffff ffffffef 00000008 816153d8 81620000
815b0d60 815bbd54 00000000 81753700 8280bf28 8280bf28 8280bf30 8280bf30
81753748 00000008 00000003 00000004 0000000c 00000100 3fffffff 8175373c
816059f0 814ddb48 00000001 8160ab30 81615488 810618bc 00000006 00000000
...
Call Trace:
[<81373684>] net_rx_action+0xfc/0x26c
[<814ddb48>] __do_softirq+0x118/0x2ec
[<810618bc>] handle_percpu_irq+0x50/0x80
[<8125ab8c>] plat_irq_dispatch+0x94/0xc8
[<81004e98>] handle_int+0x138/0x144
Code: (Bad address in epc)
---[ end trace a60d797432b656b2 ]---
The gmcc0/gmac1 must be brought in a state in which it doesn't signal a
AG71XX_INT_POLL related status bits as interrupt before registering the
interrupt handler. ag71xx_hw_start will take care of re-initializing the
AG71XX_REG_INT_ENABLE.
Fixes: f529a37420 ("surprise :p")
Signed-off-by: Sven Eckelmann <sven@narfation.org>
This was old code from the AR71XXs target days that
doesn't get compiled and used anymore.
Bringing up AR92xx and earlier chips from their
OWL-Emulator state is currently done by the upstream
ath9k-pci-owl-loader module. (see the kmod-owl-loader
package).
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
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>
Currently, the option to disable subpage writing is only set
when a HW ECC engine is used.
Some boards lack a HW ECC engine and use software for that.
In this case, this NAND option does not get set when the NAND chip
does not support it, resulting in mounting errors.
Move the setting of this option to a generic init location so it
gets set for all types where required.
While at it, also OR the option instead of just setting it
so we don't overwrite potential flags being set somewhere else.
Before:
[ 1.681273] UBI: auto-attach mtd2
[ 1.684669] ubi0: attaching mtd2
[ 1.688877] ubi0 error: validate_ec_hdr: bad VID header offset 2048, expected 512
[ 1.696469] ubi0 error: validate_ec_hdr: bad EC header
[ 1.701712] Erase counter header dump:
[ 1.705512] magic 0x55424923
[ 1.709322] version 1
[ 1.712330] ec 1
[ 1.715331] vid_hdr_offset 2048
[ 1.718610] data_offset 4096
[ 1.721880] image_seq 1462320675
[ 1.725680] hdr_crc 0x12255a15
After:
1.680917] UBI: auto-attach mtd2
[ 1.684308] ubi0: attaching mtd2
[ 2.954504] random: crng init done
[ 3.142813] ubi0: scanning is finished
[ 3.163455] ubi0: attached mtd2 (name "ubi", size 124 MiB)
[ 3.169069] ubi0: PEB size: 131072 bytes (128 KiB), LEB size: 126976 bytes
[ 3.176037] ubi0: min./max. I/O unit sizes: 2048/2048, sub-page size 2048
[ 3.182942] ubi0: VID header offset: 2048 (aligned 2048), data offset: 4096
[ 3.190013] ubi0: good PEBs: 992, bad PEBs: 0, corrupted PEBs: 0
[ 3.196102] ubi0: user volume: 3, internal volumes: 1, max. volumes count: 128
[ 3.203434] ubi0: max/mean erase counter: 2/0, WL threshold: 4096, image sequence number: 1462320675
[ 3.212700] ubi0: available PEBs: 0, total reserved PEBs: 992, PEBs reserved for bad PEB handling: 20
[ 3.222124] ubi0: background thread "ubi_bgt0d" started, PID 317
[ 3.230246] block ubiblock0_1: created from ubi0:1(rootfs)
[ 3.235819] ubiblock: device ubiblock0_1 (rootfs) set to be root filesystem
[ 3.256830] VFS: Mounted root (squashfs filesystem) readonly on device 254:0.
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Main part is copied from ar71xx original driver rb91x_nand
written by Gabor Juhos <juhosg@openwrt.org>.
What is done:
* Support of kernel 5.4 and 5.10,
* DTS support,
* New gpio API (gpiod_*) support.
Reviewed-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Signed-off-by: Denis Kalashnikov <denis281089@gmail.com>
This is a slighty modified version of ar71xx gpio-latch driver
written by Gabor Juhos <juhosg@openwrt.org>.
Changes:
* DTS support,
* New gpio API (gpiod_*).
