aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* llvm-bpf: move to staging_dir/host/llvm-bpfFelix Fietkau2021-11-042-2/+8
| | | | | | This makes it easier to package it up for the download server Signed-off-by: Felix Fietkau <nbd@nbd.name>
* include/cmake.mk: add support for overriding the host install prefixFelix Fietkau2021-11-041-1/+2
| | | | Signed-off-by: Felix Fietkau <nbd@nbd.name>
* procd: make rpcd dependency conditionalFelix Fietkau2021-11-041-1/+1
| | | | | | Avoids building rpcd when not needed Signed-off-by: Felix Fietkau <nbd@nbd.name>
* build: fix host package unpack if git-src is usedFelix Fietkau2021-11-042-2/+3
| | | | | | | If the package download step is skipped, it needs to be performed for the host build instead. Signed-off-by: Felix Fietkau <nbd@nbd.name>
* build: add SUBDIR_MAKE_DEBUG variable to make it easier to debug package ↵Felix Fietkau2021-11-041-1/+1
| | | | | | makefiles Signed-off-by: Felix Fietkau <nbd@nbd.name>
* bcm4908: fix calculation of new cferam indexSungbo Eo2021-11-041-2/+2
| | | | | | | | | | | | | The arithmetic expansion fails when idx becomes a two digit number. Fix this by relying on expr command. root@OpenWrt:/# echo $(((028 + 0) % 1000)) /bin/ash: arithmetic syntax error root@OpenWrt:/# echo $(($(expr 028 + 0) % 1000)) 28 Fixes: a6a0b252baa6 ("bcm4908: add sysupgrade support") Signed-off-by: Sungbo Eo <mans0n@gorani.run>
* omcproxy: bump to git HEADStijn Tintel2021-11-041-3/+3
| | | | | | bfba2aa groups: use uloop_timeout_remaining64 Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
* procd: bump to git HEADStijn Tintel2021-11-041-3/+3
| | | | | | 0ee8e73 trigger: use uloop_timeout_remaining64 Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
* rpcd: bump to git HEADStijn Tintel2021-11-041-3/+3
| | | | | | | 20bf958 session: use uloop_timeout_remaining64 d11ffe9 session: use blobmsg_get_u64 for RPC_DUMP_EXPIRES Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
* libubox: bump to git HEADStijn Tintel2021-11-041-3/+3
| | | | | | | | | 123e976 uloop: restore return type of uloop_timeout_remaining 3344157 uloop: add uloop_timeout_remaining64 c87d3e1 lua/uloop: use uloop_timeout_remaining64 c86a894 uloop: deprecate uloop_timeout_remaining Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
* hostapd: support qos_map_set without CONFIG_INTERWORKINGFelix Fietkau2021-11-041-0/+112
| | | | | | This feature is useful on its own even without full interworking support Signed-off-by: Felix Fietkau <nbd@nbd.name>
* bcm53xx: add first 5.17 DTS changesRafał Miłecki2021-11-045-7/+177
| | | | Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
* libubox: bump to git HEADStijn Tintel2021-11-041-4/+4
| | | | | | | be3dc72 uloop: avoid integer overflow in tv_diff Fixes: FS#3943 Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
* config: Activate SECCOMP also on MIPS 64Hauke Mehrtens2021-11-031-1/+1
| | | | | | | | This activates SECCOMP also on mips64 and mips64el. This was working fine in a basic test in qemu. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* toolchain: Allow sanitizer on mips and mipselHauke Mehrtens2021-11-031-2/+2
| | | | | | | Support for libsanitizer on MIPS 32 and MIPSEL 32 was added with GCC 9. MIPS 64 and ARC are still not supported. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* binutils: Update to version 2.37Hauke Mehrtens2021-11-031-2/+2
| | | | | | This matches the version used in the toolchain. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* strace: Update to version 5.14Hauke Mehrtens2021-11-032-8/+9
| | | | | | | Explicitly deactivate libselinux, otherwise we get a hard dependency to libselinux. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* valgrind: Activate also on MIPS 64Hauke Mehrtens2021-11-031-1/+1
| | | | | | | | This activates valgrind also on mips64 and mips64el. This was working fine in a basic test in qemu. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* valgrind: Update to version 3.18.1Hauke Mehrtens2021-11-033-8/+139
| | | | | | | | | | | | This version has some improvements for musl. This version works fine for me on MIPS 32 BE without MIPS16 even on musl. The additional patch is needed to make valgrind use the correct syscall numbers for new syscalls like clock_gettime64. The MIPS architecture uses special syscall ranges which are different from most other systems. The patch is pending upstream: https://bugs.kde.org/show_bug.cgi?id=444781 Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* gdb: Update to version 11.1Hauke Mehrtens2021-11-033-10/+10
| | | | | | GDB 11.1 now depends on gmp. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* toolchain: gdb: Update to version 11.1Hauke Mehrtens2021-11-032-22/+5
