Commit Graph

441 Commits

Author SHA1 Message Date
John Audia
51c7e1d25f kernel: bump 5.4 to 5.4.144
Manually rebased:
  backport-5.4/370-netfilter-nf_flow_table-fix-offloaded-connection-tim.patch

All other patches automatically rebased.

Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800

Signed-off-by: John Audia <graysky@archlinux.us>
2021-09-13 17:08:17 +02:00
Hauke Mehrtens
b9cc16a5e8 kernel: Set CONFIG_RCU_CPU_STALL_TIMEOUT=21
The default value for CONFIG_RCU_CPU_STALL_TIMEOUT was changed from 60
seconds to 21 seconds in 2012 in the upstream kernel. Some targets
already use 21 seconds.

This patch changes the default value in the generic configuration to 21
seconds and removes the target specific configuration options.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Acked-by: Rui Salvaterra <rsalvaterra@gmail.com>
2021-09-09 20:24:12 +02:00
Chuanhong Guo
43f0e386d4 mediatek: add support for TOTOLINK A8000RU
Specifications:
- SoC: MT7622
- RAM: 512MB
- Flash: MX35LF1GE4AB 128MB SPI NAND
- Ethernet: RTL8367S 5x1GbE
- WiFi: 2.4G: MT7622 5G: MT7615N x2
- Other ports: USB3.0 x1

Flash instruction:
*important*: upgrade vendor firmware to at least V7.1cu.643_B20200521
1. hold the reset button and power on the device. wait for about 10s
   before releasing the reset button.
2. upload sysupgrade.bin via u-boot recovery page on http://192.168.1.1

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
2021-09-03 15:53:28 +08:00
Rui Salvaterra
505b7a2d08 kernel: move two symbols to the generic kconfigs
CONFIG_RCU_{NEED_SEGCBLIST,STALL_COMMON} are set basically everywhere. Move them
to the generic kconfigs. And resort the generic kconfigs while at it.

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
2021-08-29 17:07:19 +02:00
Rui Salvaterra
7e135add24 kernel: disable CONFIG_RCU_EXPERT and friends
Based on the existing documentation [1][2], I dare anyone to demonstrate that
we need to fine-tune these RCU parameters. The (performance) breakage potential
for doing so is immense, so let's just please put down this loaded footgun.

Disable CONFIG_RCU_EXPERT and its dependent symbols. Additionally, remove the
CONFIG_RCU_EXPERT symbol from the target kconfigs which contain it.

[1] https://www.kernel.org/doc/Documentation/RCU/Design/Data-Structures/Data-Structures.html
[2] https://lwn.net/Articles/777214/

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
2021-08-29 17:03:38 +02:00
John Audia
6b1cd3e345 kernel: bump 5.10 to 5.10.61
Manually rebased:
  bcm27xx/patches-5.10/950-1031-net-lan78xx-Ack-pending-PHY-ints-when-resetting.patch

Removed upstreamed:
  mvebu/patches-5.10/101-cpufreq-armada-37xx-forbid-cpufreq-for-1.2-GHz-variant.patch

All other patches automatically rebased.

Build system: x86_64
Build-tested: bcm2711/RPi4B
Run-tested: bcm2711/RPi4B

No dmesg regressions, everything functional

Signed-off-by: John Audia <graysky@archlinux.us>
2021-08-29 16:30:20 +02:00
Chuanhong Guo
97e32e9702 mediatek: mt7623: replace kconfig symbol for snand driver
this change was missing during the spi-nand driver replacement, causing
build failure.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
2021-08-27 22:22:29 +08:00
Chuanhong Guo
472437b93f mediatek: mt7629: manually add missing kconfig
These options doesn't show up unless some kernel packages are selected,
so they are not covered by kernel_menuconfig.
Manually add them so that a build with all kmods can pass without user
interaction.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
2021-08-27 15:54:53 +08:00
Chuanhong Guo
8f2cace0b7 mediatek: mt7629: build with linux 5.10
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
2021-08-27 10:26:25 +08:00
Chuanhong Guo
9abd2fb6c4 mediatek: bmt: set default bmt oob offset to 0
there's no driver level remapping of oob data in the new spi-nand
driver and bmt oob signature starts at 0x0 of the dumped oob data.
change the default value to 0 for the new spi-nand driver.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
2021-08-27 10:26:25 +08:00
Chuanhong Guo
388127abd1 mediatek: mtk-snand: check request size against chip info
mtd->size will be overrided by BMT which makes all mtd requests made by
bmt fail in request size checking.
this commit changes the driver to check against actual chip size in chip
info as a workaround.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
2021-08-27 10:26:25 +08:00
Chuanhong Guo
b600aee3ed mediatek: attach bmt to the new snand driver
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
2021-08-27 10:26:25 +08:00
Chuanhong Guo
01b452fe2d mediatek: change dts to use the new snand driver
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
2021-08-27 10:26:25 +08:00
Chuanhong Guo
c6ed31630d mediatek: remove spi-nand hacks from 5.10
we now have a standalone mtd driver and the old spi-mem driver along
with the hack in spi-nand core can be removed.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
2021-08-27 10:26:25 +08:00
Weijie Gao
e8e8487603 mediatek: enable new spi-nand driver for kernel 5.10
This patch enables new spi-nand driver for mt7622 and mt7629.

