aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* mediatek: backport upstream mediatek patchesJohn Crispin2018-05-24130-99/+21978
| | | | Signed-off-by: John Crispin <john@phrozen.org>
* wolfssl: update to version 3.14.4Daniel Golle2018-05-243-149/+6
| | | | | | | | Use download from github archive corresponding to v3.14.4 tag because the project's website apparently only offers 3.14.0-stable release downloads. Signed-off-by: Daniel Golle <daniel@makrotopia.org>
* ustream-ssl: fix build against wolfSSLDaniel Golle2018-05-241-3/+3
| | | | | | | | | | | | commit 39a6ce205d (ustream-ssl: Enable ECDHE with OpenSSL.) broke build against wolfSSL because wolfSSL doesn't (yet) support SSL_CTX_set_ecdh_auto() of the OpenSSL API. Fix this in ustream-ssl: 189cd38b41 don't use SSL_CTX_set_ecdh_auto with wolfSSL Signed-off-by: Daniel Golle <daniel@makrotopia.org>
* tools: zlib: do not hardcode the install prefix in zlib.pcJo-Philipp Wich2018-05-241-0/+14
| | | | | | | | | | | | | | | | | Our pkg-config wrapper relies on the ability to redefine the $prefix and $exec_prefix variables in order to construct proper search paths relative to the build environment. Patch the .pc file template to construct libdir, sharedlibdir and includedir relative to the ${prefix} variable so that it can be overridden as needed. This also fixes the libxml2/host build issue raised at https://github.com/openwrt/packages/issues/6073 - it was caused by libxml2's configure picking up a wrong host search path through zlib.pc, letting it include the wrong endian.h, causing spurious member redeclaration errors in system headers. Signed-off-by: Jo-Philipp Wich <jo@mein.io>
* fstools: Add the new options available in the menuconfigPierre Lebleu2018-05-241-0/+16
| | | | | | | | Mounting using the zlib compression and mounting with full access accounting are now available in the menuconfig. Signed-off-by: Pierre Lebleu <pme.lebleu@gmail.com>
* ath79: initial tl-wr1043nd v1 supportLucian Cristian2018-05-243-14/+93
| | | | | Signed-off-by: Lucian Cristian <lucian.cristian@gmail.com> Signed-off-by: John Crispin <john@phrozen.org>
* kernel: add reset control support to rtl8366 driverJohn Crispin2018-05-244-8/+25
| | | | Signed-off-by: John Crispin <john@phrozen.org>
* ath79: fix adjust_link callback for ar9130John Crispin2018-05-241-0/+2
| | | | Signed-off-by: John Crispin <john@phrozen.org>
* ath79: make ahb wifi workJohn Crispin2018-05-245-4/+385
| | | | Signed-off-by: John Crispin <john@phrozen.org>
* Revert "dnsmasq: use "hostsdir" instead of "addn-hosts""Hans Dedecker2018-05-241-1/+1
| | | | | | | | | This reverts commit a03035dad198cd4b51645ceb43c1170f9cf95f16 as it has several issues: -Host file is located in a directory which is not unique per dnsmasq instance -odhcpd writes host info into the same directory but still sends a SIGHUP to dnsmasq Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
* base-files: rework _ucidef_set_interface to be more genericRoman Yeryomin2018-05-242-53/+27
| | | | | | | | | | | | This is a rework of previously submitted patch reworking ucidef_set_interface_raw [1]. Here, keep the idea but instead make _ucidef_set_interface more generic and use it instead of ucidef_set_interface_raw. Also change the users like ucidef_set_interface_lan and others. [1] https://patchwork.ozlabs.org/patch/844961/ Signed-off-by: Roman Yeryomin <roman@advem.lv>
* kernel: bump 4.14 to 4.14.43Koen Vandeputte2018-05-2469-966/+969
| | | | | | | | | | Refreshed all patches Compile-tested on: cns3xxx, imx6, x86_64 Runtime-tested on: cns3xxx, imx6, x86_64 Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com> Tested-by: Michael Yartys <michael.yartys@protonmail.com>
* kernel: bump 4.9 to 4.9.102Koen Vandeputte2018-05-241-2/+2
| | | | | | | | | Refreshed all patches Compile-tested on: ar71xx Runtime-tested on: ar71xx Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
* ramips: Fix up GnuBee PC1 DTS file a littleRosen Penev2018-05-241-7/+1
| | | | | | | | There's nothing connected to i2c on this board, so remove it. Also edited the gpio group to match the PC2 as they're the same. Signed-off-by: Rosen Penev <rosenp@gmail.com>
* ramips: Add back some non-mt7621 code that staging removedRosen Penev2018-05-241-1/+17