| | | | | | | | | | The removed patch was already applied upstream. gdb now mandatory depends on gmp, tell configure where to find it explicitly. We already build gmp in the tools directory for gcc. Also make it use mpfr and mpc as we also build both of them. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* ipq40xx: disable some devices due to kernel sizeRobert Marko2021-11-031-0/+9
| | | | | | | | | | Disable some of the ipq40xx devices due to their kernel size limitations. These devices fail to build with kernel 5.10 and full buildbot config. Signed-off-by: Robert Marko <robert.marko@sartura.hr> [keep gl-b1300/gl-s1300 enabled, extend commit message] Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
* ipq40xx: switch to kernel 5.10Adrian Schmutzler2021-11-031-1/+1
| | | | | | | | | | There have been enough tests and new developments require the new kernel, so let's update it. There is a bunch of devices that do not build anymore due to kernel size limitations. These are disabled in a subsequent commit. Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
* hostapd: add wmm qos map set by defaultFelix Fietkau2021-11-031-1/+6
| | | | | | | | | | This implements the mapping recommendations from RFC8325, with an update from RFC8622. This ensures that DSCP marked packets are properly sorted into WMM classes. The map can be disabled by setting iw_qos_map_set to something invalid like 'none' Signed-off-by: Felix Fietkau <nbd@nbd.name>
* vti: squash vtiv4 and vtiv6 packages into vtiFlorian Eckert2021-11-031-36/+7
| | | | | | This change adds the same package behaviour as gre package. Signed-off-by: Florian Eckert <fe@dev.tdt.de>
* qosify: add missing config option for nat supportFelix Fietkau2021-11-032-0/+2
| | | | Signed-off-by: Felix Fietkau <nbd@nbd.name>
* qosify: include nls.mk to avoid build error with full NLS support enabledFelix Fietkau2021-11-031-0/+1
| | | | Signed-off-by: Felix Fietkau <nbd@nbd.name>
* imx: remove obsolete Kernel 5.4Piotr Dymacz2021-11-0316-3122/+0
| | | | Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* imx: update subtargets default kernel configs for 5.10Piotr Dymacz2021-11-032-2/+10
| | | | Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* imx: make kernel 5.10 default versionPiotr Dymacz2021-11-031-2/+1
| | | | Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* imx: introduce 'cortexa7' subtargetPiotr Dymacz2021-11-036-1/+73
| | | | | | | | | | | This adds some essential files required by new 'cortexa7' subtarget, dedicated for Cortex-A7 based NXP i.MX series. For now, the kernel config-default focuses only on the i.MX 6UL family, as the following changeset will introduce support for i.MX 6ULL based device. Support for more platforms (e.g. i.MX 7) might be enabled later, while adding more devices. Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* uboot-imx: set BUILD_SUBTARGET to 'cortexa9'Piotr Dymacz2021-11-031-0/+3
| | | | | | All currently supported devices belong to the imx/cortexa9 subtarget. Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* uboot-envtools: move imx to imx_cortexa9Piotr Dymacz2021-11-031-0/+0
| | | | | | | | Subtarget-specific files under 'uboot-envtools' package are supported since 6f3a05ebb0 ("uboot-envtools: support uci-default config also per subtargets"). Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* imx: split into arch-specific subtargetsPiotr Dymacz2021-11-0311-333/+286
| | | | | | | | | | Modern NXP i.MX series includes several different families, based on single- or multi-core Arm Cortex-A CPUs. To be able to support more families within a single target, we split the 'imx' in arch-specific subtargets, starting with 'cortexa9' for the Cortex-A9 based i.MX 6, already supported by the original 'imx6' target. Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* uboot-imx6: rename to 'uboot-imx'Piotr Dymacz2021-11-034-1/+1
| | | | Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* uboot-envtools: rename 'imx6' to 'imx'Piotr Dymacz2021-11-031-1/+1
| | | | Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* kobs-ng: update dependencies after 'imx6' -> 'imx' renamePiotr Dymacz2021-11-031-1/+1
| | | | Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* kernel: update dependencies after 'imx6' -> 'imx' renamePiotr Dymacz2021-11-037-12/+12
| | | | Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* imx6: rename target to 'imx'Piotr Dymacz2021-11-0333-5/+5
| | | | | | | | | | | This is first step in migrating to a generic i.MX target which in the next steps will also get divided into arch-specific subtargets. In the result, this will make it possible to support, within a single target, also other modern NXP i.MX families, like the i.MX 7, i.MX 8 or recently introduced i.MX 9. Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* imx6: refresh and update target kernel configsPiotr Dymacz2021-11-032-21/+9
