Backport a hot off the press upstream kernel ATM fix:
Preserve value of skb->truesize when accounting to vcc
"There's a hack in pskb_expand_head() to avoid adjusting skb->truesize
for certain skbs. Ideally it would cover ATM too. It doesn't. Just
stashing the accounted value and using it in atm_raw_pop() is probably
the easiest way to cope."
The issue was exposed by upstream with:
commit 14afee4b6092fde451ee17604e5f5c89da33e71e
Author: Reshetova, Elena <elena.reshetova@intel.com>
Date: Fri Jun 30 13:08:00 2017 +0300
net: convert sock.sk_wmem_alloc from atomic_t to refcount_t
But an earlier commit left the ticking timebomb:
158f323b9868 ("net: adjust skb->truesize in pskb_expand_head()
Sincerest thanks to Mathias Kresin <dev@kresin.me> for debugging
assistance and to David Woodhouse <dwmw2@infradead.org> for further
guidance, cajoling & patience in interpreting the debug I was giving him
and producing a fix!
Fixes FS#1567
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
(cherry picked from commit d600de3ddd)
Shorten the version string in Netgear factory image header in order
to enable u-boot TFTP recovery flash mode to work again.
Strip 'r7210-14cb05909a' into 'r7210' in the Netgear image header
by removing the hash (anything after "-").
background:
Some Netgear routers have recently been unable to flash Openwrt
factory image with the TFTP recovery flash mode provided by Netgear
u-boot. That is due to over-long Openwrt version string overflowing
into the router type string in u-boot code. Modern git versions
produce 10-digit short hashes for the Openwrt main repo, and that
causes the version string to be too long in the image header,
breaking the image ID verification by the TFTP flash routine.
(Other option could be to force a shorter hash in scripts/getver.sh,
but as the problem only concerns Netgear routers, let's patch just
them.)
More detailed explanations in FS#1583
Tested with WNDR3800
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
(cherry picked from commit dcfe2a461e)
The kernel patch *-mips_module_reloc.patch breaks dynamic ftrace as
dynamic ftrace depends on -mlong-calls.
See http://patchwork.linux-mips.org/patch/675/
Thus we always set -mlong-calls if the kernel is being
compiled with dynamic ftrace support.
Signed-off-by: Vincent Wiemann <webmaster@codefetch.de>
(cherry picked from commit 076d2ea682)
This fixup ip align in essedma driver rx path
see cat /proc/cpu/alignment
which reports alignment-fixups without this fix.
Signed-off-by: Chen Minqiang <ptpt52@gmail.com>
(cherry picked from commit 8f804f42d5)
Marvell ahci hardware requires a workaround to prevent eSATA failures
on hotplug/reset when used with multi-bay external enclosures.
Errata Ref#226 - SATA Disk HOT swap issue when connected through Port
Multiplier in FIS-based Switching mode.
These patches backport the workaround from 4.17.
Signed-off-by: Jeremiah McConnell <miah@miah.com>
(cherry picked from commit e820455198)
Add procd file listeners to check files in `/etc/crontabs/`.
Also unified a bit the function style.
Signed-off-by: Paul Spooren <mail@aparcar.org>
(cherry picked from commit cbf69fb2ad)
zram.ko needs CONFIG_BLK_DEV activated and it is by default for all
other targets in OpenWrt.
This makes zram.ko compile again.
Compile tested only.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 6745af9a0d)
Neon and vfpv4 are mandatory extensions in the ARM64 instruction set
now, do not activate them explicitly. GCC will make use of these
extension now by default.
This makes it possible to share the toolchain with other Cortex A53
SoCs.
Compile tested only.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 10ce015c65)
Some of the Marvell targets have functional SATA port multiplier
support, which is required for multi-bay eSATA enclosures. Enable
kernel support by setting CONFIG_SATA_PMP.
Closes: FS#1232 and FS#547
Signed-off-by: Jeremiah McConnell <miah@miah.com>
(cherry picked from commit 390c4df2c0)
Without this change, ifx_mei_atm_showtime_check() will always return
"showtime" after one call of MEI_InternalXtmSwhowtimeEntrySignal()
was done, even if MEI_InternalXtmSwhowtimeExitSignal() was called
in the meantime.
The ifx_mei_atm_showtime_check() function is used by the ltq-atm and
ltq-ptm driver.
Signed-off-by: Martin Schiller <ms@dev.tdt.de>
(cherry picked from commit eee8ab59dc)
The ART partition of the Lima board stores exactly three mac addresses:
* 0x0: eth0
* 0x6: eth1
* 0x1002: wmac
The first two are correctly assigned in the mach file but the latter points
to 0x800. But this position is set to ff:ff:ff:ff:ff:ff. Luckily, the
driver falls back in ath9k_hw_init_macaddr to the EEPROM mac address when
it doesn't find a valid mac address in the platform_data.