Reviewed-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Signed-off-by: Denis Kalashnikov <denis281089@gmail.com>
This updates the NAND driver for MikroTik RB4XX series to work with
kernel 5.10, similarly to the ar934x-nand driver (fb64e2c3).
Support for kernel 5.10 was added to all ath79 subtargets except for the
mikrotik one by commit d6b785d, since patch 920-mikrotik-rb4xx.patch
needed to be reworked. Later, commit f8512661 enabled kernel 5.10 for
the mikrotik subtarget with the nand-rb4xx driver still pending, which
is updated and added back by this patch.
Compile-tested only.
Signed-off-by: Roger Pueyo Centelles <roger.pueyo@guifi.net>
The mode on the SGMII SerDes on the QCA9563 is 1000 Base-X by default.
This only allows for 1000 Mbit/s links, however when used with an SGMII
PHY in 100 Mbit/s link mode, the link remains dead.
This strictly has nothing to do with the SerDes calibration, however it
is done at the same point in the QCA reference U-Boot which is the
blueprint for everything happening here. As the current state is more or
less a hack, this should be fine.
This fixes the issues outlined above on a TP-Link EAP-225 Outdoor.
Reported-by: Tom Herbers <freifunk@tomherbers.de>
Tested-by: Tom Herbers <freifunk@tomherbers.de>
Signed-off-by: David Bauer <mail@david-bauer.net>
As mangix pointed out on IRC, ioremap and ioremap_nocache are
functionally equivalent on kenrel 5.4 and 5.10.
Therefore we can use ioremap regardless of the kernel the driver
gets compiled for.
Signed-off-by: David Bauer <mail@david-bauer.net>
Adapt the driver to make it work with the NAND subsystem changes between
kernel 5.4 and 5.10.
Tested-on: Aerohive HiveAP121
Signed-off-by: David Bauer <mail@david-bauer.net>
Add statistics to ethtool. The statistics can be useful to
debug network issues.
The code is backported from mainline ag71xx.c driver.
Signed-off-by: Leon Leijssen <leon.git@leijssen.info>
This commit add a workaround for non working SGMII link observed on some
QCA956x SoCs. The workaround originates part from the U-Boot source code
from QCA, part from the implementation from TP-Link found in the GPL
tarball for the EAP245v1.
Extends commit 0d416a8d3b for QCA956x.
Note that reset is the same on QCA955x and QCA956x, same register offset
and values.
Auto calibration is done on u-boot, but always fall back to default value
0x7. Add a DTS entry serdes-cal in case a device require another value.
Signed-off-by: Julien Dusser <julien.dusser@free.fr>
[Sander Vanheule:
Minor code style fixes,
Remove hunk adding qca956x-serdes-fixup to a missing DTS,
Remove variable err that was only assigned,
Rename function to sgmii_serdes_init,
Lower priority of serdes call message to pr_debug]
Signed-off-by: Sander Vanheule <sander@svanheule.net>
Some bootloaders do not set up gmac0 properly, leaving it disconnected
from the sgmii interface. If the user specificies phy-mode sgmii, then
use the gmac-config/device node to ensure the mux is configured
correctly.
Signed-off-by: Sander Vanheule <sander@svanheule.net>
This removes unneeded kernel version switches from the targets after
kernel 4.19 has been dropped.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Fixes:
- CVE-2020-10757
The "mtd: rawnand: Pass a nand_chip object to nand_release()" commit was
backported which needed some adaptations to other code.
Run tested: ath79
Build tested: ath79
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This adds 3 Mikrotik rb4xx series drivers as follows:
rb4xx-cpld: This is in the mfd subsystem, and is the parent CPLD device
that interfaces between the SoC SPI bus and its two children below.
rb4xx-gpio: This is the GPIO expander.
rb4xx-nand: This is the NAND driver.
The history of this code comes in three phases.
1. The first is a May 2015 attempt to push the equivalient ar71xx rb4xx
drivers upstream. See https://lore.kernel.org/patchwork/patch/940880/.
Module-author: Gabor Juhos <juhosg@openwrt.org>
Module-author: Imre Kaloz <kaloz@openwrt.org>
Module-author: Bert Vermeulen <bert@biot.com>
2. Next several ar71xx patches were applied bringing the code current.
commit 7bbf4117c6
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
commit af79fdbe4a
commit 889272d92d
commit e21cb649a2
commit 7c09fa4a74
Signed-off-by: Felix Fietkau <nbd@nbd.name>
3. Finally a heavy refactor to split the driver into the three new
subsystems, and updated to work with the device tree configuration, plus
updates and review feedback incorporated
Reviewed-by: Thibaut VARÈNE <hacks@slashdirt.org>
Signed-off-by: Christopher Hill <ch6574@gmail.com>
Currently, ag71xx will trigger a warning when TX xor RX-Delay modes are
enabled.
