arm64: dts: rockchip: overlay: yy3588: disable eDP overlay temporarily#1
Closed
SuperKali wants to merge 32 commits into
Closed
arm64: dts: rockchip: overlay: yy3588: disable eDP overlay temporarily#1SuperKali wants to merge 32 commits into
SuperKali wants to merge 32 commits into
Conversation
- this never worked, except for `123456` display on boot - was always built-in (not a module, and =y always) - This reverts commit 5e14078
- Original from armbian/build@5ee03af - backported from 7.0.y mainline version to 6.1.y - `.remove` returns void at 6.1.y The Vinka VK2C21 (and pin-compatible VK2C21A/VK2C21B) is a memory-mapping segment LCD controller that communicates over I2C. It can drive up to 128 segments (e.g. 32 SEG x 4 COM in 1/3 bias mode). Some SBCs wire this chip to GPIO pins without I2C pull-up resistors, making the standard i2c-gpio (open-drain) driver unusable. This driver therefore bit-bangs I2C in push-pull mode directly via GPIO, requiring no external pull-ups. Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
…ave it - also eat trailing whitespace
- backported driver which works - also eat trailing whitespace
… ipex. Currently the luckfox-lyra-zero-w defaults to using the ipex connector for wifi (not the onboard antenna). This has lead to user confusion. Use a gpio hog to set the antenna selection pin HIGH at boot, selecting the onboard by default. Add an overlay for switching to the ipex connector.
Adds spidev overlays for `ebyte-ecb41-pge` compatible with raspberry pi 40pin layout. Also increases CMA for Ebyte-ECB41-PGE from 4MB to 8MB to correct current CMA allocation errors when enabling these overlays + ethernet at the same time.
… (RK3588) board support (armbian#486) * easepi-a2, easepi-r2: add device tree support for vendor kernel 6.1 Add board device tree files for EasePi A2 (RK3568) and EasePi R2 (RK3588): - rk3568-easepi-a2.dts: Board DTS for EasePi A2, includes shared carrier board definition rk3568-roc-k40pro.dtsi - rk3568-roc-k40pro.dtsi: Shared hardware config for the ROC-RK3568-PC carrier platform (PMIC, regulators, GPIO, etc.) - rk3588-easepi-r2.dts: Board DTS for EasePi R2, self-contained with includes for rk3588-rk806-single.dtsi and rk3588-linux.dtsi Signed-off-by: ifroncy01 <ifroncy01@github.com> * easepi-a2: Merge roc-k40pro.dtsi into standalone DTS Combine rk3568-roc-k40pro.dtsi carrier board config into rk3568-easepi-a2.dts, making it self-contained: - Merge full root node hardware description (PMIC, regulators, leds) - Apply easepi-a2 specific root node properties - Replace ethernet aliases to use pcie_eth0 - Append all easepi-a2 label overrides (gmac0/1 disabled, pcie enabled, etc.) - Remove rk3568-roc-k40pro.dtsi as it is no longer needed * easepi-a2: fix DTC pci_device_reg warning Change PCIe bridge node address from pcie@10 to pcie@10,0 to follow standard PCI device tree address format (device,function). * easepi-a2: fix pci_device_reg DTS warning Change PCIe bridge node address from pcie@10 to pcie@10,0 to follow standard PCI device tree address format (device,function). Fixes a DTS warning: Warning (pci_device_reg): /pcie@fe260000/pcie@10: PCI unit address format error, expected "0,0" * easepi: various DTS fixes and cleanups For easepi-a2: - Fix adc-keys press-threshold-microvolt from 1750 to 175000 - Disable rk809-sound (audio card not present on board) - Formatting fix on i2s1_8ch pinctrl indentation - Translate Chinese comments to English For easepi-r2: - Fix compatible string to "easepi,r2" - Fix comment formatting and translate Chinese comments to English - Fix sdio status placement * easepi-a2: DTS fix --------- Signed-off-by: ifroncy01 <ifroncy01@github.com>
… OPP tables (armbian#484) * arm64: dts: rockchip: rk3576: remove empty rockchip,supported-hw from OPP tables The bare "rockchip,supported-hw;" property with no value makes the OPP framework require opp-supported-hw on every OPP entry, but only opp-j-m-* entries have it. This causes all generic OPP entries to be rejected, resulting in GPU stuck at a fixed ~100 MHz without devfreq. Remove the empty property so standard opp-* entries are always available for non-J/M chips. Affects: cluster0, cluster1, bus_a72, dmc, npu, gpu, vop OPP tables. Signed-off-by: mingzq <north_sea@qq.com> * arm64: dts: rockchip: keep rk3576 supported-hw outside GPU --------- Signed-off-by: mingzq <north_sea@qq.com>
PR armbian#486 is based on the v1.0 board, but the actual board sold is v1.2. Changes in v1.2: 1. Remove power supply control for the TF card to support booting from the TF card. 2. Remove power supply control for the PI6C clock, now it's directly powered. 3. Export the I2C7 pinout. Signed-off-by: Liangbin Lian <jjm2473@gmail.com>
…mputer devkits (armbian#483) The volumeup-key press-threshold-microvolt was set to 1750 instead of 1750000 (1.75V), off by a factor of 1000. Fix for both rk3576 and rk3588 recomputer devkit boards.
