- 07 Dec, 2018 2 commits
-
-
Chen-Yu Tsai authored
The RTC module on the H3 was claimed to be the same as on the A31, when in fact it is not. The A31 does not have an RTC external clock output, and its internal RC oscillator's average clock rate is not in the same range. The H5's RTC has some extra crypto-related registers compared to the H3. Their exact functions are not clear. Also the RTC-VIO regulator has different settings. This patch fixes the compatible string and clock properties to conform to the updated bindings. The device node for the internal oscillator is removed, as it is internalized into the RTC device. Clock references to the IOSC and LOSC are also fixed. Acked-by:
Maxime Ripard <maxime.ripard@bootlin.com> Tested-by:
Corentin Labbe <clabbe.montjoie@gmail.com> Signed-off-by:
Chen-Yu Tsai <wens@csie.org>
-
Chen-Yu Tsai authored
The RTC module on the A23 was claimed to be the same as on the A31, when in fact it is not. The A31 does not have an RTC external clock output, and its internal RC oscillator's average clock rate is not in the same range. The A33's RTC is the same as the A23. This patch fixes the compatible string and clock properties to conform to the updated bindings. The register range is also fixed. Acked-by:
-
- 06 Dec, 2018 2 commits
-
-
Chen-Yu Tsai authored
The R40 datasheet specifies a tolerance range for the external oscillators used. Add them to the device tree as the clock accuracy. Acked-by:
-
Chen-Yu Tsai authored
The H3 datasheet specifies a tolerance range for the external oscillators used. Add them to the device tree as the clock accuracy. The internal oscillator is left unchanged, as it will be removed later. Acked-by:
-
- 05 Dec, 2018 3 commits
-
-
Chen-Yu Tsai authored
The current oversampling rate of 512 means that for 48 kHz 16 bit stereo, the MCLK is running at the same rate as the module clock, so there is no head room to support higher sampling rates. The codec however supports up to 192 kHz for playback. This patch drops the oversampling rate from 512 to 128, so that 192 kHz audio can be played back directly without downsampling. Ideally we should be using different oversampling rates for different sampling rates, but that's not possible without a platform-specific machine driver. Fixes: 870f1bd1 ("ARM: dts: sun8i: Add audio codec, dai and card for A33") Signed-off-by:
-
Paul Kocialkowski authored
Just like on the A33, the video engine on the H3 can map any address in memory, so there is no particular need to have reserved memory at a fixed address. As a result, remove the reserved memory node and let the kernel allocate the CMA pool wherever it sees fit. Signed-off-by:
Paul Kocialkowski <paul.kocialkowski@bootlin.com> Acked-by:
-
Paul Kocialkowski authored
While we believed that the memory for the video engine had to be kept in the first 256 MiBs of DRAM, this is no longer true starting with the A33 and any address can be mapped. As a result, remove the reserved memory node and let the kernel allocate the CMA pool wherever it sees fit. Signed-off-by:
-
- 04 Dec, 2018 2 commits
-
-
Mesih Kilinc authored
Lichee Pi Nano is a F1C100s board by Lichee Pi. Add initial device tree for it. Signed-off-by:
Mesih Kilinc <mesihkilinc@gmail.com> Signed-off-by:
-
Mesih Kilinc authored
F1C100s is one product with the suniv die, which has a 32MiB co-packaged DDR1 DRAM chip. As we have the support for suniv pin controller and CCU now, add a initial DTSI for it. Signed-off-by:
-
- 30 Nov, 2018 1 commit
-
-
Oskari Lemmela authored
Add AC power supply subnode for AXP81X PMIC. Signed-off-by:
Oskari Lemmela <oskari@lemmela.net> Reviewed-by:
Quentin Schulz <quentin.schulz@bootlin.com> Reviewed-by:
Vasily Khoruzhick <anarsoul@gmail.com> Signed-off-by:
-
- 28 Nov, 2018 30 commits
-
-
Maxime Ripard authored
Our memory node will generate a warning in DTC since the unit address is not matching the reg property. However, that node will be created by the bootloader, so we can just remove it entirely in order to remove that warning. Signed-off-by:
-
Maxime Ripard authored
The MMC0 controllers have only one muxing option in the SoC. In such a case, we can just move the muxing into the DTSI, and remove it from the DTS. Signed-off-by:
-
Maxime Ripard authored
All our pinctrl nodes were using a node name convention with a unit-address to differentiate the different muxing options. However, since those nodes didn't have a reg property, they were generating warnings in DTC. In order to accomodate for this, convert the old nodes to the syntax we've been using for the new SoCs, including removing the letter suffix of the node labels to the bank of those pins to make things more readable. Signed-off-by:
-
Maxime Ripard authored
One of the usage of the LRADC is to implement buttons. The bindings define that we should have one subnode per button, with their associated voltage as a property. However, there was no reg property but we still used the voltage associated to the button as the unit-address, which eventually generated warnings in DTC. Rename the node names to avoid those warnings. Signed-off-by:
Maxime Ripard <maxime.ripard@free-electrons.com> Acked-by:
-
Maxime Ripard authored
Most of our device trees have had leading zeros for padding as part of the nodes unit-addresses. Remove all these useless zeros that generate warnings Signed-off-by:
-
Maxime Ripard authored
The i2c nodes were pre-populated to ease the use of overlays. However, now that we provide default muxing options for those nodes, the one in the DTS don't provide any content at all. Remove them. Signed-off-by:
-
Maxime Ripard authored
The I2C's and MMC0 controllers have only one muxing option in the SoC. In such a case, we can just move the muxing into the DTSI, and remove it from the DTS. Signed-off-by:
-
Maxime Ripard authored
Now that all the SoCs using the tablet reference design DTSI are using the same pinctrl naming scheme, we can move back the pinctrl phandles to the main DTSI. Signed-off-by:
-
Maxime Ripard authored