Remove this bogus offset to the ART partition to directly load the wmac via
the EEPROM data in the ART partition.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
(cherry picked from commit 4f6320704f)
ipq806x is all dual core processors. ipq807x is quad core. Removes this
from dmesg:
RCU restricting CPUs from NR_CPUS=4 to nr_cpu_ids=2.
RCU: Adjusting geometry for rcu_fanout_leaf=16, nr_cpu_ids=2
Signed-off-by: Rosen Penev <rosenp@gmail.com>
(cherry picked from commit fff65dbe24)
The mtd tool is built with different configurations depending on the
target. For example, brcm47xx adds the fixtrx subcommand, without which
an image fails when booting the second time.
Mark the mtd package as nonshared to really fix FS#484.
Signed-off-by: Mirko Parthey <mirko.parthey@web.de>
(cherry picked from commit 46d7ced9d1)
Reboot the oxnas target based on Linux 4.14 by rebasing our support on
top of the now-existing upstream kernel support.
This commit brings oxnas support to the level of v4.17 having upstream
drivers for Ethernet, Serial and NAND flash.
Botch up OpenWrt's local drivers for EHCI, SATA and PCIe based on the
new platform code and device-tree.
Re-introduce base-files from old oxnas target which works for now but
needs further clean-up towards generic board support.
Functional issues:
* PCIe won't come up (hence no USB3 on Shuttle KD20)
* I2C bus of Akitio myCloud device is likely not to work (missing
debounce support in new pinctrl driver)
Code-style issues:
* plla/pllb needs further cleanup -- currently their users are writing
into the syscon regmap after acquireling the clk instead of using
defined clk_*_*() functions to setup multipliers and dividors.
* PCIe phy needs its own little driver.
* SATA driver is a monster and should be split into an mfd having
a raidctrl regmap, sata controller, sata ports and sata phy.
Tested on MitraStar STG-212 aka. Medion Akoya MD86xxx and Shuttle KD20.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(squash-picked commit 17511a7ea8 and commit dcc34574ef from master)
TP-Link TL-WR842N v5 are simple N300 router with 5-port FE switch and
non-detachable antennas. Its very similar to TP-Link TL-MR3420 V5.
Specification:
- MT7628N/N (580 MHz)
- 64 MB of RAM (DDR2)
- 8 MB of FLASH
- 2T2R 2.4 GHz
- 5x 10/100 Mbps Ethernet
- 2x external, non-detachable antennas
- USB 2.0 Port
- UART (J1) header on PCB (115200 8n1)
- 7x LED, 2x button, power input switch
Flash instruction:
The only way to flash OpenWrt image in wr842nv5 is to use
tftp recovery mode in U-Boot:
1. Configure PC with static IP 192.168.0.225/24 and tftp server.
2. Rename "lede-ramips-mt7628-tplink_tl-wr842n-v5-squashfs-tftp-recovery.bin"
to "tp_recovery.bin" and place it in tftp server directory.
3. Connect PC with one of LAN ports, press the reset button, power up
the router and keep button pressed for around 6-7 seconds, until
device starts downloading the file.
4. Router will download file from server, write it to flash and reboot.
Signed-off-by: Maxim Anisimov <maxim.anisimov.ua@gmail.com>
With this change, the LED trigger is independent from the (wireless)
netdev name. The (wireless) netdev name can be easiliy changed in
OpenWrt and would require an update of the netdev trigger settings each
time it is done.
This change is (for now) applied only to MT7628 devices from TP-Link, as
we only had the possibility to test this change against two of those
devices, namely a TL-WR841 v13 and a Archer C50 v3.
Signed-off-by: David Bauer <mail@david-bauer.net>
I found mt7688 watchdog not working. The watchdog registers are identical
for mt7621 and mt7628/mt7688. The first watchdog related register is at
0x10000100, the last one - a 16bit sized - at 0x10000128.
Set the correct register address and size in the dtsi file to get the
watchdog working.
