SBC News

rockchip-rk3588: add -Wno-error=address to u-boot CFLAGS so it builds with gcc 12+ rockchip-rk3588: add -Wno-error=address to u-boot CFLAGS so it builds with gcc 12+ this applies to all u-boots in this family this is in addition to the the -Wno-error=enum-int-mismatch needed for gcc 13+ View the full article

rockchip-rk3588: add -Wno-error=enum-int-mismatch to u-boot CFLAGS so it builds with gcc 13+ rockchip-rk3588: add -Wno-error=enum-int-mismatch to u-boot CFLAGS so it builds with gcc 13+ this applies to all u-boots in this family View the full article

Description Hinlink H88K is a rk3588 board with a lot of IO: https://doc.hinlink.cn/ I've been building personal images: https://github.com/amazingfate/armbian-h88k-images/ I just use the same uboot defconfig with evb uboot devicetree. I may patch the evb device tree in the future since uboot patch is board specific. Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [ ] I have made corresponding changes to the documentation [x] My changes generate no new warnings [ ] Any dependent changes have been merged and published in downstream …

Description Improve fan curve from rockchip default via device-tree overlay Better operating temperatures No discernible noise during standard workloads Low always-on rotation (make sure it's not too low to buzz) Credits: incognito Source: Radxa Forum How Has This Been Tested? [x] Built and enabled on Radxa Rock 5B ${BRANCH}="current" Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] My changes generate no new warnings View the full article

Description Enable proper enumeration of usb-bluetooth modules via M2/NGFF slot - without this patch, no pcie-integrated usb-module would wourk as device-tree lacks rfkill-bt device node How Has This Been Tested? [x] Applied patch to BOARD=rock-5b BRANCH=current, got lsusb output Bus 004 Device 004: ID 0bda:b85b Realtek Semiconductor Corp. Bluetooth Radio [x] Bluetoothctl recognized bluetooth controller [x] Succesfully paired to a bluettoth device (iPhone) Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] My changes generate no…

Description I guess all these BOOTFS_TYPE come from https://github.com/armbian/build/pull/3874. Rock3a and other rk356x boards uses an old version of vendor u-boot: https://github.com/radxa/u-boot/tree/stable-4.19-rock3, which has bad support for ext4 filesystem. While rk3588 uses new version of vendor u-boot which doesn't have ext4 partition issue, so I just drop all this config from rk3588 boards. Command to query boards with this config: grep -rn rockchip-rk3588 config/boards/|awk -F ':' '{print $1}'|xargs grep BOOTFS_TYPE Here are the vendor uboot source of these boards: board vendor uboot source armsom-sige7 armsom-w3 nanopc-cm3588-nas nanopct6 nanopi-r6s r…

Description Add panthor support to rockchip-rk3588-edge with panthor v6 patch. [x] gpu node is enabled for rock5b by collabora's patch [x] ubuntu user can use mesa from oibaf's ppa How Has This Been Tested? Please describe the tests that you ran to verify your changes. Please also note any relevant details for your test configuration. [x] ./compile.sh kernel BOARD=rock-5b BRANCH=edge DEB_COMPRESS=xz GHCR_MIRROR=dockerproxy GITHUB_MIRROR=ghproxy KERNEL_GIT=shallow [x] panthor works on rock5b Checklist: Please delete options that are not relevant. [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [ ] I have c…

Description Add v3 patch of rkvdec2. Add a v4l2 patch to fit rkvdec2's 16x16 minnimal decoding size for chromium. Remove VDPU121 node because there is already a 4K H264 decoder. Gstreamer can't handle two. Add udev rule for chromium How Has This Been Tested? Please describe the tests that you ran to verify your changes. Please also note any relevant details for your test configuration. [x] ./compile.sh kernel BOARD=rock-5b BRANCH=edge DEB_COMPRESS=xz KERNEL_CONFIGURE=no KERNEL_GIT=shallow [x] gst-launch-1.0 --no-fault filesrc location=/home/jfliu/Big_Buck_Bunny_2160_10s_H264_1s.mp4 ! parsebin ! v4l2slh264dec ! autovideosink [x] chromium can decode 4K h264 video. Chec…

Description Upstreaming patch: https://patchwork.kernel.org/project/linux-rockchip/patch/20240430024002.708227-3-liujianfeng1994@gmail.com/ The current fdb50400 won't work. there are errors in dmesg when doing decoding. How Has This Been Tested? Please describe the tests that you ran to verify your changes. Please also note any relevant details for your test configuration. [x] ./compile.sh kernel BOARD=rock-5b BRANCH=edge DEB_COMPRESS=xz KERNEL_CONFIGURE=no KERNEL_GIT=shallow Checklist: Please delete options that are not relevant. [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [ ] I have commented my code, …

Description Last time when I update hdmi driver patches, I drooped this commit from collabora because it will break the system. Now I add it back with other necessary patches. Now hdmi should support more display modes. How Has This Been Tested? Please describe the tests that you ran to verify your changes. Please also note any relevant details for your test configuration. [x] ./compile.sh kernel BOARD=rock-5b BRANCH=edge DEB_COMPRESS=xz KERNEL_GIT=shallow [x] HDMI works on rock5b. I don't have monitors with resolution like 1440p to test, but this will not break current status. Checklist: Please delete options that are not relevant. [x] My code follows the style gui…