Handle them identical to the already implemented RGMII modes, as they
are only different for the attached PHY.
Signed-off-by: David Bauer <mail@david-bauer.net>
ath79 does not support kernels prior to 4.19 anymore.
Remove legacy code for those kernels from the ag71xx driver.
Signed-off-by: David Bauer <mail@david-bauer.net>
Between 4.19 and 5.4, mtd parsers have been moved to "parsers"
subdirectory. Like for myloader.c in the previous patch,
this patch moves tplinkpart.c to the kernel patches, so the
code and the kernel includes are at the same location and
the path can be adjusted per kernel.
While at it, remove some outdated kernel version switches from
the C code.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The reset assert and deassert methods currently miss
a return value, leading to a compilation warning.
Return the return-value of reset_control_assert and
reset_control_deassert to fix these warnings.
Suggested-by: Michal Cieslakiewicz <michal.cieslakiewicz@wp.pl>
Signed-off-by: David Bauer <mail@david-bauer.net>
This fixes the compilation of the AR934x NAND controller
driver for kernel 5.4 while leaving it untouched for
kernel 4.19.
This change is currently not run-tested, as i do not have such
a device at hand.
CC: Michal Cieslakiewicz <michal.cieslakiewicz@wp.pl>
CC: André Valentin <avalentin@marcant.net>
CC: WeiDong Jia <jwdsccd@gmail.com>
Signed-off-by: David Bauer <mail@david-bauer.net>
Tested-by: Michal Cieslakiewicz <michal.cieslakiewicz@wp.pl>
The ag71xx driver from Linux 5.4 currently has various shortcomings
when used with OpenWrt compared to our downstream version.
For example, the upstream driver does not support modifying the ethernet
clock and configuring RGMII delays on the MAC side.
While we should certainly switch to the upstream driver, the amount of
necessary patches would make it cumbersome to work with. It's also
highly likely we won't be able to finish patching the upstream driver in
time for a Linux 5.4 release.
Tested on Siemens WS-AP3610.
CC: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: David Bauer <mail@david-bauer.net>
Acked-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: David Bauer <mail@david-bauer.net>
[refreshed]
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
* Sync the patches with the changes done for kernel 4.19
* Use KERNEL_TESTING_PATCHVER
* Refresh the configuration
* Fix multiple compile bugs in the patches
* Only add own ag71xx files for kernel 4.19 and use upstream version for
5.4.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This new function make batch processing of network packets possible,
which slightly improves performance.
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
Tested-by: Rosen Penev <rosenp@gmail.com>
This commit adds a workaround for the loss of the SGMII link observed on
the QCA955x generation of SoCs. The workaround originates part from the
U-Boot source code, part from the implementation from AVM found in the
GPL tarball for the AVM FRITZ!WLAN Repeater 450E.
The bug results in a stuck SGMII link between the PHY device and the SoC
side. This has only been observed with the Atheros AR8033 PHY and most
likely all devices using such combination are affected.
It is worked around by reading a hidden SGMII status register and
issuing a SGMII PHY reset until the link becomes useable again.
Signed-off-by: David Bauer <mail@david-bauer.net>
This patch contains updated driver for Atheros NAND Flash Controller
written originally by Gabor Juhos for ar71xx (aka 'ar934x-nfc').
ath79 version has adapted to work with kernel 4.19 and Device Tree.
It has also been renamed to 'ar934x-nand' to avoid confusion with
Near-Field Communication technology.
Controller is present on Atheros AR934x SoCs and required for accessing
internal flash storage on routers like Netgear WNDR4300.
This port preserves all NAND programming code while moving platform
configuration to Device Tree and replacing some kernel functions marked
for retirement by 4.19.
Suitable definition is included in 'ar934x.dtsi' ('nand@1b000200' section).
Most important changes to ar71xx version are:
* old kernel sections of code removed
* 'bool swap_dma' provided by platform data is now set by boolean DT
property 'qca,nand-swap-dma'
* board-supplied (mach-*.c code) platform data removed - its elements
become either unused, redundant or replaced by DT methods (like reset)
* IRQ is reserved by devm_request_irq() so free_irq() is not needed anymore
* calls to deprecated nand_scan_ident() + nand_scan_tail() function pair
replaced by using recommended nand_scan() with attach_chip() callback
* ECC is set to hardware by default, can be overriden by standard DT
'nand-ecc-*' properties (software Hamming or BCH are other options)
This driver has been successfully tested on Netgear WNDR4300 running
experimental ath79 OpenWrt master branch.