Signed-off-by: lbzhung <gewalalb@gmail.com>
[add commit message]
Signed-off-by: Mathias Kresin <dev@kresin.me>
73edb22 mt76: discard early received packets if not running yet
0b8d1dd mt76: fix beacon timer drift
20c0766 mt7603: adjust rx hang watchdog for MT7628
664e321 mt7603: add extra PSE hang check signature for MT7628
f24b56f update MT7628 firmware to the latest version
d87e4b0 mt7603: clear PSE reset bit if PSE reset fails
0ef26ef mt76: only stop tx queues on offchannel, not during the entire scan
f399da3 mt76: prevent tx scheduling during channel change
21c1e1e mt76: move ieee80211_hw allocation to common core
730c292 mt76: wait for pending tx to complete before switching channel
fcbb49e mt76x2: use udelay instead of usleep_range in mt76x2_mac_stop
792dbe0 mt7603: do not hold dev->mutex while flushing dev->mac_work
9090f9c mt76x2: fix threshold for gain adjustment
2cbaa57 mt76x2: fix swapped values for RXO-18 in gain control
a39ab70 mt76x2: adjust AGC control register 26 based on gain for VHT80
4936c0c mt76x2: clear false CCA counters after changing gain settings
1528fe7 mt76x2: fix variable gain adjustment range
f3522e1 mt76x2: add a debugfs file to dump agc calibration information
65e161b mt76x2: fix tracking rssi for dynamic gain adjustment
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This was caused by a race condition between offload teardown and
conntrack gc bumping the timeout of offloaded connections
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Fix condir option processing allowing to use the format
"<directory>[,<file-extension>......]," as documented on the dnsmasq man
page which previously resulted into bogus dir being created.
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
On bcm53xx and brcm47xx, commands are passed to default_do_upgrade that
expect the image to be passed on stdin, rather than as an argument.
Fixes: 30f61a34b4 ("base-files: always use staged sysupgrade")
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Import a revert-commit from Stanislaw Gruszka which significantly
improves WiFi performance on rt2x00 based hardware.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry-picked from commit f4a639a3d7)
This device has only one ethernet port.
Signed-off-by: Matthias Badaire <mbadaire@gmail.com>
[add the existing eth0 as lan block, shorten commit message]
Signed-off-by: Mathias Kresin <dev@kresin.me>
The AVM package selection partially broke with the addition of the
FRITZ!Box 4020. This commit restores the intended behavior.
Signed-off-by: David Bauer <mail@david-bauer.net>
Some of the ubi-tools in the upstream mtd-utils have been
broken by a bad patch upstream. It causes major breakage
during sysupgrade when the kernel, rootfs, ... volumes
are deleted in the wrong order.
This patch therefore reverts the faulty upstream commit which
fixes the bug.
linux-mtd mailing-list thread:
<http://lists.infradead.org/pipermail/linux-mtd/2018-June/081562.html>
Cc: John Crispin <john@phrozen.org>
Reported-by: L. Wayne Leach <LLeachii@aol.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
(cherry picked from commit daf19649db)
Some devices like the Mikrotik RB912 only have 1 USB port
which is shared between an USB A type port, and the mini PCIe socket.
Toggling a gpio selects the output to which USB is connected.
Since kernel 4.9, gpio base is rounded up to a value of 32.
Commit 65da6f9ca1 ("ar71xx: fix secondary gpio controller base values") accounts correctly for that.
In this commit, rb912 sees it's value changed from AR934X_GPIO_COUNT (23) to 32
This means that the USB toggle gpio number actually also changes from 52 to 61.
But ..
Some of these GPIO numbers are also used in other locations, like the boardfile.
The author forgot to also change them over there.
Switching the USB port to mPCIe now shows my modem is correctly discovered again:
[ 2863.864471] usb 1-1: new high-speed USB device number 4 using ehci-platform
[ 2864.055303] usb 1-1: config 1 has an invalid interface number: 8 but max is 3
[ 2864.062728] usb 1-1: config 1 has no interface number 1
[ 2864.074567] qcserial 1-1:1.0: Qualcomm USB modem converter detected
[ 2864.081474] usb 1-1: Qualcomm USB modem converter now attached to ttyUSB0
[ 2864.111960] qcserial 1-1:1.2: Qualcomm USB modem converter detected
[ 2864.118976] usb 1-1: Qualcomm USB modem converter now attached to ttyUSB1
[ 2864.139808] qcserial 1-1:1.3: Qualcomm USB modem converter detected
[ 2864.146777] usb 1-1: Qualcomm USB modem converter now attached to ttyUSB2
[ 2864.165276] qmi_wwan 1-1:1.8: cdc-wdm0: USB WDM device
[ 2864.171879] qmi_wwan 1-1:1.8 wwan0: register 'qmi_wwan' at usb-ehci-platform-1, WWAN/QMI device, 02:00:44:ed:3b:11
Fixes: 65da6f9ca1 ("ar71xx: fix secondary gpio controller base values")
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Cc: Robin Leblon <robin.leblon@ncentric.com>
Cc: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 55b4b1eba0)
This reverts a change made in Sep 2017 [1] which introduced
MSG_DONTROUTE flag to prevent udhcpc from reaching out to servers on a
different subnet. That change violates RFC2131 by forcing fully
configured clients, who got their configurations through an offer
relayed by a DHCP relay, from renewing through a unicast request
directly to the DHCP server, resulting in the client resorting to
boradcasting lease extension requests instead of unicasting them,
further breaking RFC2131.