Description OrangePi 5 Plus Bump current and edge to mainline 2024.10-rc3 ensure ARMV8_CRYPTO is disabled (currently broken) enable creation of /chosen/kaslr-seed node (to enable Kernel address space layout randomization) How Has This Been Tested? [x] Built, installed, and succesfully booted trixie current and edge Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] My changes generate no new warnings View the full article

Description Radxa Rock 5B Bump current and edge to mainline 2024.10-rc3 ensure ARMV8_CRYPTO is disabled (currently broken) enable creation of /chosen/kaslr-seed node (to enable Kernel address space layout randomization) How Has This Been Tested? [x] Built, installed, and succesfully booted trixie current and edge Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] My changes generate no new warnings View the full article

rockchip-rk3588/6.11: backport Marcin Juszkiewicz's "FriendlyELEC NanoPC-T6 improvements" v6 series to fix non-LTS and split LTS version of NanoPC T6 nanopct6-lts: a separate board for the LTS version of the NanoPC T6 Split has landed upstream for 6.12, and backported in a separate commit I had already added the -lts version to the vendor (6.1) rockchip kernel a few months ago rockchip-rk3588/6.11: backport Marcin Juszkiewicz's "FriendlyELEC NanoPC-T6 improvements" v6 series to fix non-LTS and split LTS version of NanoPC T6 Marcin Juszkiewicz's series https://lore.kernel.org/all/20240829-friendlyelec-nanopc-t6-lts-v6-0-edff247e8c02@linaro.org/ has landed upstream se…

Description enable CONFIG_RTW89_8852BE=m to take advantage of https://github.com/armbian/linux-rockchip/pull/192 GitHub issue reference: https://github.com/armbian/linux-rockchip/pull/192 How Has This Been Tested? Please describe the tests that you ran to verify your changes. Please also note any relevant details for your test configuration. [x] built kernel&image for opi5plus - 8852BE chipset and bluetooth are recognized and working OOB Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] My changes generate no new warnings View the full article

Description As per subject. Kernel configs have been updated and patches have been adapted to kernel 6.6 as well Jira reference number AR-1905 How Has This Been Tested? [x] Compiled rockchip 32 bit edge kernel succesfully [x] Compiled rk322x edge kernel successfully Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description As specified in PR https://github.com/armbian/build/pull/5967, there was a nasty mistake in rockchip and rk322x source files that were attempting to create two groups (gpio and i2c) on the building host instead of the built image. This caused the build system to crash when the groups were already present on the building host, which is totally not desiderable, and the intended feature was also broken. After some inspection, the groups were leveraged by a couple of udev rules to allow non-root users have access to gpio and i2c resources out of the box. This PR fixes the group creation on the target built image. Note: gpio and i2c gids starts from 900 because g…

This PR enables kernel modules for various USB Gadget applications for rockchip 32 bit platforms (rk322x, rk3288) for both current 6.6 and edge 6.7 kernels. Also provides a bunch of patches to address issues with rk322x, whose peripheral mode does not really well with dwc2 driver yet. A couple of patches are self-produced after trial-and-error sessions, another pair are imported from rockchip 5.10 kernel and substitutes an older patch with a bit more refined code (see the "debounce" patch). edit: the patches addressing the rk322x issues probably are worth also for rk3318 and rk3328, since the controller is the very same with similar issues. During my tests, I quickly ch…

Description As per subject: edge kernel for rockchip 32 bit targets has been bumped to 6.10 Also updates mainline-kernel.conf GitHub issue reference: Jira reference number AR-2427 How Has This Been Tested? [x] Compile deb kernel packages for rockchip family [x] Live testing on rk322x system [ ] Live testing on rk3288 Asus tinkerboard Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas View the full article

Description Move rockchip 32 bit edge kernel to 6.2. Kernel works fine, but there are some kernel "error" messages that needs to be investigated. Nonetheless the base system seems to work pretty fine Jira reference number AR-1636 How Has This Been Tested? [x] Full image has been built and tested on Asus Tinkerboard Rev 1.01 Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in d…

Description Bump rockchip-edge to kernel 6.3 Jira reference number AR-1674 How Has This Been Tested? [x] Kernel patches apply correctly Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description As per subject, just bump rockchip 32 bit edge kernel to kernel 6.5, adjusting the patches and dts new directory layout Jira reference number AR-1871 How Has This Been Tested? [x] Debian bookworm image built and tested on xt-q8l-v10 board Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description As per subject, bump edge kernel to 6.8 for rockchip 32 bit family. Also include a patch to handle an apparently missing Makefile in tools/cgroup directory which makes the compilation process crash when cleaning the tools kernel subdirectory. In case other families are bumped to kernel 6.8, it should be copied as well or moved into misc directory and applied somewhere until kernel people fix that. Jira reference number AR-2111 How Has This Been Tested? [x] Kernel .deb packages compiled and tested on a rk322x-box live system Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have co…

How Has This Been Tested? [*] Turn this option off: sata hard drives are not recognized [*] Enable this option: Correctly identify sata drives Everything else is exactly the same Checklist: [*] I have performed a self-review of my own code [*] My changes generate no new warnings View the full article

