staging/stintel/package/base-files, branch master

base-files: system.sh: add mtd_get_part_erasesize()

2026-11-04T21:34:59Z

Similar to mtd_get_part_size() but for the erase sector size. Signed-off-by: Michael Pratt Link: https://github.com/openwrt/openwrt/pull/24866 Signed-off-by: Robert Marko

base-files: add mtd_get_mac_nvmem() for VLAN MAC addresses

2026-11-04T21:34:59Z

Devicetree only allows for the base physical address of a device to be set using NVMEM cells with a phandle to an NVMEM device. This leaves out the ability to set a VLAN MAC, for example, a WAN port of a switch to be a VLAN for the WAN firewall side with a separate MAC address for isolated identification. When the u-boot-env layout driver probes successfully, all of the variables are available as NVMEM cells in sysfs, not just the ones used with a phandle in devicetree. Add a new function in system.sh that reads the cells in sysfs based on the MTD device used in the NVMEM driver, similar to mtd_get_mac_binary() and to how u-boot tools are patched to detect NVMEM devices. Ref: 8142c4554ffa ("fw_env: add fallback to Linux's NVMEM based access") # uboot.git Signed-off-by: Michael Pratt Link: https://github.com/openwrt/openwrt/pull/24866 Signed-off-by: Robert Marko

base-files: add /tmp/resolv.conf.d/resolv.conf

2026-11-03T11:57:36Z

/tmp/resolv.conf links to the netifd-maintained resolv.conf.auto, and dnsmasq replaces the link by a file of its own, so the resolver configuration leaves /tmp/resolv.conf.d whenever dnsmasq runs. A jail sharing the host network can follow a file inside that directory across rename-replacement, but not one outside it. Create /tmp/resolv.conf.d/resolv.conf as the resolver configuration of the system and point /tmp/resolv.conf at it. It links to resolv.conf.auto relative to the directory, so it also resolves where a jail mounts the directory at a different path. Signed-off-by: Daniel Golle

base-files: start only enabled services on package upgrade

2026-10-28T14:46:20Z

Since commit afaa34ccd734 ("base-files: don't modify enabled state of service on upgrade"), default_postinst() starts every init script of the package on an upgrade, and leaves out only the enable. With opkg the prerm stops the service first, so the pair acts as a restart. apk runs no prerm on an upgrade, and its post-upgrade script calls default_postinst() with PKG_UPGRADE=1. A reinstall or an upgrade with apk thus starts a service that the system has disabled and stopped. On an upgrade, start an init script only where it reports itself enabled. A fresh install still enables and starts each service, and an offline root still only enables it. With opkg, a service that the system has disabled but that was started by hand now stays stopped after an upgrade, since the prerm stops it and the postinst no longer starts it. Signed-off-by: Felix Fietkau

base-files: legacy-sdcard: don't erase extra partitions on sysupgrade

2026-10-15T08:38:14Z

Writing the full image erases partitions the image lacks. Refuse that at upgrade time unless sysupgrade -p is given, so --force and failsafe cannot bypass it, and read MBR and GPT tables the same way everywhere. sysupgrade now exports SAVE_PARTITIONS to the image check and the -p help text mentions the erase. Signed-off-by: John Audia

base-files: legacy-sdcard: drop the bootloader write

2026-10-15T08:38:14Z

It writes to "$diskdev" without /dev/, so it creates a file in the working directory and has never reached the disk. Signed-off-by: John Audia

base-files: support per-color intensities for multicolor LEDs

2026-10-11T19:28:57Z

The led init script already knows how to drive a multicolor LED: it reads multi_index and writes one intensity per color to multi_intensity, taken from color_ options of the led section. Nothing fills those options in, though - generate_led only passes name, sysfs, trigger and default on to uci, so the intensities stay at zero and the LED remains dark whatever trigger is set. No board has been able to use it so far. Pass any color_* key of a board.json led object through to the led section, and add ucidef_set_led_colors() to set them from board.d: ucidef_set_led_default "led1" "LED1" "nbg7815:led1" "1" ucidef_set_led_colors "led1" red 255 green 255 blue 255 The colors are not restricted to RGB - the keys are matched against whatever the LED reports in multi_index. Signed-off-by: André Valentin

base-files: keep /var/lock volatile with persistent /var

2026-10-11T14:39:26Z

Enabling CONFIG_TARGET_ROOTFS_PERSIST_VAR leaves /var/lock on persistent storage, allowing stale lock files to survive reboots. Link /var/lock to ../tmp/lock so that lock files remain volatile. Use a relative symlink to keep build-time accesses within the target root filesystem. Create the target directory for rootfs preparation. procd already recreates /tmp/lock at boot. Fixes: 57807f50ded6 ("base-files: add option to make /var persistent") Signed-off-by: Ivan Romanov Link: https://github.com/openwrt/openwrt/pull/25718 Signed-off-by: Robert Marko

base-files: add -f to board_detect and config_generate

2026-10-08T07:50:37Z

board_detect runs board.d only when the target file is missing, and /etc/board.json is now a symlink to a copy of board-base.json. With -f, board_detect runs board.d again, also when /tmp/board.json exists, and replaces the target file. config_generate passes -f to board_detect. config_generate still creates only the UCI configuration that is missing. The regenerated copy in tmpfs has no wlan runtime data until the next 'wifi config'. Signed-off-by: Felix Fietkau

base-files: keep the board.d data in board-base.json, board.json in tmpfs

2026-10-08T07:50:37Z

wifi-detect adds runtime wlan data to /etc/board.json. It rewrites the file on flash at least twice on every boot: the run before kmodloader sees no wiphy and removes the data, and the hotplug run adds it again. board_detect now writes the board.d data to /etc/board-base.json and copies it to /tmp/sysinfo/board.json on every boot. base-files ships /etc/board.json as a symlink to the copy in tmpfs, so readers do not change. If /etc/board.json is not a symlink, for example a file kept across a reflash or a file added to the image, board_detect replaces it with the symlink. The sysupgrade overlay backup skips board-base.json, as it skips board.json. If /tmp/board.json exists when board_detect creates board-base.json, board_detect uses it instead of running board.d again. That file comes from preinit_config_board() or the x86 diag.sh in the same boot. The preinit shell holds their exported jshn state, and board.d run from it writes an empty object. Changes in behaviour: - Changes to /etc/board.json do not persist. Persistent changes go to /etc/board-base.json. - Removing /etc/board.json no longer runs board.d again. - board_detect exits 0 when the target file exists. Before, it exited 1. As a result, preinit_match_diag_led() in the x86 diag.sh now uses the /tmp/board.json from preinit_config_board(). Signed-off-by: Felix Fietkau