Zoltan HERPAI [Mon, 18 Jan 2021 08:00:42 +0000 (09:00 +0100)]
sunxi: add support for Banana Pi M2 Berry
CPU: Allwinner V40 quad-core Cortex A7 @ 1.2GHz
Memory: 1GB DDR3
Storage: SDcard, native SATA
Network: 10/100/1000M ethernet, Ampak AP6212 wifi + BT
USB: 4x USB 2.0
Installation:
Use the standard sunxi installation to an SD-card.
While the board is very similar to the M2 Ultra board
(the V40 is the automotive version of the R40), as both
the u-boot and kernel supports them separately, and some
pins are different, let's add a separate device spec.
Andrea Poletti [Fri, 23 Jul 2021 16:15:00 +0000 (18:15 +0200)]
ramips: add support for Sitecom WLR-4100 v1 002
Sitecom WLR-4100 v1 002 (marked as X4 N300) is a wireless router
Specification:
SoC: MT7620A
RAM: 64 MB DDR2
Flash: MX25L6405D SPI NOR 8 MB
WIFI: 2.4 GHz integrated
Ethernet: 5x 10/100/1000 Mbps QCA8337
USB: 1x 2.0
LEDS: 2x GPIO controlled, 5x switch
Buttons: 1x GPIO controlled
UART: row of 4 unpopulated holes near USB port, starting count from
white triangle on PCB:
VCC 3.3V
GND
TX
RX
baud: 115200, parity: none, flow control: none
Installation
Connect to one of LAN (yellow) ethernet ports,
Open router configuration interface,
Go to Toolbox > Firmware,
Browse for OpenWrt factory image with dlf extension and hit Apply,
Wait few minutes, after the Power LED will stop blinking, the router is
ready for configuration.
Known issues
Some USB 2.0 devices work at full speed mode 1.1 only
MAC addresses
factory partition only contains one (binary) MAC address in 0x4.
u-boot-env contains four (ascii) MAC addresses, of which two appear
to be valid.
The factory firmware only assigns ethaddr. Thus, we take the
binary value which we can use directly in DTS.
Additional information
OEM firmware shell password is: SitecomSenao
useful for creating backup of original firmware.
There is also another revision of this device (v1 001), based on RT3352 SoC
The nvmem feature (commit 06bb4a5) was introduced in master after the
splitting of the 21.02 branch. It need to be reverted in 21.02..
Signed-off-by: Andrea Poletti <polex73@yahoo.it>
[remove config DT label, convert to nvmem, remove MAC address
setup from u-boot-env, add MAC address info to commit message] Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
(cherry picked from commit de0c380a5f8289839ab970e794a45f0e04a466a3)
Currently, uClic++ 0.2.5 fails to compile when using a long filepath.
For example, if the openwrt directory is in the path:
/tmp/this_directory_name_is_very_long/more_long_paths/.../openwrt,
then uclibc++ will cause a very obtuse error.
Although the uclibc++ makefiles do print a "File name too long" error,
it's not the final error that's printed, so it's a bit confusing:
> /bin/sh: 1:
> cannot create src/abi/libsupc/<SNIP>_libsupc++.a.dep: File name too long
> <SNIP: some other makefile output here>
> array_type_info.o: No such file or directory
Although OpenWRT 22.03 and current master branch have removed uClib++,
I thought I'd make a PR for OpenWRT 21.02, since I encountered it
and there seems to be quite a few other people experiencing the same issue.
It especially happens when using the SDK, (or when using an encrypted fs)
since the pre-packaged SDKs have very long filenames.
This patch is already in upstream [1], but has not yet been released.
Petr Štetiar [Tue, 9 Aug 2022 05:50:19 +0000 (07:50 +0200)]
zlib: backport null dereference fix
The curl developers found test case that crashed in their testing when
using zlib patched against CVE-2022-37434, same patch we've backported
in commit 7df6795d4c25 ("zlib: backport fix for heap-based buffer
over-read (CVE-2022-37434)"). So we need to backport following patch in
order to fix issue introduced in that previous CVE-2022-37434 fix.
References: https://github.com/curl/curl/issues/9271 Fixes: 7df6795d4c25 ("zlib: backport fix for heap-based buffer over-read (CVE-2022-37434)") Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit f443e9de7003c00a935b9ea12f168e09e83b48cd)
(cherry picked from commit 707ec48ab3db6d08bd022df1bc720aee68b3b99d)
Petr Štetiar [Sat, 6 Aug 2022 12:55:07 +0000 (14:55 +0200)]
zlib: backport fix for heap-based buffer over-read (CVE-2022-37434)
zlib through 1.2.12 has a heap-based buffer over-read or buffer overflow
in inflate in inflate.c via a large gzip header extra field. NOTE: only
applications that call inflateGetHeader are affected. Some common
applications bundle the affected zlib source code but may be unable to
call inflateGetHeader.
Fixes: CVE-2022-37434
References: https://github.com/ivd38/zlib_overflow Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 7df6795d4c25447683fd4b4a4813bebcddaea547)
Paul Spooren [Tue, 7 Jun 2022 11:26:43 +0000 (13:26 +0200)]
octeon: add SUPPORTED_DEVICES to er/erlite
Using the BOARD_NAME variable results for both er and erlite devices to
identify themselfs as `er` and `erlite` (via `ubus call system board`).
This is problematic when devices search for firmware upgrades since the
OpenWrt profile is actually called `ubnt_edgerouter` and
`ubnt_edgerouter-lite`.
By adding the `SUPPORTED_DEVICE` a mapping is created to point devices
called `er` or `erlite` to the corresponding profile.