Description A user from the forum reports that a new rockchip rk3288-based board Tinkerboard S R2.0 has no wifi (see discussion here). After some talk, it seems that this board revision uses a Realtek rtl8723ds chip, whereas the R1.01 board uses rtl8723bs. This PR enabled the rtl8723ds wifi module that curiously was previously enabled in kernel 5.15 but somehow has been disabled in the advancement to 6.1. Current and edge kernel on rockchip family are both on 6.1 right now. The board actually is in a limbo where probably much of the design is shared with regular Tinkerboard/S, but other differences are unknown. Enabling a kernel module to let wifi work is a relatively s…

Description As described in the forum topic https://forum.armbian.com/topic/35677-tinkerboard-not-booting-after-updates/#comment-185896, upgrading on Asus Tinkerboard breaks the bootloader and prevents the entire system to reboot. As side effect - but not yet reported by anyone - should also break other rk3288 and rk322x board bootloaders. This PR restores the previous post-install platform_install.sh script for each board/family as it was before the family merge. Jira reference number AR-2109 How Has This Been Tested? [x] Compiled tinkerboard target, extracted the u-boot .deb package and manually inspected platform_install.sh script, not tested on a live system [x] C…

Description Fix mistake due to patch copied as-is from rk322x family without properly checking it first :pensive: How Has This Been Tested? [x] Kernel deb packages compiled Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description Hotfix for missing bluetooth on Asus Tinkerboard (rk3288) in RC build for Quoll 23.02 release. Also reinstate AUFS kernel module that was lost in some previous iteration. For reference: https://forum.armbian.com/topic/26661-armbian-2302-quoll-testings/?do=findComment&comment=160020 How Has This Been Tested? Please describe the tests that you ran to verify your changes. Please also note any relevant details for your test configuration. [x] Compiled deb packages, installed and tested successfully on a running system Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented…

Description Minimal fix for rk322x-box devices to restore the correct sdcard -> usb -> emmc boot order How Has This Been Tested? [x] Tested boot from sdcard and usb stick when armbian is installed in emmc Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code View the full article

Description This pr replaces nanopi's rtl8153 led patch with openwrt's, and removes useless led triggers. Also fixed support for orangepi-r1plus, including broken ethernet and gpio for led. How Has This Been Tested? [x] current [x] edge Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [ ] Any dependent changes have been merged and published in downstream modules View the full article

Description As per subject, move rockchip current to 6.6 and edge to 6.7. Some rework has to be done, in particular a LE patch has been fully merged (linux-0020) and can be remove and rockchip DFI driver has been heavily modified, so requires some patient work to be reinstated. Jira reference number AR-2016 How Has This Been Tested? [x] Compile rockchip 6.6 kernel [ ] Compile rockchip 6.7 kernel Checklist: [ ] My code follows the style guidelines of this project [ ] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [ ] I have made corresponding changes to the documentation [ ] My changes generate no n…

Description This PR creates a patch directory named v2022.04 and places u-boot patches for rockchip 32 bit (Tinkerboard/S) and rk322x into to leverage the new per-version patch format. Jira reference number AR-1637 How Has This Been Tested? [x] Built a fresh image and tested on rk322x device sdcard [x] Built a fresh image and tested on Asus Tinkerboard sdcard Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes…

Description Expose rk322x-dmc driver load, upthreshold and downthreshold attributes via sysfs objects to monitor and manipulate dmc behaviour via userspace. Implemented only on edge 6.6 kernel, which will become current very soon. Tunables will be available in /sys/class/devfreq/11200000.dmc/device directory How Has This Been Tested? [x] Compiled kernel and tested on live system Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] A…

Description As mentioned by @andyshrk , this gpu node patch has deleted line status = "disabled"; of node usb_host2_xhci, this will fix that so usb_host2_xhci will not get enabled on all boards. How Has This Been Tested? Please describe the tests that you ran to verify your changes. Please also note any relevant details for your test configuration. [x] ./compile.sh kernel BOARD=rock-5b BRANCH=edge DEB_COMPRESS=xz KERNEL_GIT=shallow [x] Check node usb_host2_xhci in file cache/sources/linux-kernel-worktree/6.8__rockchip-rk3588__arm64/arch/arm64/boot/dts/rockchip/rk3588s.dtsi [x] rock5b boots fine Checklist: Please delete options that are not relevant. [x] My code fol…

Description As per subject Jira reference number AR-1279 How Has This Been Tested? [x] Successfully built rockchip deb packages Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description Forgot to shift kernel configs for rockchip family when switching current to 6.6 and edge 6.7; done right now. Curiously I noticed this after hitting "buggy" behaviour on rk322x board with several rcu_scheduler debug messages while board was under high load, vnstatd process was often at 100% cpu usage and reading /proc/net/dev file (which is a simple text file with some network interface stats) stalls vnstatd. Turning off eth0 interface solved the problem but of course it was not solution. I shifted the kernel config, recompiled kernel 6.6 and this apparently solved the problem (maybe thanks to CONFIG_NEED_SRCU_NMI_SAFE=y ?) If it is so, the consideration …

Description As per subject, plus kernel configs have been updated and enabled rtl8xxx untested wifi chips How Has This Been Tested? [x] Kernel patches apply with current kernel [x] Kernel patches apply with edge kernel Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description stmmac ethernet driver is in use on all rockchip SoCs; its statistics system has been revised on kernel 6.6, but something is not working correctly right now on 32 bit platforms and a lock contemption happens very often when an userland process tries to read the statistics of the ethernet device. One of the most common victim is vnstat daemon, which often gets stuck hogging a cpu core and skyrocketing the average load. This also makes the whole kernel angry, RCU complains, processes hang, and so on... The patch provides a horrible workaround (remove the mutual exclusion in the reader) to fix the problem temporarely. This may break ethernet statistics that ca…

