aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* base-files/functions.sh: use grep -q instead of []Rosen Penev2020-01-181-4/+4
| | | | | | | | | | | | It's cleaner and faster as it does not need to do extra work. Also removed $() to avoid executing the output. The shell can handle it. https://github.com/koalaman/shellcheck/wiki/SC2143 Signed-off-by: Rosen Penev <rosenp@gmail.com> [correct || to && for one conversion] Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
* ramips: fix wps leds/btn for TP-Link TL-WA801ND v5Jan Alexander2020-01-171-4/+16
| | | | | | | | | | - fix color and active mode for existing wps led - add green wps led - add wps button Signed-off-by: Jan Alexander <jan@nalx.net> [wrap line] Signed-off-by: David Bauer <mail@david-bauer.net>
* base-files/system.sh: remove $ in $(())Rosen Penev2020-01-171-1/+1
| | | | | | | | Not needed. https://github.com/koalaman/shellcheck/wiki/Sc2004 Signed-off-by: Rosen Penev <rosenp@gmail.com>
* base-files/functions.sh: use && instead of -aRosen Penev2020-01-171-2/+2
| | | | | | | | -a is not well defined. https://github.com/koalaman/shellcheck/wiki/SC2166 Signed-off-by: Rosen Penev <rosenp@gmail.com>
* base-files/functions.sh: remove useless catRosen Penev2020-01-171-2/+2
| | | | | | | | The cut command can take a file as an input. https://github.com/koalaman/shellcheck/wiki/SC2002 Signed-off-by: Rosen Penev <rosenp@gmail.com>
* base-files/functions.sh: don't use $var in $(())Rosen Penev2020-01-171-3/+3
| | | | | | | | It's not needed. It can also lead to subtle bugs. https://github.com/koalaman/shellcheck/wiki/Sc2004 Signed-off-by: Rosen Penev <rosenp@gmail.com>
* libcxx: fix build for x86/64Stijn Tintel2020-01-171-0/+1
| | | | | | | | | | When building libcxx for x86/64, the library is installed in /usr/lib64. As the install section tries to copy the library from /usr/lib, this breaks build on x86/64. Override the lib dir suffix to fix this. Fixes: 856ea2bad3b3 ("libcxx: Add package") Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be> Acked-by: Rosen Penev <rosenp@gmail.com>
* odhcpd: update to version 2020-01-14Hans Dedecker2020-01-161-3/+3
| | | | | | | | | 6db312a dhcpv6-ia: use dhcp leasetime to set preferred/valid statefull lifetimes 2520c48 dhcpv6-ia: introduce DHCPv6 pd and ia assignments flags b413d8a dhcpv6-ia: cleanup prefix delegation routes b0902af dhcpv6-ia: remove passing interface as parameter to apply_lease Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
* ccache: update to 3.7.7DENG Qingfang2020-01-162-3/+3
| | | | | | | | | Update ccache to 3.7.7 Release notes: https://ccache.dev/releasenotes.html#_ccache_3_7_7 Signed-off-by: DENG Qingfang <dengqf6@mail2.sysu.edu.cn>
* ramips: fix HiWiFi HC5962 status LEDDENG Qingfang2020-01-161-3/+4
| | | | | | | | | Match LED behavior to stock firmware: Red: booting White: running Signed-off-by: DENG Qingfang <dengqf6@mail2.sysu.edu.cn>
* ramips: fix HiWiFi HC5962 switch configurationDENG Qingfang2020-01-161-1/+4
| | | | | | HC5962 has only 3 LAN ports, switch port 0 is unused Signed-off-by: DENG Qingfang <dengqf6@mail2.sysu.edu.cn>
* hostapd: add support for system cert bundle validationDavid Lam2020-01-162-6/+15
| | | | | | | | | | | | | | Currently, it is very cumbersome for a user to connect to a WPA-Enterprise based network securely because the RADIUS server's CA certificate must first be extracted from the EAPOL handshake using tcpdump or other methods before it can be pinned using the ca_cert(2) fields. To make this process easier and more secure (combined with changes in openwrt/openwrt#2654), this commit adds support for validating against the built-in CA bundle when the ca-bundle package is installed. Related LuCI changes in openwrt/luci#3513. Signed-off-by: David Lam <david@thedavid.net> [bump PKG_RELEASE] Signed-off-by: Jo-Philipp Wich <jo@mein.io>