Debian's changelog by Henrique de Moraes Holschuh <hmh@debian.org>:
* upstream changelog: new upstream datafile 20220207
* Mitigates (*only* when loaded from UEFI firmware through the FIT)
CVE-2021-0146, INTEL-SA-00528: VT-d privilege escalation through
debug port, on Pentium, Celeron and Atom processors with signatures
0x506c9, 0x506ca, 0x506f1, 0x706a1, 0x706a8
https://github.com/intel/Intel-Linux-Processor-Microcode-Data-Files/issues/57#issuecomment-1036363145
* Mitigates CVE-2021-0127, INTEL-SA-00532: an unexpected code breakpoint
may cause a system hang, on many processors.
* Mitigates CVE-2021-0145, INTEL-SA-00561: information disclosure due
to improper sanitization of shared resources (fast-store forward
predictor), on many processors.
* Mitigates CVE-2021-33120, INTEL-SA-00589: out-of-bounds read on some
Atom Processors may allow information disclosure or denial of service
via network access.
* Fixes critical errata (functional issues) on many processors
* Adds a MSR switch to enable RAPL filtering (default off, once enabled
it can only be disabled by poweroff or reboot). Useful to protect
SGX and other threads from side-channel info leak. Improves the
mitigation for CVE-2020-8694, CVE-2020-8695, INTEL-SA-00389 on many
processors.
* Disables TSX in more processor models.
* Fixes issue with WBINDV on multi-socket (server) systems which could
cause resets and unpredictable system behavior.
* Adds a MSR switch to 10th and 11th-gen (Ice Lake, Tiger Lake, Rocket
Lake) processors, to control a fix for (hopefully rare) unpredictable
processor behavior when HyperThreading is enabled. This MSR switch
is enabled by default on *server* processors. On other processors,
it needs to be explicitly enabled by an updated UEFI/BIOS (with added
configuration logic). An updated operating system kernel might also
be able to enable it. When enabled, this fix can impact performance.
* Updated Microcodes:
sig 0x000306f2, pf_mask 0x6f, 2021-08-11, rev 0x0049, size 38912
sig 0x000306f4, pf_mask 0x80, 2021-05-24, rev 0x001a, size 23552
sig 0x000406e3, pf_mask 0xc0, 2021-04-28, rev 0x00ec, size 105472
sig 0x00050653, pf_mask 0x97, 2021-05-26, rev 0x100015c, size 34816
sig 0x00050654, pf_mask 0xb7, 2021-06-16, rev 0x2006c0a, size 43008
sig 0x00050656, pf_mask 0xbf, 2021-08-13, rev 0x400320a, size 35840
sig 0x00050657, pf_mask 0xbf, 2021-08-13, rev 0x500320a, size 36864
sig 0x0005065b, pf_mask 0xbf, 2021-06-04, rev 0x7002402, size 28672
sig 0x00050663, pf_mask 0x10, 2021-06-12, rev 0x700001c, size 28672
sig 0x00050664, pf_mask 0x10, 2021-06-12, rev 0xf00001a, size 27648
sig 0x00050665, pf_mask 0x10, 2021-09-18, rev 0xe000014, size 23552
sig 0x000506c9, pf_mask 0x03, 2021-05-10, rev 0x0046, size 17408
sig 0x000506ca, pf_mask 0x03, 2021-05-10, rev 0x0024, size 16384
sig 0x000506e3, pf_mask 0x36, 2021-04-29, rev 0x00ec, size 108544
sig 0x000506f1, pf_mask 0x01, 2021-05-10, rev 0x0036, size 11264
sig 0x000606a6, pf_mask 0x87, 2021-12-03, rev 0xd000331, size 291840
sig 0x000706a1, pf_mask 0x01, 2021-05-10, rev 0x0038, size 74752
sig 0x000706a8, pf_mask 0x01, 2021-05-10, rev 0x001c, size 75776
sig 0x000706e5, pf_mask 0x80, 2021-05-26, rev 0x00a8, size 110592
sig 0x000806a1, pf_mask 0x10, 2021-09-02, rev 0x002d, size 34816
sig 0x000806c1, pf_mask 0x80, 2021-08-06, rev 0x009a, size 109568
sig 0x000806c2, pf_mask 0xc2, 2021-07-16, rev 0x0022, size 96256
sig 0x000806d1, pf_mask 0xc2, 2021-07-16, rev 0x003c, size 101376
sig 0x000806e9, pf_mask 0x10, 2021-04-28, rev 0x00ec, size 104448
sig 0x000806e9, pf_mask 0xc0, 2021-04-28, rev 0x00ec, size 104448
sig 0x000806ea, pf_mask 0xc0, 2021-04-28, rev 0x00ec, size 103424
sig 0x000806eb, pf_mask 0xd0, 2021-04-28, rev 0x00ec, size 104448
sig 0x000806ec, pf_mask 0x94, 2021-04-28, rev 0x00ec, size 104448
sig 0x00090661, pf_mask 0x01, 2021-09-21, rev 0x0015, size 20480
sig 0x000906c0, pf_mask 0x01, 2021-08-09, rev 0x2400001f, size 20480
sig 0x000906e9, pf_mask 0x2a, 2021-04-29, rev 0x00ec, size 106496
sig 0x000906ea, pf_mask 0x22, 2021-04-28, rev 0x00ec, size 102400
sig 0x000906eb, pf_mask 0x02, 2021-04-28, rev 0x00ec, size 104448
sig 0x000906ec, pf_mask 0x22, 2021-04-28, rev 0x00ec, size 103424
sig 0x000906ed, pf_mask 0x22, 2021-04-28, rev 0x00ec, size 103424
sig 0x000a0652, pf_mask 0x20, 2021-04-28, rev 0x00ec, size 93184
sig 0x000a0653, pf_mask 0x22, 2021-04-28, rev 0x00ec, size 94208
sig 0x000a0655, pf_mask 0x22, 2021-04-28, rev 0x00ee, size 94208
sig 0x000a0660, pf_mask 0x80, 2021-04-28, rev 0x00ea, size 94208
sig 0x000a0661, pf_mask 0x80, 2021-04-29, rev 0x00ec, size 93184
sig 0x000a0671, pf_mask 0x02, 2021-08-29, rev 0x0050, size 102400
* Removed Microcodes:
sig 0x00080664, pf_mask 0x01, 2021-02-17, rev 0xb00000f, size 130048
sig 0x00080665, pf_mask 0x01, 2021-02-17, rev 0xb00000f, size 130048
* update .gitignore and debian/.gitignore.
