Since the nvmem-based approach for retrieving MAC addresses
appears to depend on the addresses being set up after the
partitions, it is no longer possible to keep the MAC address
setup in shared DTSI files while the partitions itself are
set up in DTS files for the individual devices.
In ath79 the firmware partition is typically located somewhere
"in the middle" of the partition table. Thus, it's not trivial
to share the partitions containing MAC address information in
a common DTSI (like we did in some cases on ramips).
In this commit, MAC address setup is thus moved to the relevant
partitions, and in most cases needs to be duplicated. While
the duplication is not really nice, it eventually provides a
cleaner and more tidy setup, making the DTS(I) file
fragmentation a bit more logical. This should also help
with adding new devices, as information is distributed across
less locations.
For consistency, this commit also moves the mtd-cal-data property
"down" together with the MAC address setup, so it's not based
on a partition before the latter is defined either. (This is
only done for those files touched due to nvmem conversion.)
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The uart node is enabled on all devices except one (GL-USB150 *).
Thus, let's not have a few hundred nodes to enable it, but do not
disable it in the first place.
Where the majority of devices is using it, also move the serial0
alias to the DTSI.
*) Since GL-USB150 even defines serial0 alias, the missing uart
is probably just a mistake. Anyway, disable it for now so this
patch stays cosmetic.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Currently, we request LED labels in OpenWrt to follow the scheme
modelname:color:function
However, specifying the modelname at the beginning is actually
entirely useless for the devices we support in OpenWrt. On the
contrary, having this part actually introduces inconvenience in
several aspects:
- We need to ensure/check consistency with the DTS compatible
- We have various exceptions where not the model name is used,
but the vendor name (like tp-link), which is hard to track
and justify even for core-developers
- Having model-based components will not allow to share
identical LED definitions in DTSI files
- The inconsistency in what's used for the model part complicates
several scripts, e.g. board.d/01_leds or LED migrations from
ar71xx where this was even more messy
Apart from our needs, upstream has deprecated the label property
entirely and introduced new properties to specify color and
function properties separately. However, the implementation does
not appear to be ready and probably won't become ready and/or
match our requirements in the foreseeable future.
However, the limitation of generic LEDs to color and function
properties follows the same idea pointed out above. Generic LEDs
will get names like "green:status" or "red:indicator" then, and
if a "devicename" is prepended, it will be the one of an internal
device, like "phy1:amber:status".
With this patch, we move into the same direction, and just drop
the boardname from the LED labels. This allows to consolidate
a few definitions in DTSI files (will be much more on ramips),
and to drop a few migrations compared to ar71xx that just changed
the boardname. But mainly, it will liberate us from a completely
useless subject to take care of for device support review and
maintenance.
To also drop the boardname from existing configurations, a simple
migration routine is added unconditionally.
Although this seems unfamiliar at first look, a quick check in kernel
for the arm/arm64 dts files revealed that while 1033 lines have
labels with three parts *:*:*, still 284 actually use a two-part
labelling *:*, and thus is also acceptable and not even rare there.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The "/dts-v1/;" identifier is supposed to be present once at the
top of a device tree file after the includes have been processed.
In ath79, we therefore requested to have in the DTS files so far,
and omit it in the DTSI files. However, essentially the syntax of
the parent ath79.dtsi file already determines the DTS version, so
putting it into the DTS files is just a useless repetition.
Consequently, this patch puts the dts-v1 statement into the parent
ath79.dtsi, which is (indirectly) included by all DTS files. All
other occurences are removed.
Since the dts-v1 statement needs to be before any other definitions,
this also moves the includes to make sure the ath79.dtsi or its
descendants are always included first.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
All definitions of gpio in SoC DTSI files do not set status, i.e.
have it enabled. This drops all remaining redundant "status = okay"
definitions in descendent files (mostly older ones).
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
In ath79, for several SoCs the console bootargs are defined to the
very same value in every device's DTS. Consolidate these definitions
in the SoC dtsi files and drop further redundant definitions elsewhere.
The only device without any bootargs set has been OpenMesh OM5P-AC V2.