* hostapd: cleanup IBSS-RSNDaniel Golle2020-01-162-6/+2
| | | | | | set noscan also for IBSS and remove redundant/obsolete variable. Signed-off-by: Daniel Golle <daniel@makrotopia.org>
* ath79: ar934x: use reset for usb-phy-analogJohann Neuhauser2020-01-151-2/+2
| | | | | | | | | | | | This was already available on ar71xx, but is missing on ath79. This solves the slow usb speed on TP-Link WDR3600/WDR4300 and similar, as reported in Flyspray [0], OpenWRT Forum [1] and GitHub PR [2]. [0] https://bugs.openwrt.org/index.php?do=details&task_id=2567 [1] https://forum.openwrt.org/t/usb-wdr4300-low-speed-on-external-storage/46794 [2] https://github.com/openwrt/openwrt/pull/964 Signed-off-by: Johann Neuhauser <johann@it-neuhauser.de>
* ath79: phy-ar7200-usb: adapt old behavior of arch/mips/ath79/dev-usb.cJohann Neuhauser2020-01-151-6/+18
| | | | | | | | | | | | Do not put usb-phy into reset if clearing the usb-phy reset or setting the suspend_override has failed. Reorder (de)asserts like in arch/mips/ath79/dev-usb.c. Add an optional reset_control "usb-phy-analog", which is needed for ar934x SoCs like in the old mach-driver arch/mips/ath79/dev-usb.c. Signed-off-by: Johann Neuhauser <johann@it-neuhauser.de>
* ar71xx/mikrotik: use ath10k-ct-smallbuffers for 64 MiB devicesThomas Nixon2020-01-151-1/+1
| | | | | | | | | | This image is only needed on one device (wAP AC); since this target is going to be removed anyway it doesn't make sense to add an extra "low RAM" image. Fixes OOM issues on RouterBoard wAP AC. Signed-off-by: Thomas Nixon <tom@tomn.co.uk>
* uboot-oxnas: fix memory leak in tool mkox820crcAndrea Dalla Costa2020-01-151-0/+3
| | | | | | | | In function `main` add calls to `free` for the variable `executable`. This is needed because the variable `executable` is allocated but never freed. This cause a memory leak. Signed-off-by: Andrea Dalla Costa <andrea@dallacosta.me>
* ath79: use caldata partition label consistentlyAdrian Schmutzler2020-01-152-8/+8
| | | | | | | Change the caldata partition DTS node label to be consistent with the label property for some Netgear WNDR devices. Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
* malta: enable HighMem on MIPS32Aleksander Jan Bajkowski2020-01-152-0/+2
| | | | | | It allows to use more than 256MB memory on MIPS32. Signed-off-by: Aleksander Jan Bajkowski <A.Bajkowski@stud.elka.pw.edu.pl>
* dropbear: fix compile errorJohn Crispin2020-01-151-1/+2
| | | | | Fixes: 0da193ee6943 ("dropbear: move failsafe code out of base-files") Signed-off-by: John Crispin <john@phrozen.org>
* wireguard: skip peer config if public key of the peer is not definedFlorian Eckert2020-01-151-0/+5
| | | | | | | | | | | | | | | If a config section of a peer does not have a public key defined, the whole interface does not start. The following log is shown daemon.notice netifd: test (21071): Line unrecognized: `PublicKey=' daemon.notice netifd: test (21071): Configuration parsing erro The command 'wg show' does only show the interface name. With this change we skip the peer for this interface and emit a log message. So the other peers get configured. Signed-off-by: Florian Eckert <fe@dev.tdt.de>
* busybox: fix build issuesJohn Crispin2020-01-151-1/+3
| | | | | Fixes: f704f97e4c57 ("busybox: Include hdparm by default on nas type device") Signed-off-by: John Crispin <john@phrozen.org>