The problem with MSG_DONTROUTE appears when talking to a properly
configured DHCP server that rejects non-compliant requests. Such server
will reject lease extension attempts sent via broadcast rather than
unicast, as is the case with Finnish ISPs Telia and DNA as well as
Estonian ISP Starman. Once the lease expires without renewal, udhcpc
enters init mode, taking down the interfaces with it, and thus causing
interruption on every lease expiry. On some ISPs (such as the ones
mentioned above) that can be once every 10-20 minutes. The interruptions
appear in the logs as such:
----
udhcpc: sending renew to x.x.x.x
udhcpc: send: Network unreachable
udhcpc: sending renew to 0.0.0.0
udhcpc: sending renew to 0.0.0.0
...
udhcpc: lease lost, entering init state
Interface 'wan' has lost the connection
Interface 'wan' is now down
Network alias 'eth0' link is down
udhcpc: sending select for y.y.y.y
udhcpc: lease of y.y.y.y obtained, lease time 1200
Network alias 'eth0' link is up
Interface 'wan' is now up
----
During lease extension, a fully configured client should be able to
reach out to the server from which it recieved the lease for extension,
regardless in which network it is; that's up to the gateway to find. [2]
This patch ensures that.
[1]
http://lists.busybox.net/pipermail/busybox-cvs/2017-September/037402.html
[2]
https://www.netmanias.com/en/post/techdocs/6000/dhcp-network-protocol/
understanding-dhcp-relay-agents
Signed-off-by: Adi Shammout <adi.shammout@outlook.com>
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com> [PKG_RELEASE increase]
If log_file is specified, make sure its directory exists.
Signed-off-by: Karl Palsson <karlp@etactica.com>
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com> [PKG_RELEASE increase]
The kernel bump wrongly modified the patch
generic/pending-4.14/811-pci_disable_usb_common_quirks.patch.
Sync it from master.
Fixes: 1199a91095 ("kernel: bump 4.14 to 4.14.48 for 18.06")
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
Refreshed patches. The following patches were upstreamed and have been deleted:
* target/linux/lantiq/patches-4.14/0025-MIPS-lantiq-gphy-Remove-reboot-remove-reset-asserts.patch
* target/linux/generic/pending-4.14/101-clocksource-mips-gic-timer-fix-clocksource-counter-w.patch
* target/linux/generic/pending-4.14/103-MIPS-c-r4k-fix-data-corruption-related-to-cache-coherence.patch
* target/linux/generic/pending-4.14/181-net-usb-add-lte-modem-wistron-neweb-d18q1.patch
Compile-tested: ramips/mt7621, x86/64
Run-tested: ramips/mt7621
Signed-off-by: Stijn Segers <foss@volatilesystems.org>
Refreshed patches. The following patches were upstreamed and have been deleted:
* target/linux/ar71xx/patches-4.9/106-01-MIPS-ath79-fix-AR724X_PLL_REG_PCIE_CONFIG-offset.patch
* target/linux/generic/pending-4.9/180-net-phy-at803x-add-support-for-AT8032.patch
* target/linux/generic/pending-4.9/181-net-usb-add-lte-modem-wistron-neweb-d18q1.patch
* target/linux/generic/pending-4.9/182-net-qmi_wwan-add-BroadMobi-BM806U-2020-2033.patch
Signed-off-by: Stijn Segers <foss@volatilesystems.org>
We recently increased the kernel partition size of the CPE/WBS 210/510.
This works fine for new installations of the factory image, but on
sysupgrades, the partition table read by the bootloader is not adjusted.
This limits the maximum size of the kernel loaded by the bootloader to the
old partition size.
While adjusting the partition table would be a cleanest solution, such a
migration would have to happen before an upgrade to a new version with a
newer kernel. This is error-prone and would require a two-step upgrade, as
we mark the partition table partition read-only.
Instead, switch from the lzma-loader with embedded kernel to the
okli-loader, so only the tiny lzma-loader is loaded by the bootloader as
"kernel", and the lzma-loader will then load the rest of the kernel by
itself.
Fixes: e39847ea2f ("ar71xx: increase kernel partition size for CPE/WBS 210/510")
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
By making the kernel argv array const, the .data section can always be
omitted from the laoder binary.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