| | | | | | Staging is meant only for mt7621 but for OpenWrt more is needed. Signed-off-by: Rosen Penev <rosenp@gmail.com>
* brcm47xx: add switch port mapping to Asus WL-500WMirko Parthey2018-05-241-1/+1
| | | | | | | | | | | Switch ports 0..3 are connected to external ports LAN{1..4} in sequence, switch port 4 is not used, and switch port 5 is connected to the CPU. The WAN port is attached to the CPU's second network interface; it has no connection to the internal switch. Reuse the "Dell TrueMobile 2300" entry, which describes the same mapping. Signed-off-by: Mirko Parthey <mirko.parthey@web.de>
* kernel: mark all v3.18 and v4.4 kernels are source-onlyJohn Crispin2018-05-247-7/+7
| | | | | | These will be removed if not updated to v4.4 Signed-off-by: John Crispin <john@phrozen.org>
* dnsmasq: use "hostsdir" instead of "addn-hosts"Christian Schoenebeck2018-05-231-1/+1
| | | | | | | 1.) "addn-hosts" per default point to a file (but it supports directory) 2.) "hostsdir" only support directory with the additional benefit: New or changed files are read automatically. Signed-off-by: Christian Schoenebeck <christian.schoenebeck@gmail.com>
* toolchain/glibc: update to latest 2.26 commitHans Dedecker2018-05-231-2/+2
| | | | | | | | | | | | | | | af7519f7b3 Fix path length overflow in realpath [BZ #22786] 365722ace6 Fix stack overflow with huge PT_NOTE segment [BZ #20419] be056fae3b Fix blocking pthread_join. [BZ #23137] 02f0dd83a4 Fix signed integer overflow in random_r (bug 17343). 3241353ab2 i386: Fix i386 sigaction sa_restorer initialization (BZ#21269) 677e6d13e0 [BZ #22342] Fix netgroup cache keys. 71d339cb86 Fix i386 memmove issue (bug 22644). 31e2d15b80 Fix crash in resolver on memory allocation failure (bug 23005) 1f7c4748d6 getlogin_r: return early when linux sentinel value is set 7e7a5f0bcd resolv: Fully initialize struct mmsghdr in send_dg [BZ #23037] Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
* uboot-zynq: fix build on hosts lacking pkg-configJo-Philipp Wich2018-05-231-0/+3
| | | | | | | | | | | | The uboot-mvebu package incorrectly used the host pkg-config for the tool build parts, which broke the build on systems lacking pkg-config and only worked by accident on those that have it installed. Export the host-build specific environment variables for the uboot build to redirect pkg-config invocations to our staged host build pkg-config in buildroot. Signed-off-by: Jo-Philipp Wich <jo@mein.io>
* uboot-mxs: fix build on hosts lacking pkg-configJo-Philipp Wich2018-05-231-0/+3
| | | | | | | | | | | | The uboot-mvebu package incorrectly used the host pkg-config for the tool build parts, which broke the build on systems lacking pkg-config and only worked by accident on those that have it installed. Export the host-build specific environment variables for the uboot build to redirect pkg-config invocations to our staged host build pkg-config in buildroot. Signed-off-by: Jo-Philipp Wich <jo@mein.io>
* uboot-mvebu: fix build on hosts lacking pkg-configJo-Philipp Wich2018-05-231-0/+3
| | | | | | | | | | | | The uboot-mvebu package incorrectly used the host pkg-config for the tool build parts, which broke the build on systems lacking pkg-config and only worked by accident on those that have it installed. Export the host-build specific environment variables for the uboot build to redirect pkg-config invocations to our staged host build pkg-config in buildroot. Signed-off-by: Jo-Philipp Wich <jo@mein.io>
* ramips: fix to mt7620a and add support for i2c on WD03Matthias Badaire2018-05-221-1/+5
| | | | | | | | | There was an error on initial commit, the proper soc is mt7620n (which is more limited than mt7620a). Moreover, there is a battery management controller connected to the i2c port of the mt7620n. I have a small piece of i2c code to get battery level coming. Signed-off-by: Matthias Badaire <mbadaire@gmail.com>
* ath79: preliminary support for TP-Link WDR3600 / WDR4300 (AR9344)Johann Neuhauser2018-05-2210-2/+548
| | | | | | | | | | | | | | | | working: - leds - buttons - lan / wan - usb (hub port 1 + 2) - wifi 5g - sysupgrade - ... not working: - wifi 2g Signed-off-by: Johann Neuhauser <johann@it-neuhauser.de>
* kernel: bump 4.14 to 4.14.42Koen Vandeputte2018-05-227-23/+23