* ath79: add support for Netgear WNDR4500 v3Michal Cieslakiewicz2020-01-156-5/+63
| | | | | | | | | | | | | | | | | | | This patch introduces support for Netgear WNDR4500v3. Router is very similar to WNDR4300v2 and is based on the same PCB. Information gathered from various Internet sources (including https://patchwork.ozlabs.org/patch/809227/) shows following differences to WNDR4300v2: * two USB 2.0 ports with separate LEDs * USB LEDs soldered to secondary pads * WPS and RFKILL buttons soldered to secondary pads * described as N900 device with 3x3:3 MIMO for 2.4GHz radio * power supply requirement is DC 12V 2.5A * vendor HW ID suffix differs in one digit * bigger chassis Signed-off-by: Michal Cieslakiewicz <michal.cieslakiewicz@wp.pl>
* ath79: add support for Netgear WNDR4300 v2Michal Cieslakiewicz2020-01-157-5/+297
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch introduces support for Netgear WNDR4300v2. Specification ============= * Description: Netgear WNDR4300 v2 * Loader: U-boot * SOC: Qualcomm Atheros QCA9563 (775 MHz) * RAM: 128 MiB * Flash: 2 MiB SPI-NOR + 128 MiB SPI-NAND - NOR: U-boot binary: 256 KiB - NOR: U-boot environment: 64 KiB - NOR: ART Backup: 64 KiB - NOR: Config: 64 KiB - NOR: Traffic Meter: 64 KiB - NOR: POT: 64 KiB - NOR: Reserved: 1408 KiB - NOR: ART: 64 KiB - NAND: Firmware: 25600 KiB (see notes for OpenWrt) - NAND: Language: 2048 KiB - NAND: mtdoops Crash Dump: 128 KiB - NAND: Reserved: 103296 KiB * Ethernet: 5 x 10/100/1000 (4 x LAN, 1 x WAN) (AR8337) * Wireless: - 2.4 GHz b/g/n (internal) - 5 GHz a/n (AR9580) * USB: yes, 1 x USB 2.0 * Buttons: - Reset - WiFi (rfkill) - WPS * LEDs: - Power (amber/green) - WAN (amber/green) - WLAN 2G (green) - WLAN 5G (blue) - 4 x LAN (amber/green) - USB (green) - WPS (green) * UART: 4-pin connector JP1, 3.3V (Vcc, TX, RX, GND), 115200 8N1 * Power supply: DC 12V 1.5A * MAC addresses: LAN=WLAN2G on case label, WAN +1, WLAN5G +2 Important Notes =============== 0. NOR Flash (2 MiB) is not touched by OpenWrt installation. 1. NAND Flash (128 MiB) layout under OpenWrt is changed as follows: all space is split between 4 MiB kernel and 124 MiB UBI areas; vendor partitions (language and mtdoops) are removed; kernel space size can be further expanded if needed; maximum image size is set to 25600k for compatibility reasons and can also be increased. 2. CPU clock is 775 MHz, not 750 MHz. 3. 5 GHz wireless radio chip is Atheros AR9580-AR1A with bogus PCI device ID 0xabcd. For ath9k driver to load successfully, this is overriden in DTS with correct value for this chip, 0x0033. 4. RFKILL button is wired to AR9580 pin 9 which is normally disabled by chip definition in ath9k code (0x0000F4FF gpio mask). Therefore 'qca,gpio-mask=<0xf6ff>' hack must be used for button to work properly. 5. USB port is always on, no GPIO for 5V power control has been identified. Installation ============ * TFTP recovery * TFTP via U-boot prompt * sysupgrade * Web interface Test build configuration ======================== CONFIG_TARGET_ath79=y CONFIG_TARGET_ath79_nand=y CONFIG_TARGET_ath79_nand_DEVICE_netgear_wndr4300-v2=y CONFIG_ALL_KMODS=y CONFIG_DEVEL=y CONFIG_CCACHE=y CONFIG_COLLECT_KERNEL_DEBUG=y CONFIG_IMAGEOPT=y Signed-off-by: Michal Cieslakiewicz <michal.cieslakiewicz@wp.pl>
* ath79: WNDR4300: increase kernel partition to 4MMichal Cieslakiewicz2020-01-152-6/+6
| | | | | | | | | Increase kernel partition from 2 MiB to 4 MiB for Netgear WNDR routers with NAND flash. Change affects following devices: * Netgear WNDR3700 v4 * Netgear WNDR4300 Signed-off-by: Michal Cieslakiewicz <michal.cieslakiewicz@wp.pl>