Signed-off-by: Weijie Gao <hackpascal@gmail.com>
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
2021-08-27 10:26:25 +08:00
Weijie Gao
050621aa01 mediatek: add a new spi-nand driver for kernel 5.10
This patch adds a new spi-nand driver which implements the SNFI of mt7622
and mt7629.

Unlike the existing snfi driver which makes use of the spi-mem framework
and the spi-nand framework with modified ecc support, this driver is
implemented directly on the mtd framework with other components untouched,
and provides better performance, and behaves exactly the same as the nand
framework.

Signed-off-by: Weijie Gao <hackpascal@gmail.com>
2021-08-27 10:26:24 +08:00
Daniel Golle
f4493c647a
mediatek: mt7623: enable clock drivers for Mali-450 and audio
Built clock drivers for G3DSYS and AUDSYS into the kernel to allow
multimedia features (GPU and audio) to work if they exist.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-08-22 14:34:29 +01:00
Daniel Golle
4f1c5b01c1
mediatek: mt7623: backport musb, improve HDMI console
Backport support for dual-role USB 2.0 as that's what is actually
built-into MT7623.
Improve HDMI console by enabling VT and setting up tty1..tty6.
Re-add accidentally removed CONFIG_ARM_ARCH_TIMER.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-08-21 17:23:36 +01:00
John Audia
f25cebc43c kernel: bump 5.4 to 5.4.142
Removed upstreamed:
  hack-5.4/991-platform-x86-pcengines-apuv2-revert-simswitch.patch

All other patches automatically rebased.

Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800

No dmesg regressions, everything functional

Signed-off-by: John Audia <graysky@archlinux.us>
2021-08-21 15:59:39 +02:00
Daniel Golle
0d3f3323a2
mediatek: mt7623: enable more hardware features
In order to make HDMI console available on the BananaPi BPi-R2 select
various Kconfig symbols which are useful for systems with graphics.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-08-21 04:55:24 +01:00
Adrian Schmutzler
05ccca031e mediatek: convert mtd-mac-address to nvmem for MT7629 RFB
This conversion appears to have been overlooked since it's in a
kernel patch.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
2021-08-20 18:13:55 +02:00
John Audia
ed9341dd78 kernel: bump 5.4 to 5.4.140
Removed upstreamed bcm27xx/patches-5.4:
    950-0977-USB-gadget-f_hid-avoid-crashes-and-log-spam.patch
    950-0980-SQUASH-USB-gadget-f_hid-remove-more-spam.patch

All other patches automatically rebased.

Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800

No dmesg regressions, everything functional

Signed-off-by: John Audia <graysky@archlinux.us>
2021-08-14 20:25:25 +02:00
John Audia
02e2723ef3 kernel: bump 5.4 to 5.4.139
All patches automatically rebased.

Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800

No dmesg regressions, everything functional

Signed-off-by: John Audia <graysky@archlinux.us>
2021-08-14 20:25:19 +02:00
Rui Salvaterra
6ec859d231 kernel: bump 5.10 to 5.10.58
No deleted or manually refreshed patches.

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
2021-08-14 16:32:34 +01:00
David Bauer
101c0c00a7 mediatek: add missing config symbols
Signed-off-by: David Bauer <mail@david-bauer.net>
2021-08-13 00:52:30 +02:00
David Bauer
4c2a8b00fe mediatek: add missing config symbol
Signed-off-by: David Bauer <mail@david-bauer.net>
2021-08-11 02:15:13 +02:00
David Bauer
1c9a9f7c7a treewide: add various missing config symbols
Fixes commit 91a52f22a1 ("treewide: backport support for nvmem on non platform devices")

Signed-off-by: David Bauer <mail@david-bauer.net>
2021-08-10 20:22:59 +02:00
Daniel Golle
a0b827b916
kirkwood: import patch to fix build and refresh patches
kirkwood build broke due to missing include needed for ETH_ALEN.
Add patch (sent upstream as well) to address that.
Refresh patches for 5.4 and 5.10.