Add some missing items from .gitignore and debian/.gitignore.
* ucode-blacklist: do not late-load 0x406e3 and 0x506e3.
When the BIOS microcode is older than revision 0x7f (and perhaps in some
other cases as well), the latest microcode updates for 0x406e3 and
0x506e3 must be applied using the early update method. Otherwise, the
system might hang. Also: there must not be any other intermediate
microcode update attempts [other than the one done by the BIOS itself],
either. It must go from the BIOS microcode update directly to the
latest microcode update.
* source: update symlinks to reflect id of the latest release, 20220207
sdk: add spidev-test to the bundle of userspace sources
moves and extends the current facilities, which have been
added some time ago for the the usbip utility, to support
more utilites that are shipped with the Linux kernel tree
to the SDK.
this allows to drop all the hand-waving and code for
failed previous attempts to mitigate the SDK build failures.
*) AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
implementation would not encrypt the entirety of the data under some
circumstances. This could reveal sixteen bytes of data that was
preexisting in the memory that wasn't written. In the special case of
"in place" encryption, sixteen bytes of the plaintext would be revealed.
Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
they are both unaffected.
(CVE-2022-2097)
[Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño]
Chukun Pan [Fri, 10 Jun 2022 15:54:40 +0000 (23:54 +0800)]
mediatek: mt7623: fixes kconfig for hwcrypto
The MediaTek's Crypto Engine driver has been replaced with the upstream
Inside Secure's SafeXcel cryptographic engine driver, however kconfig
has not been changed accordingly, this commit fixes it.
Fixes: 127ad76 ("mediatek: switch over to extended upstream eip97
driver")
Andre Heider [Thu, 23 Jun 2022 07:08:07 +0000 (09:08 +0200)]
openssl: bump to 1.1.1p
Changes between 1.1.1o and 1.1.1p [21 Jun 2022]
*) In addition to the c_rehash shell command injection identified in
CVE-2022-1292, further bugs where the c_rehash script does not
properly sanitise shell metacharacters to prevent command injection have been
fixed.
When the CVE-2022-1292 was fixed it was not discovered that there
are other places in the script where the file names of certificates
being hashed were possibly passed to a command executed through the shell.
This script is distributed by some operating systems in a manner where
it is automatically executed. On such operating systems, an attacker
could execute arbitrary commands with the privileges of the script.
Use of the c_rehash script is considered obsolete and should be replaced
by the OpenSSL rehash command line tool.
(CVE-2022-2068)
[Daniel Fiala, Tomáš Mráz]
*) When OpenSSL TLS client is connecting without any supported elliptic
curves and TLS-1.3 protocol is disabled the connection will no longer fail
if a ciphersuite that does not use a key exchange based on elliptic
curves can be negotiated.
[Tomáš Mráz]
bcm53xx: use -falign-functions=32 for kernel compilation
Northstar SoCs have pretty small CPU caches and their performance is
heavily affected by cache hits & misses. It means that all kind of
random code changes can affect performance as they often reorganize
(change alignment & possibly reorder) kernel symbols.
It was discussed in ARM / net mailinglists:
1. ARM router NAT performance affected by random/unrelated commits [1] [2]
2. Optimizing kernel compilation / alignments for network performance [3] [4]
It seems that -falign-functions can be used as a partial workaround. It
doesn't solve all cases (e.g. documented watchdog one [5]) but it surely
helps with many of them.
A complete long term solution may be PGO (profile-guided optimization)
but it isn't available at this point.
Rafał Miłecki [Fri, 10 Jun 2022 08:51:23 +0000 (10:51 +0200)]
bcm53xx: enable & setup packet steering
Packet steering can improve NAT masquarade performance on Northstar by
40-50%. It makes reaching 940-942 Mb/s possible on BCM4708 (and
obviously BCM47094 too). Add scripts setting up the most optimal
Northstar setup.
Below are testing results for running iperf TCP traffic from LAN to WAN.
They were used to pick up golden values.
Above tests were performed with all eth0 interrupts handled by CPU0.
Setting "echo 2 > /proc/irq/38/smp_affinity" was tested on BCM4708 but
it didn't increased speeds (just required different steering):
Rafał Miłecki [Mon, 20 Jun 2022 08:21:20 +0000 (10:21 +0200)]
bcm53xx: disable GRO by default at kernel level
This improves NAT masquarade network performance.
An alternative to kernel change would be runtime setup but that requires
ethtool and identifying relevant network interface and all related
switch ports interfaces.
Rafał Miłecki [Wed, 15 Jun 2022 10:36:37 +0000 (12:36 +0200)]
kernel: drop patch adding hardcoded kernel compilation flags
1. KCFLAGS should be used for custom flags
2. Optimization flags are arch / SoC specific
3. -fno-reorder-blocks may *worsen* network performace on some SoCs
4. Usage of flags was *reversed* since 5.4 and noone reported that
If we really need custom flags then CONFIG_KERNEL_CFLAGS should get
default value adjusted properly (per target).
This patch was backported to kernel 5.4.200, but without the wireguard
change, because wireguard is not available in upstream kernel 5.4.
This adds the missing changes for wireguard too.
mvebu: move upstreamed DTS files (ESPRESSObin) to files-5.4
Since kernel 5.5-rc1 [1], there are upstreamed DTS files related to ESPRESSObin
variants. Move these to files-5.4.