* mac80211: ath9k: add GPIO mask dts propertyMichal Cieslakiewicz2020-01-151-0/+25
| | | | | | | | | | | | | | | This patch adds 'qca,gpio-mask=<u32>' device tree property to ath9k node. This optional setting is a hack and should only be used in very special (and rare) cases when a button or LED is wired to a GPIO pin normally masked out (due to being one-way etc). Netgear WNDR4300 v2 is one such example - it uses GPI9 for RFKILL. See ath9k/reg.h *_GPIO_MASK constants. Use with caution and expect to see stream of kernel warnings if wrong mask value is provided. Signed-off-by: Michal Cieslakiewicz <michal.cieslakiewicz@wp.pl>
* zram-swap: support swap priorityMaxim Storchak2020-01-152-2/+4
| | | | | | | | If zram-backed swap is added after an existing swap, it gets a lower priority. Assiming that usually all other swaps are slower, there should be a way to assign a higher priority to zram swap. Signed-off-by: Maxim Storchak <m.storchak@gmail.com>
* ath79: add support for D-Link DIR-505David Bauer2020-01-154-0/+164
| | | | | | | | | | | | | | | | | | | | | | | This commit adds support for the D-Link DIR-505, previously supported in ar71xx. Hardware -------- SoC: Atheros AR9330 FLASH: 8M SPI-NOR RAM: 64M WIFI: 1T1R 1SS Atheros AR9330 LED: Power green, Status red BTN: WPS, Reset Installation ------------ Currently, installation is only possible by sysupgrading from an earlier OpenWrt version, U-Boot TFTP or a modded U-Boot. I do not have the original bootloader from D-Link on my device anymore, so i cannot test the factory image. Signed-off-by: David Bauer <mail@david-bauer.net>
* perf: Add libunwind only if selectedRosen Penev2020-01-151-2/+2
| | | | | | | | | The depends are totally wrong. libunwind does not work with powerpc and i386 as it needs glibc. Instead of duplicating the platforms, just change the dependency. Signed-off-by: Rosen Penev <rosenp@gmail.com>
* uhttpd: add enable instance optionFlorian Eckert2020-01-152-1/+5
| | | | | | | | With this change it is now possible to switch off single instances of the uhttpd config. Until now it was only possible to switch all instances of uhttpd on or off. Signed-off-by: Florian Eckert <fe@dev.tdt.de>
* hostapd: add wpa_strict_rekey supportKyle Copperfield2020-01-151-1/+3
| | | | | | | | | | | The sender domain has a DMARC Reject/Quarantine policy which disallows sending mailing list messages using the original "From" header. To mitigate this problem, the original message has been wrapped automatically by the mailing list software. Rekey GTK on STA disassociate Signed-off-by: Kyle Copperfield <kmcopper@danwin1210.me>
* hostapd: add dtim_period, local_pwr_constraint, spectrum_mgmt_requiredKyle Copperfield2020-01-152-3/+12
| | | | | | | | | | | | | The sender domain has a DMARC Reject/Quarantine policy which disallows sending mailing list messages using the original "From" header. To mitigate this problem, the original message has been wrapped automatically by the mailing list software. Allows dtim_period to be configurable, the default is from hostapd. Adds additional regulatory tunables for power constraint and spectrum managment. Signed-off-by: Kyle Copperfield <kmcopper@danwin1210.me>
* ath79: GL-AR750S (NOR/NAND): limit factory.img kernel size to 2 MBJeff Kletsky2020-01-151-1/+4