Fixes: 91a52f22a1 ("treewide: backport support for nvmem on non platform devices")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-08-05 23:53:07 +01:00
David Bauer
c3d99ae792 mediatek: mt7622: enable HW RNG
MT7622 provides are hardware RNG with upstream Linux driver. Enable
compilation of this driver to make use of the hardware RNG.

Signed-off-by: David Bauer <mail@david-bauer.net>
2021-08-05 01:48:55 +02:00
John Audia
28ef764026 kernel: bump 5.4 to 5.4.135
All patches automatically rebased.

Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800

No dmesg regressions, everything functional

Signed-off-by: John Audia <graysky@archlinux.us>
2021-07-31 19:13:00 +02:00
Rui Salvaterra
a20e6149b3 kernel: bump 5.10 to 5.10.54
Deleted (upstreamed) patches:
mediatek/patches-5.10/000-spi-fix-fifo.patch

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
2021-07-29 19:50:16 +01:00
Nick Hainke
e8e2b88f5f mediatek: mt7623: import patch to fix msi warning
The 1st generation MediaTek PCIe host bridge cannot handle Message
Signaled Interrupts (MSIs). The core PCI code is not aware that MSI is
not available. This results in warnings of the form:

WARNING: CPU: 2 PID: 112 at include/linux/msi.h:219
  pci_msi_setup_msi_irqs.constprop.8+0x64/0x6c
Modules linked in: ahci(+) libahci libata sd_mod scsi_mod
  gpio_button_hotplug
CPU: 2 PID: 112 Comm: kmodloader Not tainted 5.10.52 #0
Hardware name: Mediatek Cortex-A7 (Device Tree)

Import patches that introduce the 'no_msi' attribute to signal missing
MSI support to the core PCI.

Refresh patches:
- 000-spi-fix-fifo.patch
- 330-mtk-bmt-support.patch
- 510-net-mediatek-add-flow-offload-for-mt7623.patch
- 601-PCI-mediatek-Use-regmap-to-get-shared-pcie-cfg-base.patch
- 610-pcie-mediatek-fix-clearing-interrupt-status.patch
- 700-net-ethernet-mtk_eth_soc-add-support-for-coherent-DM.patch
- 710-pci-pcie-mediatek-add-support-for-coherent-DMA.patch

Signed-off-by: Nick Hainke <vincent@systemli.org>
2021-07-29 15:08:17 +01:00
Hauke Mehrtens
8c2509dc5f mediatek: Fix Fidelix-FM35X1GA patch
This patch did not apply cleanly any more after support for the XTX
flash was added to the generic patches.

Fixes: 92012dd867 ("kernel: Add support for XTX XT26G02A SPI NAND")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
2021-07-25 16:11:29 +02:00
Nick Hainke
90db922f88 mediatek: mt7623: import patch to fix spi
The patch fixes the fifo rx mode for the mt7623. It is already accepted
upstream for Linux Kernel 5.15.

To test the spi we can change the dts file to

&spi0 {
	pinctrl-names = "default";
	pinctrl-0 = <&spi0_pins_a>;
	status = "okay";
	spidev: spidev@0 {
		compatible = "linux,spidev";
		spi-max-frequency = <1000000>;
		reg = <0>;
	};
};

Afterwards we should see a spidev device under /dev/.
To test it we can further use spidev-test.