This helps if you want to use a newer kernel version than used
in OpenWrt 21.02 (= LTS kernel 5.4), you would end up with duplicate files
(one outdated, one up to date from newer Linux versions).
Fixes:
Error: arch/arm64/boot/dts/marvell/armada-3720-espressobin-v7.dts:19.1-7:
Label or path ports not found
FATAL ERROR: Syntax error parsing input tree
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
(cherry picked from commit 7be8ab4f7b582924bca6594103735d888989d804) Signed-off-by: Josef Schlehofer <pepe.schlehofer@gmail.com>
[reword commit subject and commit description]
Chukun Pan [Wed, 8 Jun 2022 15:05:03 +0000 (23:05 +0800)]
exfat: update to 5.19.1
Major changes are:
4 cleanups & typos fixes.
Add keep_last_dots mount option to allow access to paths
with trailing dots.
Avoid repetitive volume dirty bit set/clear to improve
storage life time.
Fix ->i_blocks truncation issue caused by wrong 32bit mask.
Fix ->i_blocks truncation issue that still exists elsewhere.
Fix missing REQ_SYNC in exfat_update_bhs().
Fix referencing wrong parent directory information during rename.
Fix slab-out-bounds in exat_clear_bitmap() reported from syzbot.
Improve performance while zeroing a cluster with dirsync mount option.
Introduce a sys_tz mount option to use system timezone.
Move super block magic number to magic.h
Tiago Gaspar [Wed, 4 May 2022 09:36:07 +0000 (10:36 +0100)]
firewall: config: remove restictions on DHCPv6 allow rule
Remove restrictions on source and destination addresses, which aren't
specified on RFC8415, and for some reason in openwrt are configured
to allow both link-local and ULA addresses.
As cleared out in issue #5066 there are some ISPs that use Gloabal
Unicast addresses, so fix this rule to allow them.
Alban Bedel [Sat, 30 Apr 2022 08:42:33 +0000 (10:42 +0200)]
ramips: zbt-wg2626: Add the reset gpio for PCIe port 1
The 2.4GHz interface doesn't come up properly with the log showing:
mt7621-pci 1e140000.pcie: pcie1 no card, disable it (RST & CLK)
As seen on other MT7621 boards this is caused by a missing reset GPIO.
The MT7621 dtsi set GPIO 19 as PCIe reset GPIO, which on this board
reset the 5GHz interface on port 0. Add GPIO 8 to the PCIe reset GPIO
list to also reset the 2.4GHz interface on port 1.
Nick Hainke [Mon, 18 Apr 2022 13:04:25 +0000 (15:04 +0200)]
ipq40xx: fix ar40xx driver
This commit is completely based on the work of adron-s:
https://github.com/openwrt/openwrt/pull/4721#issuecomment-1101108651
The commit fixes the data corruption on TX packets. Packets are
transmitted, but their contents are replaced with zeros. This error is
caused by the lack of guard (50 ms) intervals between calibration phases.
This error is treated by adding mdelay(50) to the calibration function
code. In the original qca-ssda code [0], these mdelays were existing, but
in the ar41xx.c they are gone.
Suggested-by: Serhii Serhieiev <adron@mstnt.com> Reviewed-by: Robert Marko <robimarko@gmail.com> Signed-off-by: Nick Hainke <vincent@systemli.org>
(cherry picked from commit ab7e53e5cce703c7a62efbe1d41fb94c2228a178)
[Deleted 5.10 from commit title] Signed-off-by: Nick Hainke <vincent@systemli.org>
David Bauer [Tue, 29 Mar 2022 22:31:26 +0000 (00:31 +0200)]
hostapd: add ubus method for requesting link measurements
Add a ubus method to request link-measurements from connected STAs.
In addition to the STAs address, the used and maximum transmit power can
be provided by the external process for the link-measurement. If they
are not provided, 0 is used as the default value.
Felix Fietkau [Thu, 10 Jun 2021 06:52:28 +0000 (08:52 +0200)]
iwinfo: update to the latest version
aa0e3c4bbe12 iwinfo: nl80211: add support for printing the device path for a phy dd6d6d2dec35 iwinfo: nl80211: use new path lookup function for nl80211_phy_idx_from_uci_path 268bb26d2e2a iwinfo: nl80211: support looking up phy by path=.. and macaddr=... c0414642fead iwinfo: nl80211: fix typo
Felix Fietkau [Tue, 19 Apr 2022 19:14:42 +0000 (21:14 +0200)]
kernel: add missing config symbols
MPLS feature symbols are normally only set when kmod-mpls is enabled, but the
CONFIG_MPLS symbol they depend on could also have been selected by openvswitch
instead
This patch fixes an invalid TX PA DC bias level on QCA9561, which
results in a very low output power and very low throughput as devices
are further away from the AP (compared to other 2.4GHz APs),
following a suggestion from nbd[1].
Michael Pratt [Tue, 20 Apr 2021 16:35:47 +0000 (12:35 -0400)]
ramips: mt7620: disable SOC VLANs for external switches
These boards have AR8327 or QCA8337 external ethernet switch.
The SOC also has it's own internal switch
where VLAN is now enabled by default.
Changes to preinit caused all switches to have VLANs enabled by default
even if they are not configured with a topology in uci_defaults
(see commit f017f617aecbd47debd4d3a734dc0e471342db96)
When both internal and external switches have VLANs,
and the external switch has both LAN and WAN,
the TX traffic from the SOC cannot flow to the tagged port on the external switch
because the VLAN IDs are not matching.
So disable the internal switch VLANs by default on these boards.
Also, add a topology for the internal switch,
so that on LuCI there is not an "unknown topology" warning.
In theory, it may be possible to have LAN ports on both switches
through internal and external PHYs, but there are no known boards that have this.
Michael Pratt [Mon, 31 May 2021 07:56:20 +0000 (03:56 -0400)]
ramips: mt7620: add ephy-disable option to switch driver
Add back the register write to disable internal PHYs
as a separate option in the code that can be set using a DTS property.