| | | | | | | | | Refreshed all patches Compile-tested on: cns3xxx, imx6, x86_64 Runtime-tested on: cns3xxx, imx6, x86_64 Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
* kernel: bump 4.9 to 4.9.101Koen Vandeputte2018-05-224-17/+17
| | | | | | | | | Refreshed all patches Compile-tested on: ar71xx Runtime-tested on: ar71xx Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
* ustream-ssl: update to latest git HEADJohn Crispin2018-05-221-4/+4
| | | | | | | 5322f9d mbedtls: Fix setting allowed cipher suites e8a1469 mbedtls: Add support for a session cache Signed-off-by: John Crispin <john@phrozen.org>
* mbedtls: Deactivate platform abstractionHauke Mehrtens2018-05-221-0/+9
| | | | | | | | This makes mbedtls use the POSIX API directly and not use the own abstraction layer. The size of the ipkg decreased by about 100 bytes. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* mbedtls: Activate the session cacheHauke Mehrtens2018-05-221-10/+0
| | | | | | | | | | This make sit possible to store informations about a session and reuse it later. When used by a server it increases the time to create a new TLS session from about 1 second to less than 0.1 seconds. The size of the ipkg file increased by about 800 Bytes. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* mbedtls: update to version 2.9.0Hauke Mehrtens2018-05-226-31/+34
| | | | | | | | The soversion was changed in this version again and is now aligned with the 2.7.2 version. The size of the ipkg file stayed mostly the same. Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
* package sysfsutils: add support for sysfs settings at bootRodolfo Giometti2018-05-224-0/+83
| | | | | | This patch is based on sysfsutils package's behaviour on Debian OS. Signed-off-by: Rodolfo Giometti <giometti@linux.it>
* ipq806x: drop "mtd: nand: add Winbond manufacturer and chip"Stefan Lippers-Hollmann2018-05-221-38/+0
| | | | | | The W25N01GV NAND is currently not used in any ipq806x device. Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: drop linux 4.9 supportStefan Lippers-Hollmann2018-05-2287-29438/+0
| | | | Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: switch to linux 4.14Stefan Lippers-Hollmann2018-05-221-1/+1
| | | | Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: remove spi-nor from r7800 dtsPavel Kubelun2018-05-222-90/+0
| | | | | | | | There's no spi-nor in R7800, so disable unequipped interfaces. Signed-off-by: Pavel Kubelun <be.dissent@gmail.com> [slh: rebase for kernel v4.14 as well] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: remove rpm pinctrl from board dtsPavel Kubelun2018-05-2216-80/+0
| | | | | | | | | These pins seem to be used by hw exclusively, so claiming it in kernel causes an error in syslog in k4.14+. Signed-off-by: Pavel Kubelun <be.dissent@gmail.com> [slh: rebase for kernel v4.14 as well] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: fix EA8500 switch controlPavel Kubelun2018-05-222-8/+36
| | | | | | | | | | | | | | | | | EA8500 has pcie2 slot unequipped. By EA8500 hw design default pcie2 reset gpio (gpio63) is used to reset the switch. That's why enabling pcie2 brings the switch into a working state. So let's just control the gpio63 without enabling the pcie2 slot. We have to remove the pcie2_pins node so the gpio63 is not defined twice. Because pcie2 node has a reference to pcie2_pins we have to remove it as well. Signed-off-by: Pavel Kubelun <be.dissent@gmail.com> [slh: rebase for kernel v4.14 as well] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: move mmc specific nodes into v1.0 dtsiPavel Kubelun2018-05-224-150/+150
| | | | | | | | | These nodes are common for all revisions so put it into SoC v1.0 dtsi file. Signed-off-by: Pavel Kubelun <be.dissent@gmail.com> [slh: rebase for kernel v4.14 as well] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: fix pcie tx termination offsetPavel Kubelun2018-05-222-0/+24
| | | | | | | | | | | | | | According to GPL tarballs and QSDK related branch tx termination offset for ipq8064 SoC version >= 2.0 should be equal to 0 and not 7. https://github.com/paul-chambers/netgear-r7800/blob/master/git_home/linux.git/sourcecode/arch/arm/mach-msm/board-ipq806x.c#L1682-L1685 Fix this. Signed-off-by: Pavel Kubelun <be.dissent@gmail.com> [slh: rebase for kernel v4.14 as well] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: apply updated USB PHY settings to v2.0 SoCPavel Kubelun2018-05-224-22/+34