…i5U (armbian#488) * media: rockchip: hdmirx: add edid version to disable NV24 Add two new EDID variants (HDMIRX_EDID_340M_NV12 = 3 and HDMIRX_EDID_600M_NV12 = 4) that omit the YCbCr 4:4:4 (NV24) capability bit and let the driver fix up the EDID checksum accordingly. Read the selected variant from the new `edid-version` DT property on the hdmirx_ctrler node so per-board defaults can be chosen without requiring userspace to poke the edid sysfs file at boot. This is a direct prerequisite for board DTS files that want to advertise a default EDID other than the hard-coded 340MHz variant. Signed-off-by: jensen <jensenhuang@friendlyarm.com> [smazurov@gmail.com: cherry-picked unmodified from friendlyarm/kernel-rockchip commit ee391fb22 on branch nanopi6-v6.1.y. Applies cleanly on top of armbian/linux-rockchip rk-6.1-rkr5.1 (with offsets, due to subsequent unrelated drift). Tested on Orange Pi 5 Ultra (RK3588) with both 1080p60 and 4K60 HDMI sources; required by the companion DTS patch which sets edid-version = <2> for OPi5U.] Tested-by: Stepan Mazurov <smazurov@gmail.com> Signed-off-by: Stepan Mazurov <smazurov@gmail.com> * arm64: dts: rockchip: opi5-ultra: default hdmirx edid-version to 600M Set edid-version = <2> (HDMIRX_EDID_600M) on the hdmirx_ctrler node so the default EDID advertised to upstream HDMI sources includes the TMDS-up-to-600MHz block (covering 4K@60 modes), instead of the driver's built-in HDMIRX_EDID_340M default which advertises only up to TMDS 340MHz (1080p60 / 4K@30). Symptom without this property on Orange Pi 5 Ultra: a 4K-capable HDMI source (tested with a Linux desktop pushing 3840x2160@60 from a discrete GPU) cycles indefinitely between TMDS-valid and TMDS-invalid states. The hdmirx PHY rejects the source's 4K stream because the EDID the source read did not advertise 4K support, yet the source insists on 4K based on its own display configuration. Writing "2" to /sys/devices/platform/fdee0000.hdmirx-controller/hdmirx/hdmirx/edid at runtime resolves the symptom immediately and stably; this DTS change makes the runtime fix the default at boot. The Rock 5B+ and NanoPC-T6 board DTS files already set similar edid-version properties; Orange Pi 5 Ultra was missing the equivalent. Depends on the rk_hdmirx driver reading edid-version from DT (prerequisite patch: cherry-pick of friendlyarm/kernel-rockchip commit ee391fb22 in the companion patch). Tested-by: Stepan Mazurov <smazurov@gmail.com> Signed-off-by: Stepan Mazurov <smazurov@gmail.com> --------- Signed-off-by: jensen <jensenhuang@friendlyarm.com> Signed-off-by: Stepan Mazurov <smazurov@gmail.com> Co-authored-by: jensen <jensenhuang@friendlyarm.com>
…n#490) Signed-off-by: mingzq <north_sea@qq.com>
which cause initramfs load rootfs failed Signed-off-by: Liangbin Lian <jjm2473@gmail.com>
Signed-off-by: Liangbin Lian <jjm2473@gmail.com>
Signed-off-by: Liangbin Lian <jjm2473@gmail.com>
Signed-off-by: Liangbin Lian <jjm2473@gmail.com>
Read r8125 LED register/value pairs from the PCIe controller DT "realtek,led-data" property, instead of hardcoding in the shared driver path. Walk up the device parent chain to find the PCIe platform device's of_node, since PCI devices on Rockchip platforms do not have their of_node set directly. Parse the DT property once during probe and cache it in the driver private struct. Apply the cached config from rtl_hw_start_8125_common only, eliminating redundant calls and repeated memory allocation.