Set the option to true by default
when an external mt7530 switch is identified.
This makes the driver more in sync with original SDK code
while keeping the lines separated into different options
to accommodate any board with any PHY layout.
Michael Pratt [Sun, 30 May 2021 06:02:47 +0000 (02:02 -0400)]
ramips: mt7620: move mt7620_mdio_mode() to ethernet driver
The function mt7620_mdio_mode is only called once
and both the function and mdio_mode block have been named incorrectly,
leading to confusion and useless commits.
These lines in the mdio_mode block of mt7620_hw_init
are only intended for boards with an external mt7530 switch.
(see commit 194ca6127ee18cd3a95da4d03f02e43b5428c0bb)
Therefore, move lines from mdio_mode to the place in soc_mt7620.c
where the type of mt7530 switch is identified,
and move lines from mt7620_mdio_mode to a main function.
mt7620_mdio_mode was called from mt7620_gsw_init
where the priv struct is available,
so the lines must stay in mt7620_gsw_init function.
In order to keep things as simple as possible,
keep the DTS property related function calls together,
by moving them from mt7620_gsw_probe to init.
Remove the now useless DTS properties and extra phy nodes.
Fixes: 5a6229a93df8 ("ramips: remove superfluous & confusing DT binding") Fixes: b85fe43ec8c4 ("ramips: mt7620: add force use of mdio-mode") Signed-off-by: Michael Pratt <mcpratt@pm.me>
(cherry picked from commit 6972e498d33ec896c676b7af91e3bfb00aa846a1)
Michael Pratt [Sat, 3 Apr 2021 18:42:51 +0000 (14:42 -0400)]
ramips: mt7620: use DTS to set PHY base address for external PHYs
Set the PHY base address to 12 for mt7530 and 8 for others,
which is based on the default setting for some devices
from printing the register with the following command
after it is written to by uboot during the boot cycle.
`md 0x10117014 1`
PHY_BASE option only uses 5 bits of the register,
bits 16 to 20, so use 8-bit integer type.
Set the option using the DTS property mediatek,ephy-base
and create the gsw node if missing.
Also, added a kernel message to display the EPHY base address.
Note:
If anything is written to a PHY address that is greater than 1 hex char (greater than 0xf)
then there is adverse effects with Atheros switches.
Michael Pratt [Tue, 20 Apr 2021 22:45:15 +0000 (18:45 -0400)]
ramips: mt7620: allow both internal and external PHYs
When the new variable ephy_base was introduced,
it was not applied to the if block for mdio_mode.
The first line in the mdio_mode if block
sets the EPHY base address to 12 in the SOC by writing a register,
but the corresponding variable in the driver
was still set to the default of 0.
This causes subsequent lines that write registers with the function
_mt7620_mii_write
to write to PHY addresses 0 through 4
while internal PHYs have been moved to addresses 12 through 16.
All of these lines are intended only for PHYs on the SOC internal switch,
however, they are being written to external ethernet switches
if they exist at those PHY addresses 0 through 4.
This causes some ethernet ports to be broken on boards with AR8327 or QCA8337 switch.
Other suggested fixes move those lines to the else block of mdio_mode,
but removing the else block completely also fixes it.
Therefore, move the lines to the mt7620_hw_init function main block,
and have only one instance of the function mtk_switch_w32
for writing the register with the EPHY base address.
In theory, this also allows for boards that have both external switches
and internal PHYs that lead to ethernet ports to be supported.
Fixes: 391df3782914 ("ramips: mt7620: add EPHY base mdio address changing possibility") Signed-off-by: Michael Pratt <mcpratt@pm.me>
(cherry picked from commit de5394a29dae9356a830d043e76591698411e97a)
Michael Pratt [Wed, 2 Jun 2021 03:01:26 +0000 (23:01 -0400)]
ramips: mt7620: fix ethernet driver GMAC port init
A workaround was added to the switch driver
to set SOC port 4 as an RGMII GMAC interface
based on the DTS property mediatek,port4-gmac.
(previously mediatek,port4)
However, the ethernet driver already does this,
but is being blocked by a return statement
whenever the phy-handle and fixed-link properties
are both missing from nodes that define the port properties.
Revert the workaround, so that both the switch driver
and ethernet driver are not doing the same thing
and move the phy-handle related lines down
so nothing is ending the function prematurely.
While at it, clean up kernel messages
and delete useless return statements.
Fixes: f6d81e2fa1f1 ("mt7620: gsw: make IntPHY and ExtPHY share mdio addr 4 possible") Signed-off-by: Michael Pratt <mcpratt@pm.me>
(cherry picked from commit afd60d650e769e9578eac5bb3647807f683aaf7c)
Michael Pratt [Fri, 11 Jun 2021 18:54:43 +0000 (14:54 -0400)]
ramips: mt7620: remove useless GMAC nodes
These nodes are used for configuring a GMAC interface
and for defining external PHYs to be accessed with MDIO.
None of this is possible on MT7620N, only MT7620A,
so remove them from all MT7620N DTS.
When the mdio-bus node is missing, the driver returns -NODEV
which causes the internal switch to not initialize.
Replace that return so that everything works without the DTS node.
Also, an extra kernel message to indicate for all error conditions
that mdio-bus is disabled.
Fixes: d482356322c9 ("ramips: mt7620n: add mdio node and disable port4 by default") Fixes: aa5014dd1a58 ("ramips: mt7620n: enable port 4 as EPHY by default") Signed-off-by: Michael Pratt <mcpratt@pm.me>
(cherry picked from commit a2acdf9607045e5669c305c57dd7c77be8351ba0)
Let the DTS property be boolean instead of string
where EPHY mode is the default.
Fix how the properties are written
for all DTS that use them,
and add missing nodes where applicable,
and remove useless nodes,
and minor DTS formatting.