Description Due to some findings reported in previous PR (https://github.com/armbian/build/pull/3830#issuecomment-1181587003), forcing cgroups to v1 is insecure and break things. Also rockchip.txt default env file turned out to be in use for rockchip64 family too. In this PR, the cgroups workaround is removed for rockchip family and a proper rockchip64.txt default env file is added and set into include config file. Jira reference number AR-1259 How Has This Been Tested? Untested Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x…

rockchip64 / edge 6.6: add rk3566-orangepi-3b-sata DTB rockchip64 / edge 6.6: add rk3566-orangepi-3b-sata DTB rockchip64 / edge 6.6: add overlay rockchip-rk3566-sata2.dts this would also apply to 3568 View the full article

rockchip64 / edge 6.6: bump to 6.6-rc5; rebase patches rockchip64 / edge 6.6: bump to 6.6-rc5; rebase patches View the full article

Description As per subject, bump rockchip64 edge kernel to 6.8 Some patches have been reworked (helios64, nanopi-r2s, rockpi-s) to accomodate minor changes. Also this PR carries the same Makefile patch that was missing in tools/cgroup directory as well as https://github.com/armbian/build/pull/6429 Jira reference number AR-2112 How Has This Been Tested? [x] Kernel .deb compilation [ ] Live test kernel on rk3399 board [ ] Live test kernel on rk3328 board Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] My changes generate no n…

rockchip64 edge 6.6: bump to 6.6-rc6; rebase patches rockchip64 edge 6.6: bump to 6.6-rc6; rebase patches View the full article

Description Addresses the RK3399-only overlays by adding -rk3399 to them. Sort the Makefile to have some order to the chaos. (https://github.com/armbian/build/issues/5778) Jira reference number AR-1875 How Has This Been Tested? Build test Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description This patch PCIE initialization error after warm reboot. Typical error log after warm reboot: [ 2.510082] rockchip-pcie f8000000.pcie: host bridge /pcie@f8000000 ranges: [ 2.510139] rockchip-pcie f8000000.pcie: MEM 0x00fa000000..0x00fbdfffff -> 0x00fa000000 [ 2.510160] rockchip-pcie f8000000.pcie: IO 0x00fbe00000..0x00fbefffff -> 0x00fbe00000 [ 2.510748] rockchip-pcie f8000000.pcie: no bus scan delay, default to 0 ms [ 2.510795] rockchip-pcie f8000000.pcie: no vpcie12v regulator found [ 3.011274] rockchip-pcie f8000000.pcie: PCIe link training gen1 timeout! [ 3.011373] rockchip-pcie: probe of f8000000.pcie failed with erro…

Description As per the discovery made by @brentr , the current kernel for rockchip64 become hugely bigger than usual (13M vs. 18M, gzipped). After some checks, it turns out that what happened before with the appearance of useless and unwanted cruft (PATA, ACPI, SCSI :roll_eyes: , whatever... see linked PR) happened again. This PR restores the a sane kernel config for current 6.6 and edge 6.7, rebasing over kernel 6.1. I double checked against latest sane 6.6 I could find, which is 6.6-rc4 and differences are minor. I attach a "patch" file with the differences from sane 6.6-rc4 and 6.6.13 in this PR (use less -R to see with colors) in case someone wants to check it out…

Description RK3328 dmc driver was added a few kernels ago, it is broken with 6.1 causing kernel panic and failure to boot. Remove driver, disable boards that still depend on it but can't be tested. Patches need rebased for nanopi r2 (all) rk3328-roc-pc nanopi neo3 orangepi R1 How Has This Been Tested? [X] Build passes [X] Renegade boots successfully Checklist: [X] My code follows the style guidelines of this project [X] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [ ] I have made corresponding changes to the documentation [X] My changes generate no new warnings [X] Any dependent changes have be…

rockchip64_common: default, but do not overwrite, BL31_BLOB and MINILOADER_BLOB rockchip64_common: default, but do not overwrite, BL31_BLOB and MINILOADER_BLOB rockchip64_common: move the TB2's special DDR_BLOB out from family file into board file View the full article

rockchip64_common: use ttyFIQ0 as default SERIALCON for both legacy and vendor BRANCHes rockchip64_common: use ttyFIQ0 as default SERIALCON for both legacy and vendor BRANCHes View the full article

Description This PR adds missing Operating Performance Point (OPP) nodes for lower frequencies to the RK3588 device tree. by enabling the CPU clusters to scale down to lower frequencies when under light loads. The changes add OPP nodes for 408MHz, 600MHz, 816MHz, and 1008MHz (for cluster1 and cluster2 only, as cluster0 already had 1008MHz). How Has This Been Tested? [x] Tested with RK3588-based board (NanoPC T6-LTS) and verified CPU frequency scaling works correctly Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] My changes generate no new warnings View the full article

Description Fix Type-C DP altmode on some rk3399 boards. Currently only on rockchip64-6.15. The old implementation is tricky, and only works in normal orientation. If the plug is flipped, there will be only USB2.0 connectivity. With the proposed changes, the type-c ports on related rk3399 boards should work in both orientations. A few patches from Megous are picked(some of them already present in sunxi64 kernel patches). The tricky part is how to notify rk3399 cdn-dp driver about DP HPD(hot plug detection), and a workaround is written. A caveat is that the c port has to be OTG mode, otherwise it will behave very similar to the old implementation if DP is enabled. I'm …