| | | | | | | | | This is a minor extension of the commit 26ae69fd03 ("imx6: refresh kernel config with 5.10 symbols"), with correct and full disable of the Arm Cortex-A7 based i.MX 6UL and 6UL{L,Z} families support and re-enable of the Cortex-A9 based i.MX 6L{S,X}. Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* imx6: image: fix Gateworks Ventana boot script filenamePiotr Dymacz2021-11-032-1/+1
| | | | | | | | | | | | | U-Boot for the Gateworks Ventana includes filename of the boot script in the default/embedded environment (see 'include/configs/gw_ventana.h' in the U-Boot sources). This restores the old boot script filename ('6x_bootscript-ventana'), making Ventana boards boot again. Fixes: 8dba71dd33 ("imx6: image: drop BOOT_SCRIPT and fix DEVICE_NAME") Reported-by: Koen Vandeputte <koen.vandeputte@ncentric.com> Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* ath79: add support for Netgear R6100Piotr Dymacz2021-11-035-0/+253
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Netgear R6100 is a dual-band Wi-Fi 5 (AC1200) router based on Qualcomm Atheros (AR9344 + QCA9882) platform. Support for this device was first introduced in 15f6f67d18 (ar71xx). FCC ID: PY312400225. Specifications: - Atheros AR9344 (560 MHz) - 128 MB of RAM (DDR2) - 128 MB of flash (parallel NAND) - 2T2R 2.4 GHz Wi-Fi (AR9344) - 2T2R 5 GHz Wi-Fi (QCA9882) - 5x 10/100 Mbps Ethernet (AR9344) - 4x internal antenna - 1x USB 2.0 (GPIO-controlled power) - 6x LED, 3x button (reset, Wi-Fi, WPS) - UART (4-pin, 2.54 mm pitch) header on PCB - 1x mechanical power switch - DC jack for main power input (12 V) WARNING: sysupgrade from older stable releases is not possible, fresh installation (via vendor's GUI or TFTP based recovery) is required. Reason for that is increased kernel partition size. Installation: Use the 'factory' image under vendor's GUI or via TFTP U-Boot recovery. You can use the 'nmrpflash' tool at a boot time, before kernel is loaded or start it manually by pressing the 'reset' button for ~20 seconds from powering up the device (U-Boot will start TFTP server on 192.168.1.1, use TFTP client to send the image). Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com> Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu> Signed-off-by: Michael Pratt <mcpratt@pm.me> Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
* bcm4908: add testing support for kernel 5.10Rafał Miłecki2021-11-032-1/+3
| | | | | | | | | | CONFIG_COMPAT_32BIT_TIME is a new symbol and has to be set to avoid: Provide system calls for 32-bit time_t (COMPAT_32BIT_TIME) [N/y/?] (NEW) CONFIG_RELOCATABLE needs to be enabled to make kernel start booting. That raises a question: do we really need CONFIG_EXPERT=y ? Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
* bcm4908: prepare kernel 5.10 supportRafał Miłecki2021-11-0364-0/+7477
| | | | | | It compiles but *doesn't* boot so it isn't enabled yet. Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
* pistachio: Take bootargs from device treeHauke Mehrtens2021-11-021-2/+2
| | | | | | | | | | | | | | The boot arguments are copied into the device tree by the boot loader and taken from the device tree by the kernel. The code which takes the boot arguments from the different sources was reworked with kernel 5.5. We have to activate CONFIG_MIPS_CMDLINE_DTB_EXTEND to take the boot arguments from the device tree. This makes the system boot on the board again. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* pistachio: Make kernel 5.10 patches applyHauke Mehrtens2021-11-026-94/+37
| | | | | | | Make the patches apply on kernel 5.10 and refresh the patches and the kernel configuration on top of kernel 5.10. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* pistachio: Copy kernel 5.4 patches to 5.10Hauke Mehrtens2021-11-0215-0/+1240
| | | | | | | This just copies the patches and the configuration from kernel 5.4 to kernel 5.10. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* pistachio: Fix FIT image configuration nameHauke Mehrtens2021-11-021-0/+1
| | | | | | | | | | The pistachio U-Boot expects a default configuration with the name config@1 in the FIT image. The default was changed in OpenWrt some months ago. This makes the board boot again. Fixes: 9f714398e060 ("build: use config-1 instead of config@1 as default") Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* kernel: Set some options in generic configurationHauke Mehrtens2021-11-0210-9/+2
| | | | | | | Add CONFIG_USB_ETH and CONFIG_PWM_JZ4740 to generic kernel configuration. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* strace: remove code coverage makefile varRosen Penev2021-11-022-1/+11
| | | | | | | | | It relies on a custom ax_code_coverage.m4 file included with strace. Unfortunately, this conflicts with the one included with autoconf-macros. Instead of creating a huge patch to fix it, just remove the variable as code coverage is not used here. Signed-off-by: Rosen Penev <rosenp@gmail.com>