Ben Gainey [Sun, 31 Jan 2021 11:15:38 +0000 (11:15 +0000)]
ramips: add support for the Wavlink WL-WN579X3
About the device
----------------
SoC: MediaTek MT7620a @ 580MHz
RAM: 64M
FLASH: 8MB
WiFi: SoC-integrated: MediaTek MT7620a bgn
WiFi: MediaTek MT7612EN nac
GbE: 2x (RTL8211F)
BTN: - WPS
- Reset
- Router/Repeater/AP (3-way slide-switch)
LED: - WPS (blue)
- 3-segment Wifi signal representation (blue)
- WiFi (blue)
- WAN (blue)
- LAN (blue)
- Power (blue)
UART: UART is present as Pads with through-holes on the PCB. They are
located next to the reset button and are labelled Vcc/TX/RX/GND as
appropriate. Use 3.3V, 57600-8N1.
Installation
------------
Using the webcmd interface
--------------------------
Warning: Do not update to the latest Wavlink firmware (version 20201201) as this removes the webcmd console and you will need to
use the serial port instead.
You will need to have built uboot/sqauashfs image for this device,
and you will need to provide an HTTP service where the image can
be downloaded from that is accessible by the device.
You cannot use the device manufacturers firmware upgrade interface
as it rejects the OpenWrt image.
1. Log into the device's admin portal. This is necessary to
authenticate you as a user in order to be able to access the
webcmd interface.
2. Navigate to http://<device-ip>/webcmd.shtml - you can access
the console directly through this page, or you may wish to
launch the installed `telnetd` and use telnet instead.
* Using telnet is recommended since it provides a more
convenient shell interface that the web form.
* Launch telnetd from the form with the command `telnetd`.
* Check the port that telnetd is running on using
`netstat -antp|grep telnetd`, it is likely to be 2323.
* Connect to the target using `telnet`. The username should
be `admin2860`, and the password is your admin password.
3. On the target use `curl` to download the image.
e.g. `curl -L -O http://<some-other-lan-ip>/openwrt-ramips-mt7620-\
wavlink_wl-wn579x3-squashfs-sysupgrade.bin`.
Check the hash using `md5sum`.
4. Use the mtd_write command to flash the image.
* The flash partition should be mtd4, but check
/sys/class/mtd/mtd4/name first. The partition should be
called 'Kernel'.
* To flash use the following command:
`mtd_write -r -e /dev/mtd<n> write <image-file> /dev/mtd<n>`
Where mtd<n> is the Kernel partition, and <image-file> is
the OpenWrt image previously downloaded.
* The command above will erase, flash and then reboot the
device. Once it reboots it will be running OpenWrt.
Connect via ssh to the device at 192.168.1.1 on the LAN port.
The WAN port will be configured via DHCP.
Using the serial port
---------------------
The device uses uboot like many other MT7260a based boards. To
use this interface, you will need to connect to the serial
interface, and provide a TFTP server. At boot follow the
bootloader menu and select option 2 to erase/flash the image.
Provide the address and filename details for the tftp server.
The bootloader will do the rest.
Once the image is flashed, the board will boot into OpenWrt. The
console is available over the serial port.
Descriptions:
Previous supported device YOUKU yk1 is actually Youku YK-L1. Though they look
really different, the only hardware difference between the two models is flash
size, YK-L1 has 32 MiB flash but YK-L1c has 16MiB. It seems that YK-L1c can
compatible with YK-L1's firmware but it's better to split it to different models.
It is easy to identify the models by looking at the label on the bottom of the
device. The label has the model number "YK-L1" or "YK-L1c". Due to different flash
sizes, YK-L1c that using previous YK-L1's firmware needs to apply "force update"
to install compatible firmware, so please backup config file before system upgrade.
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
[use more specific name for DTSI] Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
(cherry picked from commit 4a9f389ed2dcee18a5c5e1f0d4e5c406f9290579)
Shiji Yang [Sun, 4 Apr 2021 03:55:41 +0000 (11:55 +0800)]
ramips: improve pinctrl for Youku YK-L1
1. rename led pin "air" to a more common name "wlan" and use "phy0tpt" to trigger it.
2. led "wan" can be triggered by ethernet pinctrl by default so just drop it.
Shiji Yang [Sun, 21 Mar 2021 14:43:59 +0000 (22:43 +0800)]
ramips: speed up spi frequency for Youku YK-L1
Youku YK-L1 has a huge storage space up to 32 MB. It is better to
use a higher spi clock to read or write serial nor flash chips.
Youku YK-L1 has Winbond w25q256fvfg on board that can support
104 MHz spi clock so 48 MHz is safe enough.
The real frequency can only be sysclk(580MHz ) /3 /(2^n) so 80 MHz
defined in dts file will set only 48 MHz in spi bus.
Josef Schlehofer [Thu, 18 Nov 2021 15:37:17 +0000 (16:37 +0100)]
mvebu: kernel: enable CONFIG_BLK_DEV_NVME
CONFIG_BLK_DEV_NVME [1]:
- This is a kernel driver for SSD connected to PCI or PCIe bus [1].
By default, it is enabled for targets "ipq807x", "rockchip/armv8"
and "x86/64".
With miniPCIe adapter, there is a possibility to connect NVMe disk
to Turris Omnia (cortex-a9), Turris MOX (cortex-a53).
It allows to boot system from NVMe disk, because of that it can not
be kmod package as you can not access the disk to be able to boot from
it.
CONFIG_NVME_CORE [2]:
- This is selected by CONFIG_BLK_DEV_NVME
It does not need to be explicitly enabled, but it is done for "ipq807",
"x64_64" and rockchip/armv8", which has also enabled the previous config
option as well.
Joe Mullally [Mon, 30 Aug 2021 21:35:05 +0000 (22:35 +0100)]
ath79: Move TPLink WPA8630Pv2 to ath79-tiny target
These devices only have 6MiB available for firmware, which is not
enough for recent release images, so move these to the tiny target.