Changes to the add-board patch did not get added to the makefile patch. Fixed. Description Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context. List any dependencies that are required for this change. Jira reference number AR-1258 How Has This Been Tested? Build and boot tested Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [ ] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged…

rockchip64-edge (6.8.y): enable CONFIG_RTW88_PCI and all the RTW88 PCI cards rockchip64-edge (6.8.y): enable CONFIG_RTW88_PCI and all the RTW88 PCI cards needed for the Tinkerboard-2 (8822CE) didn't touch SDIO or USB ones View the full article

Description We can also use rkvdec2 on rk356x. Remove all non-coherent-dst-bufs patches because they are suggested by a chromium developer when using libyuvimageprocessor to do NV12->AR24 convert. Now we use wayland gbm to render NV12 directly, so these patches are not necessary. Add rkvdec2 nodes to devicetree of rk356x. Enable rkvdec2 in kernel config and change CMA size from 128M to 384M for chromium's 4K h264 decoding. Change hantro g1 vpu compatible string from rockchip,rk3568-vpu to rockchip,rk3328-vpu because they have exactly the same vpu variant data. Disable h264 decoder of rk3328-vpu because there are aleady rkvdec/rkvdec2 h264 decoders on rk3328 and rk356…

Description Was having boot issues with RK3568 on edge 6.7.... Per @amazingfate 's advise.. updating montjoie's Rk35xx crypto patch from our Rk3588 patchset resolved the issue. Tested on: RK3568 - nanopi-r5s RK3399 - nanopc-t4 View the full article

Description Cut and sewing work around mailing list/libreelec patches to include a minimum set of patches that attempt to solve the various issues with HDMI on rockchip64 family. This affects both current kernel 6.6 and edge kernel 6.8. References on the forums: https://forum.armbian.com/topic/26818-opi-4-lts-no-hdmi-output/ https://forum.armbian.com/topic/37166-hdmi-issues-on-6x-kernel/ One kind guy tested the kernel package on rockpro64 and reported success: https://forum.armbian.com/topic/26818-opi-4-lts-no-hdmi-output/?do=findComment&comment=189646 Also remove a patch named "rk356x-vop2-support", that was not about vop2 support but rather tweaks the HDMI phy …

Description ssv6051p is found as cheap wifi/802.11n only on some older rk3328 boards; some community members asked for it and here it is. Also make some tidying up for device tree overlays for rk3318-box and updates kernel configurations for current and edge rockchip64 kernels. Driver is complied as module and should not be harmful for any other board. How Has This Been Tested? Please describe the tests that you ran to verify your changes. Please also note any relevant details for your test configuration. [x] Compiled deb packages for edge kernel and tested on live rk3318-box installation [x] Compiled deb packages for current kernel and tested on live rk3318-box instal…

Description Due to changes in https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/drivers/usb/typec/tcpm/tcpm.c?h=v6.6.31&id=789326cafbd1f67f424436b6bc8bdb887a364637 we need to adjust patch How Has This Been Tested? [x] Patch applies well [ ] Build test Checklist: [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description Restore CONFIG_OF_CONFIGFS=y accidentally(?) lost in https://github.com/armbian/build/commit/fae4f664397656e5a432a24b8adc5c237a53227e How Has This Been Tested? [x] Compile current&edge kernels, check overlays works Checklist: [ ] My code follows the style guidelines of this project [ ] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [ ] I have made corresponding changes to the documentation [ ] My changes generate no new warnings [ ] Any dependent changes have been merged and published in downstream modules View the full article

Description bump atf to LTS v2.8 for Rockchip64 How Has This Been Tested? Rockpro64 boots Checklist: [ ] My code follows the style guidelines of this project [ ] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [ ] I have made corresponding changes to the documentation [ ] My changes generate no new warnings [ ] Any dependent changes have been merged and published in downstream modules View the full article

Description Bump rockchip64 edge kernel to 6.9. The same fix that has been proposed in https://github.com/armbian/build/pull/6690 has been ported here too and it works as well. Note: at the current status, helios64 board patches have been disabled because the base device tree does not apply on kernel 6.9. Jira reference number AR-2348 How Has This Been Tested? [x] Compiled successfully [x] Upgrade the kernel (and device trees obviously) on a rk3318/rk3328 live system, dtbo applies correctly and system boots fine [ ] Upgrade the kernel on a live rk3399 system Checklist: Please delete options that are not relevant. [x] My code follows the style guidelines of this pr…

Description As per subject: bump rockchip64 current kernel to 6.6 and edge kernel to 6.7 Jira reference number AR-2017 How Has This Been Tested? [ ] Compile kernel deb packages for 6.6 [ ] Compile kernel deb packages for 6.7 [ ] Build image and test with kernel 6.7 for rk3328 board [ ] Build image and test with kernel 6.7 for rk3399 board Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and…