Signed-off-by: Michal Cieslakiewicz <michal.cieslakiewicz@wp.pl>
[add reset control]
Signed-off-by: David Bauer <mail@david-bauer.net>
tx_size was just declared above and set to BIT(tx->order)
Use the declaration instead, which could avoid a pointer deref
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
All other instances of this identical declaration fetch the
value directly from the ring_order.
Also do it here.
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Upstream commit 246902bdf562d45ea3475fac64c93048a7a39f01
Which contains following explanation:
--
There is no need to use GFP_ATOMIC here, GFP_KERNEL should be enough.
The 'kcalloc()' just a few lines above, already uses GFP_KERNEL.
--
Looking at the code, all other descriptors also use plain GFP_KERNEL
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
The ag71xx code did not include a case for the "rgmii-id" PHY mode in the
code. There are devices that need this mode, so I'm adding it.
Signed-off-by: Daniel Gimpelevich <daniel@gimpelevich.san-francisco.ca.us>
gmac0 may need a phy on builtin switch, which can be unavailable
if gmac0 is probed before builtin switch.
Return -EPROBE_DEFER in this case so that gmac0 can be probed
later.
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
ethtool doesn't work currently as phy_ethtool_ioctl expects user space
pointer, but it's being passed kernel one. Fixing it doesn't make sense
as {s,g}et_settings were deprecated anyway. So let's rather remove
phy_ethtool_ioctl and use new {s,g}et_link_ksettings instead. While at
it, update nway_reset as well.
Cc: John Crispin <john@phrozen.org>
Ref: https://bugs.openwrt.org/index.php?do=details&task_id=1982
Signed-off-by: Petr Štetiar <ynezz@true.cz>
This ioctl is currently routed through generic interface code:
dev_ioctl
dev_ethtool
__ethtool_get_link_ksettings
phy_ethtool_ioctl
Cc: John Crispin <john@phrozen.org>
Cc: Chuanhong Guo <gch981213@gmail.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
In ar71xx we check for stuck DMA on devices which fall in the is_ar724x
SoC group (ar724x, ar933x, ar934x, qca9533, tp9343, qca955x, qca956x).
In ath79 we're currently performing this check only for devices with
ar7240 SoC, so this patch tries to sync the dma stuck checking behavior
with what is being done in ar71xx.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
linux 4.19 doesn't accept a NULL device for these functions.
It also complains that the device struct in net_device doesn't have
a dma_mask set.
Pass the device struct from platform_device for these functions.
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
Kernel newer than 4.15 dropped "data" field and used from_timer
to cast out the parent struct pointer for current timer.
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
It is unused by default and upstream is trying to remove it as it has
negative effects when the driver is under load. Upstream explanation:
netpoll: avoid capture effects for NAPI drivers
As diagnosed by Song Liu, ndo_poll_controller() can
be very dangerous on loaded hosts, since the cpu
calling ndo_poll_controller() might steal all NAPI
contexts (for all RX/TX queues of the NIC).
This capture, showing one ksoftirqd eating all cycles
can last for unlimited amount of time, since one
cpu is generally not able to drain all the queues under load.
It seems that all networking drivers that do use NAPI
for their TX completions, should not provide a ndo_poll_controller() :
Most NAPI drivers have netpoll support already handled
in core networking stack, since netpoll_poll_dev(
uses poll_napi(dev) to iterate through registered
NAPI contexts for a device.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
While converting Nanostation M XW from current ar71xx code to ath79 I've
hit one issue, where the ethernet networking wasn't working, so I was
checking every bit in the networking setup path between ar71xx and
ath79.
I've came to the following code in ar71xx/mach-ubnt-xm.c:
static void __init ubnt_xw_init(void) {
...
ath79_setup_ar934x_eth_cfg(AR934X_ETH_CFG_MII_GMAC0 |
AR934X_ETH_CFG_MII_GMAC0_SLAVE);
...
}
Where this code is setting AR934X_ETH_CFG_MII_GMAC0_SLAVE bit in
AR934X_GMAC_REG_ETH_CFG register, but I couldn't find a way of setting
this bit from DTS, so this patch adds `mii-gmac0-slave` DTS property
which allows setting of this bit in `gmac-config`, which is then used in
Nanostation M XW DTS.
Tested-by: Joe Ayers <ae6xe@arrl.net>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