Note for users sysupgrading from the previous ath79-generic snapshot
images:
The tiny target kernel has a 4Kb flash erase block size instead
of the generic target's 64kb. This means the JFFS2 overlay partition
containing settings must be reformatted with the new block size or else
there will be data corruption.
To do this, backup your settings before upgrading, then during the
sysupgrade, de-select "Keep Settings". On the CLI, use "sysupgrade -n".
If you forget to do this and your system becomes unstable after
upgrading, you can do this to format the partition and recover:
* Reboot
* Press RESET when Power LED blinks during boot to enter Failsafe mode
* SSH to 192.168.1.1
* Run "firstboot" and reboot
Signed-off-by: Joe Mullally <jwmullally@gmail.com> Tested-by: Robert Högberg <robert.hogberg@gmail.com> Signed-off-by: Petr Štetiar <ynezz@true.cz> [commit message facelift]
(cherry picked from commit 44e1e5d) Signed-off-by: Petr Štetiar <ynezz@true.cz>
bcm27xx: add AMP2 to HifiBerry DAC+ / DAC+ Pro package
According to the vendor [1] these HATs share the same DT overlay:
hifiberry-dacplus. The PCM512x-compatible control unit is attached to
I2C, so the additional snd-soc-pcm512x-i2c kernel module is required.
Also explicitly note the Amp2 support to reduce confusion for those
users.
Flashing:
TFTP boot initramfs image and then perform sysupgrade. Follow common
MikroTik procedure as in https://openwrt.org/toh/mikrotik/common.
Note: following 781d4bfb397cdd12ee0151eb66c577f470e3377d
The network setup avoids using the integrated switch and connects the
single Ethernet port directly. This way, link speed (10/100 Mbps) is
properly reported by eth0.
- QCA9533 (650 MHz), 64 or 128MB RAM, 16MB SPI NOR
- 2x 10/100 Mbps Ethernet, with 802.3at PoE support (WAN)
- 2T2R 802.11b/g/n 2.4GHz
Flash instructions:
If your device comes with generic QSDK based firmware, you can login
over telnet (login: root, empty password, default IP: 192.168.188.253),
issue first (important!) 'fw_setenv' command and then perform regular
upgrade, using 'sysupgrade -n -F ...' (you can use 'wget' to download
image to the device, SSH server is not available):
In case your device runs firmware with YunCore custom GUI, you can use
U-Boot recovery mode:
1. Set a static IP 192.168.0.141/24 on PC and start TFTP server with
'tftp' image renamed to 'upgrade.bin'
2. Power the device with reset button pressed and release it after 5-7
seconds, recovery mode should start downloading image from server
(unfortunately, there is no visible indication that recovery got
enabled - in case of problems check TFTP server logs)
Signed-off-by: Clemens Hopfer <openwrt@wireloss.net> Signed-off-by: Thibaut VARÈNE <hacks@slashdirt.org>
(cherry-picked from commit a05dcb07241aa83a4416b56201e31b4af8518981)
[switch to mtd-mac-address instead of nvmem-cells]
- QCA9563 (775MHz), 128MB RAM, 16MB SPI NOR
- 2T2R 802.11b/g/n 2.4GHz
- 2T2R 802.11n/ac 5GHz
- 2x 10/100/1000 Mbps Ethernet, with 802.3at PoE support (WAN port)
LED for 5 GHz WLAN is currently not supported as it is connected directly
to the QCA9882 radio chip.
Flash instructions:
If your device comes with generic QSDK based firmware, you can login
over telnet (login: root, empty password, default IP: 192.168.188.253),
issue first (important!) 'fw_setenv' command and then perform regular
upgrade, using 'sysupgrade -n -F ...' (you can use 'wget' to download
image to the device, SSH server is not available):
In case your device runs firmware with YunCore custom GUI, you can use
U-Boot recovery mode:
1. Set a static IP 192.168.0.141/24 on PC and start TFTP server with
'tftp' image renamed to 'upgrade.bin'
2. Power the device with reset button pressed and release it after 5-7
seconds, recovery mode should start downloading image from server
(unfortunately, there is no visible indication that recovery got
enabled - in case of problems check TFTP server logs)
- CVE-2022-25640: A TLS v1.3 server who requires mutual authentication
can be bypassed. If a malicious client does not send the
certificate_verify message a client can connect without presenting a
certificate even if the server requires one.
- CVE-2022-25638: A TLS v1.3 client attempting to authenticate a TLS
v1.3 server can have its certificate heck bypassed. If the sig_algo in
the certificate_verify message is different than the certificate
message checking may be bypassed.
Tony Ambardar [Mon, 12 Apr 2021 23:46:22 +0000 (16:46 -0700)]
bpftools: fix feature override for masking clang
Rename feature variable clang-bpf-global-var following upstream changes.
This restores the HAVE_CLANG feature override and should avoid rare build
errors where a recent host clang and BTF-enabled host kernel are present.
Fixes: 23be333401f0 ("bpftools: update to 5.10.10") Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
(cherry picked from commit cf20f1bb5f0479c2509dd651d08e235a3b9e8755)
Michael Pratt [Sat, 11 Sep 2021 02:06:10 +0000 (22:06 -0400)]
ramips: fix reboot for remaining 32 MB boards
The following devices have a Winbond W25Q256FV flash chip,
which does not have the RESET pin enabled by default,
and otherwise would require setting a bit in a status register.
Before moving to Linux 5.4, we had the patch:
0053-mtd-spi-nor-add-w25q256-3b-mode-switch.patch
which kept specific flash chips with explicit 3-byte and 4-byte address modes
to stay in 3-byte address mode while idle (after an erase or write)
by using a custom flag SPI_NOR_4B_READ_OP that was part of the patch.
this was obsoleted by the patch:
481-mtd-spi-nor-rework-broken-flash-reset-support.patch
which uses the newer upstream flag SNOR_F_BROKEN_RESET
for devices with a flash chip that cannot be hardware reset with RESET pin
and therefore must be left in 3-byte address mode when idle.