Description Bump rockchip64 family edge kernel to 6.4. All enabled patches from 6.3 have been ported from to 6.4, some of them have been reworked/adapted. Jira reference number AR-1810 How Has This Been Tested? [x] Patches apply cleanly [ ] kernel compiles Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description These are dead kernels with no ties to either vendor or mainline support. Remove patchsets to reduce clutter and confusion for any/all future contribution. (Example: Fixes should be posted to all LTS and current stable mainline kernel, not to these dead ones) How Has This Been Tested? Kernels have to be manually selected be editing config files, they aren't being used in the build system itself, any niche use for them can be accomplished by pulling them from history by the user. Checklist: [ ] My code follows the style guidelines of this project [ ] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-underst…

Description pbp dp port alt mode patch breaking on 6.6.31 https://paste.armbian.com/udedocucek View the full article

Description A change in kernel 5.19.2 causes kernel oops in inno usb2 phy initialization (at least) when there is an usb3 type-c connector device (like fairchild usb3 controller) trying to initialize the usb2 phy. My Orange PI 4 LTS was entering kernel panic when the monitor was brought on by effect of fusb302 kernel module initializing the whole type-c subsystem (display port, usb2 and usb3 phys). This patch removes the init code introduced in 5.19.2 How Has This Been Tested? [x] Produced kernel debs and installed on living Orange PI 4 LTS system Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x…

Description Enabling CPU idle can lower CPU power at low load. How Has This Been Tested? [x] Bootup test [x] Stress test Checklist: [ ] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [ ] I have made corresponding changes to the documentation [x] My changes generate no new warnings [ ] Any dependent changes have been merged and published in downstream modules View the full article

Description Orange Pi 4 LTS has DDR4 memory banks that always run at full speed. This PR enables the DMC driver in kernel 6.1 to lower the frequency to reduce power usage during idle periods. Benchmarks with a digital multimeter show that the power usage decreases by 200mW when the board is idling. Jira reference number AR-1584 How Has This Been Tested? This is a pre-armbian-next era patch and has been tested to be fully functional. I rebased it successfully on main branch, but requires to be tested not to break the compilation. [ ] compile to produce kernel deb packages Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-…

Description Due to "recent" fixes to the rk3328 dram memory controller driver (see https://github.com/armbian/build/pull/4774) and tests/benchmarks being made on Rock Pi E board, enable the dmc driver on such CSC board for a nice and free speed bump Jira reference number AR-1585 How Has This Been Tested? The patch is from the pre-armbian-next era and has been tested there to be fully functional, I will check that it does not break compilation on main branch as soon as possible. [ ] compile deb kernel packages on main branch Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my co…

Description Small followup fix https://github.com/armbian/build/pull/6006 How Has This Been Tested? Please describe the tests that you ran to verify your changes. Please also note any relevant details for your test configuration. [x] Compile and boot rk3318-box from second partition Checklist: [ ] My code follows the style guidelines of this project [ ] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [ ] I have made corresponding changes to the documentation [ ] My changes generate no new warnings [ ] Any dependent changes have been merged and published in downstream modules View the full article

Description For some reason, the same config overhaul happened again in upgrading from 6.0 to 6.1, just like https://github.com/armbian/build/pull/4442 This fixes again the rockchip64 config with inherited options, plus enables MGLRU by default for increased performance. How Has This Been Tested? [x] Built deb packages for edge kernel and tested on a live rk3318/rk3328 system Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any …

Description Kernel config for rockchip64 for kernel 6.0 got a sudden overhaul I don't know it was wanted or not. That caused the addition of a huge number of unneeded modules (Parallel ATA, Intel enterprise things, PCI drivers, NAND support and drivers, and other crap) that causes a grow in compressed modules size of 20 megabytes and 2 megabyte more on the kernel image plus several other insane settings, like CMA set to 384 megabytes. This PR restores the same "shape" of the previous kernel config and fixes ssv6051 and uwe5622 patches to compile without errors with v6.0. How Has This Been Tested? [x] kernel deb packages have been compiled, verified that final size is i…

Description Orange PI 4 LTS was flawed with sdcard detection issues, sometimes happening on current 5.15 kernel, practically always happening. All the changes have been taken from xunlong orangepi repositories device tree for kernel 5.17 Jira reference number AR-1369 How Has This Been Tested? [x] Compiled a device tree with the changes and applied to existing system with kernel 5.19.3. After this modification, the board rebooted from sdcard 3 times out of 3. Previously the board would boot 1 time out of dozen of attempts. Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code…

Description Hotfix for rk3328 dram memory controller driver has some issues and causes crashes on board on which is activated (see https://github.com/armbian/build/pull/4767) on edge 6.1 kernel. This pull request fixes the issue, plus removes some cruft and warnings from the existing code. How Has This Been Tested? [x] Compile kernel deb packages successfully [x] Install on a rk3318 system [ ] Install on rk3328 system Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My…

Description Driver does not compile on kernel 6.7 due to minor changes in kernel API. The PR just fixes the changes. How Has This Been Tested? [x] Compiled rockchip64 edge 6.7 kernel Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description Fix incorrect data read sometimes in last 4kb of compressed files on btrfs. In my case only last 2560 bytes of dtb file was broken, so the only issue was warnings in dmesg on boot. Well, last 2560 of vmlinuz was also broken, but fortunately there was only '\0' bytes, so system was functional and this problem was unnoticed. Patch taken from upstream uboot: https://github.com/u-boot/u-boot/commit/b1d3013d024086c042dbae4ddd99db56bb55b5e7 How Has This Been Tested? [x] Boot from zstd compressed btrfs partition [x] Check uboot loaded kernel sha1sum in uboot console Checklist: [ ] My code follows the style guidelines of this project [ ] I have performed a self-…