Signed-off-by: Ken Wang <ken@radxa.com>
Signed-off-by: Ken Wang <ken@radxa.com>
The 7" 1024x600 DSI kit on Youyeetoo R1 and YY3588 is a TPC2059 panel whose GT911 needs the vendor cfg blob, and the 11.6" eDP kit ships a GT928 wired portrait. Add both blobs from the youyeetoo yy3588_linux tree as new tp-size cases 2059 and 928 so existing boards keep their current behavior. Signed-off-by: SuperKali <hello@superkali.me>
Vendor evb7-v11 wires HDMI-RX hotplug detect on GPIO1_D5, not GPIO1_C6. GPIO1_C6 is the pwm15m2 pad used by the eDP kit backlight, so keeping detect there breaks the eDP overlay pinmux. Signed-off-by: SuperKali <hello@superkali.me>
tp-size 911 programs the WGJ89006B 4096x4096 blob, which does not match the 7" TPC2059 sensor matrix and leaves dead zones at the panel edges. Switch both DSI overlays to the new 2059 case that sends the vendor 1024x600 blob with 1:1 coordinates. Signed-off-by: SuperKali <hello@superkali.me>
…lays Wire the three official Youyeetoo display kits from the vendor rk3588-lcd-yyt.dtsi: 7" 1024x600 TPC2059 panels on dsi0 and dsi1 and the 11.6" 1920x1080 eDP panel with GT928 touch. The previous dsi1 overlay described a non-existent Innolux panel. eDP runs on VP1 because dp0 already owns VP2 on this board. Signed-off-by: SuperKali <hello@superkali.me>
The 7" kit backlight driver is active-low: with normal polarity raising brightness dims the panel, confirmed on R1 hardware. The vendor Android DTB programs the same PWM inverted. Flip polarity on all four DSI overlays; the eDP backlight stays normal per vendor. Signed-off-by: SuperKali <hello@superkali.me>
eDP on VP1 fails: VP1 shares its PLL with VP0 so the 141.5MHz dclk gets clamped to 132MHz and the VOP2 port mux times out on the second modeset. Move eDP to VP2 as the vendor does and hand DP VP1 instead. Use the BOE QV116FHB-N81 datasheet typical timing and drop the panel enable GPIO, GPIO0_C4 is already held high by the always-on wifi_disable regulator in the board DTS. Signed-off-by: SuperKali <hello@superkali.me>
…lays With the vendor kernel the display-subsystem bind stalls at boot when a display overlay is enabled without its panel attached, on eDP the forced HPD keeps link training against a missing sink. State the constraint in the overlay descriptions. Signed-off-by: SuperKali <hello@superkali.me>
The 11.6 inch eDP kit panel currently available shows severe image retention and can hang the display-subsystem bind as it degrades, so the overlay cannot be validated on known-good hardware right now. Rename it to .disabled and drop it from the overlay Makefile so it is not built or offered until a working panel is available to re-verify the path. Signed-off-by: SuperKali <hello@superkali.me>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
10 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The 11.6 inch eDP kit panel we have shows severe image retention that gets worse with use, and as it degrades it can hang the display-subsystem bind at boot. That leaves the eDP overlay impossible to validate on a known-good panel right now.
This renames the overlay to .disabled and drops it from the overlay Makefile so it is not built or offered until a working panel is available to re-verify the path. The DSI overlays are untouched and keep working.
The pending backlight-PWM and regulator fixes for this overlay are parked and will land together with the re-enable once the eDP path can be tested on good hardware.