| | | | | | | | | | | | | | | | The present U-Boot for GL-AR750S has a limit of 2 MB for kernel size. While sysupgrade can manage kernels up to the present limit of 4 MB, directly flashing a factory.img with a kernel size greater than 2 MB through U-Boot will result in an unbootable device. This commit uses the newly-introduced check-kernel-size build operation to prevent the output of factory.img when the kernel exceeds 2 MB in size, yet permits output of sysupgrade.img as long as the kernel is within KERNEL_SIZE := 4096k Cc: Chuanhong Guo <gch981213@gmail.com> Signed-off-by: Jeff Kletsky <git-commits@allycomm.com>
* build: define check-kernel-size to remove unflashable imagesJeff Kletsky2020-01-151-0/+7
| | | | | | | | | | | | | | | | | | | | | Certain boards have limitations on U-Boot that prevent flashing of images where the kernel size exceeds a threshold, yet sysupgrade can sucessfully manage larger kernels. The current check-size will remove the target artifact if its total size exceeds the threshold. If applied after append-kernel, it will remove the kernel, but the remaining image-assembly steps will continue, resulting in an image without a kernel that is likely unbootable. By defining check-kernel-size, it is now possible to prevent release of such unbootable images through a construct similar to: IMAGE/factory.img := append-kernel | pad-to $$$$(GL_UBOOT_UBI_OFFSET) | \ append-ubi | check-kernel-size $$$$(GL_UBOOT_UBI_OFFSET) Cc: Chuanhong Guo <gch981213@gmail.com> Signed-off-by: Jeff Kletsky <git-commits@allycomm.com>
* dropbear: move failsafe code out of base-filesKyle Copperfield2020-01-154-8/+11
| | | | | | | | | | | | The sender domain has a DMARC Reject/Quarantine policy which disallows sending mailing list messages using the original "From" header. To mitigate this problem, the original message has been wrapped automatically by the mailing list software. Failsafe code of dropbear should be in the dropbear package not the base-files package. Signed-off-by: Kyle Copperfield <kmcopper@danwin1210.me>
* busybox: Include hdparm by default on nas type deviceLinus Walleij2020-01-151-0/+3
| | | | | | | | | NAS devices certainly need to have hdparm to configure things like spin-down time or their disks will be constantly spinning. Just catenate CONFIG_HDPARM=y on these configs. Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
* ath79: wlr-7100: use ath10k-ct smallbuffers package variantTomasz Maciej Nowak2020-01-151-1/+1
| | | | | | | | The memory hacks got removed from ath10k with 1e27bef ("mac80211: remove ath10k_pci memory hacks"). As this device has low amount of RAM, switch to ath-10k-ct small buffers variant, to avoid the OOM Reaper. Signed-off-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
* cryptodev-linux: remove DEFAULT redefinitionEneas U de Queiroz2020-01-151-1/+0
| | | | | | | | The 'DEFAULT:=m if ALL' line prevents the phase1 buildbots from building the package, and users from downloading it, since they use 'ALL_KMODS=y' but 'ALL' is not set. Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
* mac80211: fix MAC address allocations if the local bit is set on the base addrFelix Fietkau2020-01-151-1/+1
| | | | | | | If it's set, don't subtract 1 from the interface index encoded into the first byte of the address Signed-off-by: Felix Fietkau <nbd@nbd.name>
* lantiq: reorganize 02_network board.d filesAdrian Schmutzler2020-01-145-276/+378
| | | | | | | | | | This reorganizes 02_network board.d files based on what's done for ath79 and ramips: Instead of putting all settings into a single big case, the interface/dsl/MAC address setup is put into separate functions with a specific switch case for each of them. This makes grouping of devices much easier and should be easier to read, too. Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