Some GPIO pinctrl nodes cannot be easily removed, because they would also change the pin configuration, for example to add a pull resistor or change the current delivered by the pin. Those nodes still have underscores and unit-addresses in their node names in our DTs, so adjust their name to remove the warnings. Use that occasion to also fix some poorly chosen node-names. Signed-off-by:
-
Maxime Ripard authored
All our pinctrl nodes were using a node name convention with a unit-address to differentiate the different muxing options. However, since those nodes didn't have a reg property, they were generating warnings in DTC. In order to accomodate for this, convert the old nodes to the syntax we've been using for the new SoCs, including removing the letter suffix of the node labels to the bank of those pins to make things more readable. Signed-off-by:
-
Maxime Ripard authored
Boards usually have an external pull-up on the card-detect signal, so there's no need to add another one. This also removes a DTC warning. Signed-off-by:
-
Maxime Ripard authored
The pin groups are supposed to be in alphabetical order, and they aren't. Fix this. Signed-off-by:
-
Maxime Ripard authored
One of the usage of the LRADC is to implement buttons. The bindings define that we should have one subnode per button, with their associated voltage as a property. However, there was no reg property but we still used the voltage associated to the button as the unit-address, which eventually generated warnings in DTC. Rename the node names to avoid those warnings. Signed-off-by:
-
Maxime Ripard authored
The gpio pinctrl nodes are redundant and as such useless most of the times. Since they will also generate warnings in DTC, we can simply remove most of them. Signed-off-by:
-
Maxime Ripard authored
Some boards override the MMC pin muxing settings in order to enable the pull-ups and change the drive strength to a value higher than the default. While this was needed in the earlier days, this is now the default setting for those pins, and therefore we don't need those board-specific settings anymore. Signed-off-by:
-
Maxime Ripard authored
The simple-framebuffer nodes have a unit address, but no reg property which generates a warning when compiling it with DTC. Change the simple-framebuffer node names so that there is no warnings on this anymore. Signed-off-by:
-
Maxime Ripard authored
Most of our device trees have had leading zeros for padding as part of the nodes unit-addresses. Remove all these useless zeros that generate warnings Signed-off-by:
-
Maxime Ripard authored
The #address-cells and #size-cells are only relevant for nodes that have childs with reg properties. Otherwise, DTC will emit a warning saying that those properties are unnecessary. Remove them when needed. Signed-off-by:
-
Maxime Ripard authored
DTC will emit a warning on our OPPs nodes for the common DTSI between the A23 and A33 since those nodes use the frequency as unit addresses, but don't have a matching reg property. Fix this by moving the frequency to the node name instead. Signed-off-by:
-
Maxime Ripard authored
Our main node for all the in-SoC controllers used to have a unit name. The unit-name, in addition to being actually false, would not match any reg property, which generates a warning. Remove it in order to remove those warnings. Signed-off-by:
-
Maxime Ripard authored
Using skeleton.dtsi will create a memory node that will generate a warning in DTC. However, that node will be created by the bootloader, so we can just remove it entirely in order to remove that warning. Signed-off-by:
-
Maxime Ripard authored
The #address-cells and #size-cells are only relevant for nodes that have childs with reg properties. Otherwise, DTC will emit a warning saying that those properties are unnecessary. Remove them when needed. Signed-off-by:
-
Maxime Ripard authored
Some boards override the MMC pin muxing settings in order to enable the pull-ups and change the drive strength to a value higher than the default. While this was needed in the earlier days, this is now the default setting for those pins, and therefore we don't need those board-specific settings anymore. Signed-off-by:
-
Maxime Ripard authored
The I2C and MMC controllers have only one muxing option in the SoC. In such a case, we can just move the muxing into the DTSI, and remove it from the DTS. Signed-off-by:
-
Maxime Ripard authored
Our HDMI output endpoint on the A10s DTSI has a warning under DTC: "graph node has single child node 'endpoint', #address-cells/#size-cells are not necessary". Fix this by removing those properties. Signed-off-by:
-
Maxime Ripard authored
Some GPIO pinctrl nodes cannot be easily removed, because they would also change the pin configuration, for example to add a pull resistor or change the current delivered by the pin. Those nodes still have underscores and unit-addresses in their node names in our DTs, so adjust their name to remove the warnings. Use that occasion to also fix some poorly chosen node-names. Signed-off-by:
-
Maxime Ripard authored
The SOM204-EVB doesn't use the CTS pin, and thus was defining its own pinctrl node for the UART3 muxing. Since we split away the TX and RX pin, we can use the global node now, and only have the RTS pin in our local node. Signed-off-by:
-
Maxime Ripard authored
Some UART nodes on the A20 DTSI do not share the same pattern that we use everywhere else, with the RTS and CTS pins split away from the TX and RX pins. Make those pin groups consistent with the rest of our DT. Signed-off-by:
-
Maxime Ripard authored
All our pinctrl nodes were using a node name convention with a unit-address to differentiate the different muxing options. However, since those nodes didn't have a reg property, they were generating warnings in DTC. In order to accomodate for this, convert the old nodes to the syntax we've been using for the new SoCs, including removing the letter suffix of the node labels to the bank of those pins to make things more readable. Signed-off-by:
-
Maxime Ripard authored
Some gpio-keys definitions in our DTs were having buttons defined with a unit-address and that would generate a DTC warning. Change the buttons node names to remove the warnings. Signed-off-by:
-