Signed-off-by: Nick Hainke <vincent@systemli.org>
2021-07-24 13:10:53 +01:00
Daniel Golle
089c2bb217 mediatek: mt7623: yet another legacy image variable restored
Restore BOOTFS_BLOCKS variable until U7623 gets brushed, so things
hopefully build now.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-07-21 16:48:35 +01:00
Daniel Golle
5096f0513b mediatek: mt7623: allow build without ext4 rootfs
The image generation code for the U7623 board expects ext4 filesystem
to be selected in menuconfig and CONFIG_TARGET_ROOTFS_PARTSIZE to be
defined. Now that ext4 isn't enabled any more, the variable was missing
and broke the build.
Set the default (104) instead of using the config variable to fix that.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-07-21 15:34:21 +01:00
Daniel Golle
c5b3c13b60 mediatek: mt7623: add back needed build steps for U7623 board
When reworking the BPi-R2 the mtk-mmc-img build step was removed
despite it was still needed to build the image for the UniElec U7623
board. Add it back for now until U7623 gets its facelift.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-07-21 10:26:44 +01:00
Daniel Golle
fe7596192e mediatek: remove ext4 images from mt7623 build
A full read-write rootfs cannot work inside uImage.FIT as the hash
will obviously change once writing to it. Disable generating ext4
rootfs images.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-07-20 18:43:56 +01:00
Daniel Golle
27deb35af3 mediatek: fix typo in kernel mt7623/config-5.10
Fix typo reported by github user @mans0n.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-07-20 17:03:44 +01:00
Daniel Golle
ddec752890 mediatek: drop kmod-mmc
Support for MMC is built-into the kernel anyway, no need to select the
(empty/stub) kmod package.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-07-20 17:01:02 +01:00
Daniel Golle
86a61e716e mediatek: rework Bananapi BPi-R2
* introduce recovery (=initramfs) vs. production dual boot scheme
 * make use of uImage.FIT (instead of FAT partition)
 * generate images using build steps (instead of external scripts)
 * simplify sysupgrade and config restore (thanks to uImage.FIT)
 * make sure mmc devices are ordered persistently (set DT aliases)

This commit breaks sysupgrade from existing installations, you will
have to re-install using the sdcard.img.gz image.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-07-20 04:16:48 +01:00
Daniel Golle
932366fdd9 mediatek: mt7623: import patch to fix flow offloading
Hardware flow offloading was reported to work when setting the right
version identifier. Import a patch from Frank Wunderlich doing that.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-07-20 04:11:05 +01:00
Daniel Golle
bce7777222 mediatek: mt7623: build with Linux 5.10
Update Kernel config and set Linux 5.10 for mediatek/mt7623.
(patches have already been updated to 5.10 when mt7622 was bumped)
Tested on Bananapi BPi-R2.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2021-07-20 04:11:05 +01:00
Ansuel Smith
1e6f330ccf mediatek: convert mtd-mac-address to nvmem implementation
Define nvmem-cells and convert mtd-mac-address to nvmem implementation.
The conversion is done with an automated script.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
2021-07-19 14:51:22 +02:00
Rui Salvaterra
b71767d8e0 kernel: bump 5.10 to 5.10.50
No deleted or manually refreshed patches.

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
2021-07-17 13:27:03 +02:00
John Audia
2fe1b32156 kernel: bump 5.4 to 5.4.132
Manually rebased:
  layerscape/patches-5.4/805-display-0002-drm-rockchip-prepare-common-code-for-cdns-and-rk-dpi.patch

All other patches automatically rebased.

Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800

No dmesg regressions, everything functional

Signed-off-by: John Audia <graysky@archlinux.us>
2021-07-17 13:21:03 +02:00
John Audia
d8dc9f108a kernel: bump 5.4 to 5.4.126
Manually rebased:
  bcm27xx/patches-5.4/950-0089-cgroup-Disable-cgroup-memory-by-default.patch

All other patches automatically rebased.

Signed-off-by: John Audia <graysky@archlinux.us>
2021-06-26 12:49:15 +02:00
Rui Salvaterra
b92794f4d0 kernel: bump 5.10 to 5.10.46
Add the new symbol to the generic kconfig.

No deleted or manually refreshed patches.

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
2021-06-26 12:43:54 +02:00
John Audia
752cd29904 kernel: bump 5.4 to 5.4.124
Manually rebased:
  generic/hack-5.4/662-remove_pfifo_fast.patch

All other patches automatically rebased.

Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800

No dmesg regressions, everything functional

Signed-off-by: John Audia <graysky@archlinux.us>
[manual changes to ramips/patches-5.10/835-asoc-add-mt7620-support.patch]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
2021-06-06 00:25:25 +02:00
Rui Salvaterra
da77b9c2f1 kernel: bump 5.10 to 5.10.42
Deleted (reverse-appliable):
generic/pending-5.10/110-perf-jevents-fix-getting-maximum-number-of-fds.patch

Manually refreshed:
ramips/patches-5.10/835-asoc-add-mt7620-support.patch

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
2021-06-06 00:24:07 +02:00