* lantiq: move common DSL setup into lantiq.shAdrian Schmutzler2020-01-146-52/+20
| | | | | | | | | DSL setup consists of the same commands for all subtargets, so move it into a helper function. While at it, remove shebang from library file. Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
* lantiq: split base-files into subtargetsAdrian Schmutzler2020-01-1417-214/+523
| | | | | | | | | | | | | | | | | | This splits the device-dependent base-files into subtarget directories, like done recently for ath79 and ramips. While this increases the overall lines of codes, it will make the code per subtarget smaller and easier to keep track of features and devices. While at it, several variables at the top of 02_network are removed, as they were never changed. The values are put directly into the function calls where they are used. Remove unneeded LED setup from 01_leds, and remove 01_leds entirely for falcon subtarget (as it is not used there). Applies alphabetic reordering to device cases in base-files. Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
* valgrind: do not strip internal preload libraries and executablesJo-Philipp Wich2020-01-141-2/+20
| | | | | | | | | | Implement the suggestions laid out in README_PACKAGERS, mainly by preventing the stripping of the internal vgpreload*.so libraries. Also retain the symbol information of valgrind's private helper executables and enable LTO as suggested in the packagers readme. Signed-off-by: Jo-Philipp Wich <jo@mein.io>
* base-files: fix build for /sbin/pkg_checkXu Wang2020-01-142-3/+3
| | | | | | | | | | Setting CONFIG_IPK_FILES_CHECKSUMS=y causes sha256 checksum files to be included with the packages to check for corruption. This commit fixes two issues: - /sbin/pkg_check was being removed incorrectly if IPK_FILES_CHECKSUMS=y - checksums were being saved in the wrong file Signed-off-by: Xu Wang <xwang1498@gmx.com>
* wrt350nv2-builder: Fix memory leakAndrea Dalla Costa2020-01-141-0/+1
| | | | | | | Add missing call to `free` for variable `buffer` in function `create_bin_file`. Signed-off-by: Andrea Dalla Costa <andrea@dallacosta.me>
* firmware-utils/mktitanimg: fix possible resource leakAndrea Dalla Costa2020-01-141-0/+1
| | | | | | Add missing call to `fclose` for file pointer `nsp_image`. Signed-off-by: Andrea Dalla Costa <andrea@dallacosta.me>
* firmware-utils/mksenaofw: fix possible memory leakAndrea Dalla Costa2020-01-141-0/+2
| | | | | | Add missing calls to `free` for variable `pmodel`. Signed-off-by: Andrea Dalla Costa <andrea@dallacosta.me>
* firmware-utils/mkfwimage: fix possible memory and resource leakAndrea Dalla Costa2020-01-142-0/+6
| | | | | | | | | Add missing calls to `free` for variable `mem`. Add missing call to `fclose` for variable `f`. The same changes were made in both `mkfwimage.c` and `mkfwimage2.c`. Signed-off-by: Andrea Dalla Costa <andrea@dallacosta.me>
* firmware-utils/mkchkimg: fix possible resource leaksAndrea Dalla Costa2020-01-141-0/+8
| | | | | | | | Add missing `fclose` calls for file pointers `kern_fp`, `fs_fp` and `out_fp`. Not closing files could lead to resource leaks. Signed-off-by: Andrea Dalla Costa <andrea@dallacosta.me>
* firmware-utils: fix possible memory leak and resource leakAndrea Dalla Costa2020-01-141-0/+3
| | | | | | | | | | Add missing calls to `free` for variable `buffer`. This could lead to a memory leak. Add missing call to `close` for file pointer `fdin`. This could lead to a resource leak. Signed-off-by: Andrea Dalla Costa <andrea@dallacosta.me>
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-05 10:26:51 +0000