Description Fix uboot booting from zstd compressed BTRFS. uboot logs before fix: Found U-Boot script /boot/boot.scr ZSTD_decompressStream error 7 3252 bytes read in 41 ms (77.1 KiB/s) ... SCRIPT FAILED: continuing... Error 7 means "frame parameter window too large". It looks like uboot's max window size specification on decompressor init is totally wrong (compressed data size is used as max window size). Zstd default max window size is 128MB, but I'm unsure is it good idea to malloc 128MB in uboot on each BTRFS extent decompression. So I have capped minimum window size to 128KB - it should be sufficient for BTRFS as max compressed extent size is 128KB. How Has This Be…

Description Add patches to support HEVC hardware video decoder provided by rkvdec in rockchip64 socs like rk3328 and rk3399. This leverages the hardware features reducing cpu usage by a wide margin. Also provides some patches and fixes for v4l2 whose patch files seems big, but most is just description text. Can be tested along with the repository published by me on this forum post. Jira reference number AR-1997 How Has This Been Tested? [x] Compile current 6.1 rockchip64 kernel [x] Compile edge 6.6 rockchip64-kernel [x] Test current kernel on existing rk3399 installation on OrangePi4 LTS board [x] Test edge kernel on existing rk3399 installation on OrangePi4 LTS boar…

Description There is an orbiting patch that adds True Random Number Generator (TRNG) completing the rockchip mainline crypto driver. This PR introduces the patch (which is in RFC condition, thus passible of changes) with these benefits: provides better entropy (see below) removes the memory region claim error in dmesg is almost in mainline Some caveats: the existing driver covers rk3288 (32 bit, not applicable in this context), rk3328 and rk3399, but only rk3399 seems to have working hardware. For the other two chips, the TRNG is just disabled. How Has This Been Tested? [x] Current 6.6 kernel packages built and installed on live system (rk3399) [x] Edge 6.7 kernel pa…

Description Current rockchip64 bootscript always load kernel/etc from the first device partition, but not from partition bootscript itself was loaded. This PR makes bootscript load kernel/etc from the same partition. This allows to easily have multiple boot partitions (primary/backup) and switch them with MBR boot flags or GPT legacy_boot flags. How Has This Been Tested? [x] Compile and boot rk3318-box Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [ ] I have commented my code, particularly in hard-to-understand areas [ ] I have made corresponding changes to the documentation [ ] My changes genera…

Description In order to fullfil the Jira task AR-1988 (Resolve Rockchip patch maintenance nightmare), remove rockpis family in favor of rockchip64 and merge the configuration bits from the family .conf file into the board rockpi-s.conf file. Also removes the legacy patches from the archive Jira reference number AR-1988 How Has This Been Tested? [x] A current kernel image has been produced [x] idbloader has been verified to be matching against the pre-cleanup image [x] /boot/armbianEnv.txt contains the overlay_prefix=rk3308 [x] /boot/boot.cmd matches config/bootscripts/boot-rockpis.cmd [ ] /boot/armbianEnv.txt contains /config/bootenv/rockpi-s.txt Please @brentr check …

Description Some patches were triggering warning in the armbian build system for kernel 6.1 in rockchip64 family. This PR rework those patches to avoid warnings during kernel patching, does not introduce new features, just cleans some things up. Also the Pinebook Pro patch that was addressing the external control (extcon) for usb3 type-c connector is now split into a patch that adds the device tree bits for pinebook and another patch that implements the feature, since the feature is used in other rk3399 boards. How Has This Been Tested? [x] Debian bullseye minimal image has been built and tested on Orange Pi 4 LTS Checklist: [x] My code follows the style guidelines o…

Description rework to apply cleanly on 6.10.5 and 6.6.46 - root cause GitHub issue reference: #7087 Jira reference number AR-2467 special thanks: @EvilOlaf How Has This Been Tested? [x] Patch applies cleanly to rockchip64 edge 6.10.5 - compiling ok [x] Patch applies cleanly to rockchip64 current 6.6.46 - compiling ok Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] My changes generate no new warnings View the full article

Description Current kernel on rockchip64 does not compile because some realtek drivers want ipx.h kernel header. Add a patch (taken from sunxi-5.15) to put back ipx.h in place. How Has This Been Tested? [x] Built kernel deb packages Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] I have made corresponding changes to the documentation [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

Description They don't apply as they have been merged upstream and build fails. View the full article

Description As long as the rockchip-inno-usb3 phy driver has been removed from current kernel, remove it from edge kernel as well. This affects only rk3328 devices, but so far none of them has enabled the driver except roc-pc and station-m1; station-m1 dts file has been removed since it does not appear in CURRENT kernel patch directory, roc-pc device tree has been adapted. So far I tested the phy driver a bit going into the OTG journey. It turned out that does not work at all with the dwc3 USB3 core driver; The phy driver presence is somewhat necessary to attach the proper OTG vbus regulator for OTG function, thus removing the driver will probably hinder future OTG func…

Description Quick fix to hdmi patch due to mainlined hunks. How Has This Been Tested? [x] New fixed patch apply on 6.8.9 edge kernel [x] Compiled rockchip drm code on the fly to check it works Did not check the real functionality, but I hope I did not break anything, the patch is minimal... Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [x] My changes generate no new warnings View the full article