This will now inherit the setting from qca955x.dtsi
Note that while this tidies up master a lot, it might develop into a
frequent pitfall for backports.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The "/dts-v1/;" identifier is supposed to be put once at the beginning
of a device tree file. Thus, it makes no sense to provide it a second
time in to-be-included DTSI files.
This removes the identifier from all DTSI files in /target/linux.
Most of the DTS files in OpenWrt do contain the "/dts-v1/;". It is
missing for most of the following targets, though:
mvebu, ipq806x, mpc85xx, ipq40xx
This does not touch ipq806x for now, as the bump to 4.19 is close.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This applies several style adjustments that have been requested in
recent reviews to older DTS files. Despite making the code base more
consistent, this will also help to reduce review time when DTSes
are copy/pasted.
Applied changes:
- Rename gpio-keys/gpio-leds to keys/leds
- Remove node labels that are not used
- Use label property for partitions
- Prefix led node labels with "led_"
- Remove redundant includes
- Harmonize new lines after status property
- Several smaller style fixes
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This patch adds the label MAC address for several devices in
ath79.
Some devices require setting the MAC address in 02_network:
For the following devices, the netif device can be linked in
device tree, but the MAC address cannot be read:
- alfa-network,ap121f
- avm,fritz300e
- ubnt-xm devices
For the following devices, label MAC address is tied to lan or
wan, so no node to link to exists in device tree:
- adtran,bsap1800-v2
- adtran,bsap1840
- dlink,dir-842-c1/-c2/-c3
- engenius,ecb1750
- iodata,etg3-r
- iodata,wn-ac1167dgr
- iodata,wn-ac1600dgr
- iodata,wn-ac1600dgr2
- iodata,wn-ag300dgr
- nec,wg800hp
- nec,wg1200cr
- trendnet,tew-823dru
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Commit "generic: ar8216: add mib_poll_interval switch attribute" sets
mib-poll-interval as disabled by default (was set to 2s), so it makes
switch LEDs trigger disfunctional on devices which don't have
mib-poll-interval set.
So this patch sets mib-poll-interval to 500ms on devices which have
ar83xx switch connected to mdio0 bus, as the same value was set for
built in switches in 443fc9ac35 ("ath79: use ar8216 for builtin
switch").
Some measurements performed on TP-Link Archer C7-v5:
mib-type=0, mib-poll-interval=500ms (10s pidstat)
Average: %usr %system %guest %wait %CPU CPU Command
Average: 0.00 1.93 0.00 0.00 1.93 - kworker/0:2
iperf3 (30s): 334 Mbits/sec
mib-type=0, mib-poll-interval=2s (10s pidstat)
Average: %usr %system %guest %wait %CPU CPU Command
Average: 0.00 1.14 0.00 0.00 1.14 - kworker/0:2
iperf3 (30s): 334 Mbits/sec
So it seems like we get 4x faster LED refresh rate for additional 0.8%
CPU load.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
>From the Documentation/devicetree/bindings/leds/common.txt:
- default-state : The initial state of the LED. Valid values are "on", "off",
and "keep". If the LED is already on or off and the default-state property is
set the to same value, then no glitch should be produced where the LED
momentarily turns off (or on). The "keep" setting will keep the LED at
whatever its current state is, without producing a glitch. The default is
off if this property is not present.
So setting the default-state of the LEDs to `off` is redundant as `off`
is default LED state anyway. We should remove it as almost every new
PR/patch submission contains this property by default which seems to be
just copy&paste from some DTS file already present in the tree.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Parsing "firmware" partition (to create kernel + rootfs) was implemented
using OpenWrt downstream code enabled by CONFIG_MTD_SPLIT_FIRMWARE. With
recent upstream mtd changes we can do it in a more clean way for DTS
targets. It just requires adding a proper "compatible" string to the
"firmware" partition node.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Use diag.sh version used for apm821xx, ipq40xx and ipq806x, which
supports different leds for the different boot states.
The existing led sequences should be the same as before.
Signed-off-by: Dmitry Tunin <hanipouspilot@gmail.com>
[reword commit message]
Signed-off-by: Mathias Kresin <dev@kresin.me>
Fix all issues found by the devicetree compiler like wrong address/size
cells as well as wrong/missing/superfluous unit addresses.
Signed-off-by: Mathias Kresin <dev@kresin.me>