| | | | | | | | | | | | | | | | | | | USB PHY power settings introduced for ipq8065 SoC with commit 644a0d5 "ipq8065: adjust SS USB PHY power settings" According to that commit msg and in correspondence to GPL tarballs and related QSDK branch those settings are applied to ipq8064 SoCs of version >= 2.0. https://github.com/paul-chambers/netgear-r7800/blob/master/git_home/linux.git/sourcecode/arch/arm/mach-msm/board-ipq806x.c#L2507-L2514 Now as we have clarified that mass market boards are of SoC v2.0 move those USB PHY settings from ipq8065 (v3.0 SoC) dtsi to ipq8064 v2.0 dtsi. Signed-off-by: Pavel Kubelun <be.dissent@gmail.com> [slh: rebase for kernel v4.14 as well] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: reference ipq8065 as ipq8064 v3.0Pavel Kubelun2018-05-228-8/+6
| | | | | | | | | | | ipq8065 is ipq8064 v3.0 > socinfo_init: v6, id=280, ver=3.0, raw_id=17, raw_ver=17, hw_plat=0, hw_plat_ver=65536 Include dtsi accordingly and remove the unneeded qcom-ipq8065-v1.0.dtsi Signed-off-by: Pavel Kubelun <be.dissent@gmail.com> [slh: rebase for kernel v4.14 as well] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: move mass market ipq8064 to v2 dtsiPavel Kubelun2018-05-2210-10/+10
| | | | | | | | | | | | | | | | | According to OEM bootlog entry mass market devices are ipq8064 SoC v2.0: > socinfo_init: v6, id=202, ver=2.0, raw_id=2064, raw_ver=2064, hw_plat=0, hw_plat_ver=65536 I've checked C2600, EA8500 and VR2600v but couldn't find other boards bootlog. I think it's safe to assume that other boards are also v2.0. R7500 may be an exception because it was the first device to hit the market. So switch to v2.0 dtsi. Signed-off-by: Pavel Kubelun <be.dissent@gmail.com> [slh: rebase for kernel v4.14 as well] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: introduce ipq8064 SoC v2 dtsiPavel Kubelun2018-05-222-0/+2
| | | | | | | | | | | | | | | According to QCA internal numbering there are 3 versions of ipq8064/5 SoC: ipq8064 v1.0 - probably ipq8064 evaluation boards only ipq8064 v2.0 - probably ipq8064 mass market boards only ipq8064 v3.0 - aka ipq8065, boards based on ipq8065. Each next revision includes configuration differences from previous revision and adds something new. Signed-off-by: Pavel Kubelun <be.dissent@gmail.com> [slh: rebase for kernel v4.14 as well] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* kernel: iqp806x low latency kernel does not bootMarc Benoit2018-05-222-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | It keeps failing (R7800) with the stack trace below BUG: scheduling while atomic: kworker/0:1/26/0x00000002 (unwind_backtrace) from [<c02121d0>] (show_stack+0x10/0x14) (show_stack) from [<c03932e4>] (dump_stack+0x7c/0x9c) (dump_stack) from [<c0239b90>] (__schedule_bug+0x5c/0x80) (__schedule_bug) from [<c05b7260>] (__schedule+0x50/0x3f4) (__schedule) from [<c05b76a8>] (schedule+0xa4/0xd4) (schedule) from [<c05ba430>] (schedule_hrtimeout_range_clock+0xc8/0x100) (schedule_hrtimeout_range_clock) from [<c05ba480>] (schedule_hrtimeout_range+0x18/0x20) (schedule_hrtimeout_range) from [<c05b9f78>] (usleep_range+0x48/0x50) (usleep_range) from [<c03f333c>] (__clk_hfpll_enable+0x44/0xd0) (__clk_hfpll_enable) from [<c03f3474>] (clk_hfpll_set_rate+0xac/0xc4) (clk_hfpll_set_rate) from [<c03ec390>] (clk_change_rate+0xf4/0x1fc) (clk_change_rate) from [<c03ec510>] (clk_core_set_rate_nolock+0x78/0x94) (clk_core_set_rate_nolock) from [<c03ec54c>] (clk_set_rate+0x20/0x30) (clk_set_rate) from [<c0424168>] (dev_pm_opp_set_rate+0x190/0x26c) (dev_pm_opp_set_rate) from [<c04a8548>] (set_target+0x40/0x108) (set_target) from [<c04a4.140>] (__cpufreq_driver_target+0x3f4/0x488) (__cpufreq_driver_target) from [<c04a7494>] (od_dbs_timer+0xcc/0x154) (od_dbs_timer) from [<c04a7998>] (dbs_work_handler+0x2c/0x54) (dbs_work_handler) from [<c02309e8>] (process_one_work+0x1c0/0x2f0) (process_one_work) from [<c02319a8>] (worker_thread+0x2a4/0x404) (worker_thread) from [<c0235944>] (kthread+0xd8/0xe8) (kthread) from [<c020eef0>] (ret_from_fork+0x14/0x24) Signed-off-by: Marc Benoit <marcb62185@gmail.com> [slh: rebase for kernel v4.14 as well] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: define KERNEL_SIZE in KB instead of byte for the image generationStefan Lippers-Hollmann2018-05-221-5/+5
| | | | Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: increase kernel partition size for the Netgear Nighthawk X4 R7500v2Stefan Lippers-Hollmann2018-05-223-7/+7
| | | | | | | | | | | | | | | | | | Starting with kernel 4.14 and gcc 7, the kernel doesn't fit into the 2 MB reserved for the kernel partition by the OEM firmware anymore. This patch increases the kernel partition from 2 MB to 4 MB, at the expense of the rootfs, for all supported kernels. WARNING: This breaks sysupgrade! Flashing a firmware containing this changed partitioning from an older image with 2 MB reserved for the kernel partion requires the tftp recovery procedure, thereby losing all settings. This change follows the functional example of the Netgear r7800, but has not been runtime tested on a Netgear Nighthawk X4 R7500v2. Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
* ipq806x: increase kernel partition size for the Netgear Nighthawk X4 R7500Stefan Lippers-Hollmann2018-05-223-7/+7
| | | | | | | | | | | | | | | | | | | Starting with kernel 4.14 and gcc 7, the kernel doesn't fit into the 2 MB reserved for the kernel partition by the OEM firmware anymore. This patch increases the kernel partition from 2 MB to 4 MB, at the expense of the rootfs, for all supported kernels. WARNING: This breaks sysupgrade! Flashing a firmware containing this changed partitioning from an older image with 2 MB reserved for the kernel partion requires the tftp recovery procedure, thereby losing all settings. This change follows the functional example of the Netgear r7800, but has not been runtime tested on a Netgear Nighthawk X4 R7500 Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Jonas Gorski <jonas.gorski@gmail.com>
* ipq806x: increase kernel partition size for the Netgear Nighthawk X4 D7800Stefan Lippers-Hollmann2018-05-223-7/+7
| | | | | | | | | | | | | | | | | | | Starting with kernel 4.14 and gcc 7, the kernel doesn't fit into the 2 MB reserved for the kernel partition by the OEM firmware anymore. This patch increases the kernel partition from 2 MB to 4 MB, at the expense of the rootfs, for all supported kernels. WARNING: This breaks sysupgrade! Flashing a firmware containing this changed partitioning from an older image with 2 MB reserved for the kernel partion requires the tftp recovery procedure, thereby losing all settings. This change follows the functional example of the Netgear r7800, but has not been runtime tested on a Netgear Nighthawk X4 D7800. Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Tathagata Das <tathagata@alumnux.com>
* ipq806x: increase kernel partition size for the TP-Link Archer VR2600vStefan Lippers-Hollmann2018-05-223-7/+7
| | | | | | | | | | | | | | | | | | The default image does not fit 2 MB anymore, expand kernel partition to 3 MB. Upgrading should work transparently via sysupgrade in both directions. Another option would be to merge "kernel" and "rootfs" into a single "firmware" partition using MTD_SPLIT_TPLINK_FW, but just changing the sizes of the existing partitioning has been deemed safer in the absence of an actual runtime test on an affected device; the maximum for rootfs changes from 10.4 MB to 9.4 MB. This change follows the example for the TP-Link Archer C2600, but has not been runtime tested on a TP-Link Archer VR2600v. Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> Cc: Sebastian Quilitz <zeraphim@x-pantion.de>
* ipq806x: increase kernel partition size for the TP-Link Archer C2600Stefan Lippers-Hollmann2018-05-222-5/+10
| | | | | | | | | | | | | | | | | | The default image does not fit 2 MB anymore, expand os-image partition to 4 MB. Upgrading works transparently via sysupgrade in both directions. Another option would have been to merge "os-image" and "rootfs" into a single "firmware" partition using MTD_SPLIT_TPLINK_FW, but just changing the sizes of the existing partitioning has been deemed safer and actually tested on an affected device; the maximum for rootfs changes from 27 MB to 25 MB. Run-tested on TP-Link Archer C2600. Signed-off-by: Joris de Vries <joris@apptrician.nl> [slh: extend comments and commit message, rename rootfs] Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>