Description A patch for Pinebook Pro was not applying cleaning on current 5.15 rockchip64 kernel flavour, turns out that Orange Pi 4 LTS has no more HDMI video output and dmesg is lamenting about a missing extcon connector over fusb driver. fusb driver has the control over video output since it has the capacity to steer video output from HDMI to DP via USB type-c connector. A broken fusb driver causes broken video output; this problem can be shared among all boards that depend upon it. This pull request fixes the application of the Pinebook Pro patch, and perhaps fixes the Orange Pi 4 LTS broken video output. edit: I confirm, fixing the patch fixes also video output on…

Description Orange Pi 4 and Orange Pi 4 LTS were the last boards using staging USB3 Type-C controller. This PR moves both of them to use mainline driver, so staging driver can be removed from patchset (as suggested by @Tonymac32 here: https://github.com/armbian/build/pull/3953#issuecomment-1172962963) Most of the changes have been adapted upon the same nodes from tinkerboard 2 device tree and GPIOs have been adjusted. Since both opi4 and opi4 LTS share the same hardware, I felt it could be useful to change on both boards. Probably if it works for my opi4 LTS also works for regular opi4; I will do my tests, if @piter75 wants to do private tests, in compatibility with hi…

Patch is obsolete. The rk3399 has a crypto IP handled by the rk3288 crypto driver so since 6.2 there is a node for it. Description Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context. List any dependencies that are required for this change. Jira reference number [AR-9999] How Has This Been Tested? Please describe the tests that you ran to verify your changes. Please also note any relevant details for your test configuration. on RockPi 4b crypto is able to complete init, which otherwise fails Checklist: [ ] My code follows the style guidelines of this project [ ] I have performed a self-review of my ow…

Description Move stack pointer further from base address to allow bigger uboot images. Currently uboot stack pointer located 1MiB after text base, and uboot size with defconfig is barely less than 1MiB (1001472). So any small modification in code/config/compiler-version could overflow text area (my uboot blown up after I enabled verbose logging to debug btrfs boot problem). How Has This Been Tested? [x] Build and boot rk3318-box Checklist: [x] I have performed a self-review of my own code View the full article

Description The legacy U-Boot can't be built anymore and 4.4 is EOL. As discussed in https://github.com/armbian/build/issues/6821#issuecomment-2205234200 I didn't see any kernel or uboot patch folder. If I forgot something related to legacy, please let me know :) GitHub issue reference: https://github.com/armbian/build/issues/6821 Jira reference number AR-2392 Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas View the full article

Description Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context. List any dependencies that are required for this change. Jira reference number AR-1438 Tony McKahan an issue complaining that rockpis, among others, was overwriting the mainline kernels devtree. This patch replaces the existing sequence of board-rockpis-* patches doing that with a single one that starts from the rockpi-s devtree newly included in the mainline kernel. In reviewing the devtree, I noticed that the Bluetooth items for UART4 had somehow been lost. So, I restored them. How Has This Been Tested? Build new image. Verified WiFi, eth…

Description Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context. List any dependencies that are required for this change. Jira reference number AR-1268 AR-1269 AR-1270 The fixMACaddress udev rules are generic enough to be used on all boards where random MAC addresses are a problem. However, for now, the feature is only enabled as part of the RockPi S BSP. Other BSPs could also copy the directory package/bsp/fixMACaddress to the $destination if desired. How Has This Been Tested? Generated new images, verified that WiFi address is now fixed across reboots. Verified that applying SDIO clock speed DT overlay …

Description This backports a commit from the Linux kernel into version 5.15 in order to enable the automatic control of the fan, based on the CPU temperature. Otherwise, the fan would run at maximum speed all the time (per default). How Has This Been Tested? I've built the kernel and installed the resulting .deb packages on my RockPro64 (running Bullseye). I monitored the CPU temperature and verified that the fan speeds up as soon as it hits the trip points defined in the device tree. Checklist: [x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] My changes generate no new warnings [x] Any dependent changes hav…

rockpro64: bump ATF to lts-2.8.8 rockpro64: bump ATF to lts-2.8.8 View the full article

rockpro64: bump u-boot to 2024.07; make fancy configs; drop bin size limit as fix landed upstream rockpro64: bump u-boot to 2024.07; make fancy configs; drop bin size limit as fix landed upstream View the full article

rockpro64: bump u-boot to v2024.04-rc4; use binman-produced bins rockpro64: bump u-boot to v2024.04-rc4; use binman-produced bins mmc-sdhci-allow-disabling-sdma-in-spl.patch has landed in https://github.com/u-boot/u-boot/commit/3cd664dc92ca832506f1a4e7769cb5ee6a88137d use binman-produced binaries use flashcp -p to write fast to spi also: bump ATF (TF-A) to lts-v2.8.16 also: enable TCP networking cmds in u-boot also: configure /etc/fw_env.config userspace to SPI env coordinates configure /etc/fw_env.config userspace to SPI env coordinates include libubootenv-tool userspace for fw_printenv and fw_setenv View the full article

Description armbian-install should mask installing to root partition. How Has This Been Tested? [x] Manual run To do: fix labels change executing bit after transferring script to the target. View the full article

Currently it doesn't provide any data to what failed, this attempts to address it View the full article

I believe this is what causes the red systemd error on console-setup during boot View the full article