The new patch requires that the DTS of affected devices
have the property "broken-flash-reset", which was not yet added for most of them.
This commit adds the property for remaining affected devices in ramips target,
specifically because of the flash chip model.
However, it is possible that there are other devices
where the flash chip uses an explicit 4-byte address mode
and the RESET pin is not connected to the SOC on the board,
and those DTS would also need this property.
Ref: 22d982ea0033 ("ramips: add support for switching between 3-byte and 4-byte addressing")
Ref: dfa521f12953 ("generic: spi-nor: rework broken-flash-reset") Signed-off-by: Michael Pratt <mcpratt@pm.me>
[pepe2k@gmail.com: backported to 21.02] Fixes: #9655, #9636, #9547 Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
(backported from commit 74516f4357d281f093f0daac01c4c5c239acc443)
Petr Štetiar [Thu, 24 Mar 2022 05:52:37 +0000 (06:52 +0100)]
imagebuilder: fix broken image generation with external targets
When using external targets there is a symlink being created for the
target under target/linux which then becomes dangling under Image
Builder. Fix it by dereferencing the possible symlink.
Tested on IB with external target, ipq40xx and mvebu.
ath79: fix label MAC address for Ubiquiti UniFi AP Outdoor+
The label has the MAC address of eth0, not the WLAN PHY address. We can
merge the definition back into ar7241_ubnt_unifi.dtsi, as both DTS
derived from it use the same interface for their label MAC addresses
after all.
This turned out this is an issue with upstream changing
ATA_TAG_INTERNAL's value from 31 to 32 during 4.18 release.
Update "SATA_DWC_QCMD_MAX" to account for that.
Marek Behún [Sat, 8 Jan 2022 18:56:02 +0000 (19:56 +0100)]
mvebu: SFP backports for GPON modules
This backports the following upstream Linux patches
net: sfp: add mode quirk for GPON module Ubiquiti U-Fiber Instant
net: sfp: relax bitrate-derived mode check
net: sfp: cope with SFPs that set both LOS normal and LOS inverted
for 5.4 for mvebu platform.
Josef Schlehofer [Sun, 20 Mar 2022 22:28:41 +0000 (23:28 +0100)]
cypress-firmware: drop several packages
1. Drop package: cypress-firmware-4359-pcie
This binary is no longer provided and there are not many details what
happened.
2. Drop package: cypress-firmware-4359-sdio
This binary is no longer provided, but in this case, to compare it with
PCIe package mention as first, there was added
support in Linux-firmware [1], but no sign of firmware file.
4. Drop package: cypress-firmware-89459-pcie [2]
According to Infineon: "CYW89459 is an automotive Wi-Fi chip which is not
supported in the broad market community."
Josef Schlehofer [Wed, 16 Mar 2022 12:14:57 +0000 (13:14 +0100)]
cypress-firmware: update it to version 5.4.18-2021_0812
- Binary files were renamed to cyfmac from brcmfmac, but the files needs
to be on the router with the previous naming
[ 6.656165] brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43455-sdio for chip BCM4345/6
[ 6.665182] brcmfmac mmc1:0001:1: Direct firmware load for brcm/brcmfmac43455-sdio.bin failed with error -2
[ 6.674928] brcmfmac mmc1:0001:1: Falling back to sysfs fallback for: brcm/brcmfmac43455-sdio.bin
- Cypress were acquired by Infineon Technologies
Thus change the project URL and switch to download files from their
GitHub repository. This is much better than the previous solution, which
requires finding new threads on their community forum about new driver
updates, and it will be necessary to change the URL each time.
Unfortunately, it seems that there is not published changelog, but
according to this forum thread [1], be careful by opening the link from
solution since it contains ending bracket ), it brings fixes for various
security vulnerabilities, which were fixed in 7_45_234.
Fixes:
- FragAttacks
- Kr00k
Also add LICENSE file
Run tested on Seeedstudio router powered by Raspberry Pi 4 CM with
package cypress-firmware-43455-sdio.
Petr Štetiar [Thu, 24 Mar 2022 05:45:04 +0000 (06:45 +0100)]
zlib: backport security fix for a reproducible crash in compressor
Tavis has just reported, that he was recently trying to track down a
reproducible crash in a compressor. Believe it or not, it really was a
bug in zlib-1.2.11 when compressing (not decompressing!) certain inputs.
Tavis has reported it upstream, but it turns out the issue has been
public since 2018, but the patch never made it into a release. As far as
he knows, nobody ever assigned it a CVE.
Petr Štetiar [Tue, 8 Mar 2022 09:41:24 +0000 (10:41 +0100)]
sunxi: cortexa7: fix ethernet link detection on a20-olinuxino-lime2
a20-olinuxino-lime2 is currently having hard time with link detection of
certain 1000Mbit partners due to usage of generic PHY driver, probably
due to following missing workaround introduced in upstream in commit 3aed3e2a143c ("net: phy: micrel: add Asym Pause workaround"):
The Micrel KSZ9031 PHY may fail to establish a link when the Asymmetric
Pause capability is set. This issue is described in a Silicon Errata
(DS80000691D or DS80000692D), which advises to always disable the
capability. This patch implements the workaround by defining a KSZ9031
specific get_feature callback to force the Asymmetric Pause capability
bit to be cleared.
This fixes issues where the link would not come up at boot time, or when
the Asym Pause bit was set later on.
As a20-olinuxino-lime2 has Micrel KSZ9031RNXCC-TR Gigabit PHY since
revision H, so we need to use Micrel PHY driver on those devices.
projects / thirdparty / openwrt.git / log
