You digging too deep to tell me your current position by few words 
resets are not found by ryders code…because they are deleted…i added them this way:
281 resets = <&hifsys HIFSYS_PCIE0_RST>,
282 <&hifsys HIFSYS_PCIE1_RST>,
283 <&hifsys HIFSYS_PCIE2_RST>;
but now nothing is printed on pci-enum (not the debug-messages i see before)
https://github.com/frank-w/u-boot/commits/sata_test_ryder
accidentaly removed my debug-messages by “git reset”
now i see much more, but resets still 0
DEBUG: Passed mtk_pcie_probe 247
DEBUG: Passed mtk_pcie_probe 252
DEBUG: Passed mtk_pcie_probe 258
DEBUG: Passed mtk_pcie_probe 264
DEBUG: Passed mtk_pcie_probe 269
DEBUG: Passed mtk_pcie_probe 273
DEBUG: Passed mtk_pcie_probe 278
DEBUG: Passed mtk_pcie_parse_port 207
DEBUG: Passed mtk_pcie_parse_port 211
DEBUG: Passed mtk_pcie_parse_port 216
DEBUG: Passed mtk_pcie_parse_port 221 0,0,0
DEBUG: Passed mtk_pcie_parse_port 225
DEBUG: Passed mtk_pcie_parse_port 229
DEBUG: Passed mtk_pcie_parse_port 232
DEBUG: Passed mtk_pcie_parse_port 235
DEBUG: Passed mtk_pcie_probe 269
DEBUG: Passed mtk_pcie_probe 273
DEBUG: Passed mtk_pcie_probe 278
DEBUG: Passed mtk_pcie_parse_port 207
DEBUG: Passed mtk_pcie_parse_port 211
DEBUG: Passed mtk_pcie_parse_port 216
DEBUG: Passed mtk_pcie_parse_port 221 0,0,0
DEBUG: Passed mtk_pcie_parse_port 225
DEBUG: Passed mtk_pcie_parse_port 229
DEBUG: Passed mtk_pcie_parse_port 232
DEBUG: Passed mtk_pcie_parse_port 235
DEBUG: Passed mtk_pcie_probe 269
DEBUG: Passed mtk_pcie_probe 273
DEBUG: Passed mtk_pcie_probe 278
DEBUG: Passed mtk_pcie_parse_port 207
DEBUG: Passed mtk_pcie_parse_port 211
DEBUG: Passed mtk_pcie_parse_port 216
DEBUG: Passed mtk_pcie_parse_port 221 0,0,0
DEBUG: Passed mtk_pcie_parse_port 225
DEBUG: Passed mtk_pcie_parse_port 229
DEBUG: Passed mtk_pcie_parse_port 232
DEBUG: Passed mtk_pcie_parse_port 235
DEBUG: Passed mtk_pcie_probe 285
DEBUG: Passed mtk_pcie_enable_port 172
DEBUG: Passed mtk_pcie_enable_port 184
DEBUG: Passed mtk_pcie_enable_port 192
DEBUG: Passed mtk_pcie_enable_port 172
DEBUG: Passed mtk_pcie_enable_port 184
DEBUG: Passed mtk_pcie_enable_port 192
DEBUG: Passed mtk_pcie_enable_port 172
DEBUG: Passed mtk_pcie_enable_port 184
DEBUG: Passed mtk_pcie_enable_port 192
ok let’s look at pinmux in dts…
should be this: https://github.com/frank-w/BPI-R2-4.14/blob/6a8cd4974ab40922aa8320f7eee1c8cc26b084de/arch/arm/boot/dts/mt7623.dtsi#L1469
seems dt-bindings for pinctrl are not there…
dt-bindings/pinctrl/mt7623-pinfunc.h
$ ls -1 ./arch/arm/dts/include/dt-bindings/pinctrl/
am33xx.h
am43xx.h
at91.h
bcm2835.h
dra.h
hisi.h
k3.h
omap.h
pads-imx8qm.h
pads-imx8qxp.h
pinctrl-snapdragon.h
pinctrl-tegra.h
pinctrl-tegra-xusb.h
pins-imx8mq.h
r7s72100-pinctrl.h
rockchip.h
stm32-pinfunc.h
sun4i-a10.hbecause you lack of pinmux setting in dts and phy driver
and if i read the PCI core correctly - pci_uclass_pre_probe() in pci-uclass.c will parse range automatically.
--- a/arch/arm/dts/mt7623n-bananapi-bpi-r2.dts
+++ b/arch/arm/dts/mt7623n-bananapi-bpi-r2.dts
@@ -172,6 +172,13 @@
};
};
+ pcie_default: pcie-default {
+ mux {
+ function = "pcie";
+ groups = "pcie0_0_perst", "pcie1_0_perst";
+ };
+ };
+
uart0_pins_a: uart0-default {
mux {
function = "uart";
@@ -201,6 +208,28 @@
};
};
+&pcie {
+ pinctrl-names = "default";
+ pinctrl-0 = <&pcie_default>;
+ status = "okay";
+
+ pcie@0,0 {
+ status = "okay";
+ };
+
+ pcie@1,0 {
+ status = "okay";
+ };
+};
+
+&pcie0_phy {
+ status = "okay";
+};
+
+&pcie1_phy {
+ status = "okay";
+};
Just note that the form of uboot pinmux is different than linux mt7623 dts since we wrote a new pinctrl core after 4.19, and this new format is adopted by uboot 7623 at that time.
2 Likes
it’s time to copy complete DTS with all dependencies
that brings nothing… i tried to copy relavant parts an hung on dt-bindings…
that was the point i hung 
maybe we should make a separate thread for pcie/ahci?
that was the point i hung
use the version i posted and add a phy driver
i have started new thread to outsource discussion about this
http://forum.banana-pi.org/t/bpi-r2-uboot-pcie-sata-ahaci/9613
Hi,there was released an mtk xhci driver for uboot and i need help to port dts-nodes for r2 and r64
https://patchwork.ozlabs.org/patch/1252654/
Linux:
- https://github.com/frank-w/BPI-R2-4.14/blob/5.4-main/arch/arm/boot/dts/mt7623.dtsi#L1146
- https://github.com/frank-w/BPI-R2-4.14/blob/c62152c808341ab630a4af9bae2e623c83a449ab/arch/arm64/boot/dts/mediatek/mt7622.dtsi#L720
My start on r2/mt7623:
https://github.com/frank-w/u-boot/commit/09a9b9dabf067ce71b46171bf0982cc4982ca47c
Compiles,but crashes on “usb start”
There are 2 main differences between linux and uboot driver:
- uboot seems to use relative addresses for ports where linux uses absolute addresses
- uboot uses ranges in phy where linux defines reg
Anyone can help?
1 Like
Already done…r2 has xhci v0.96 which is not yet supported by driver,but dts works. On r64 my dts is still broken…don’t get the xhci version info
Please feel free to ask him for any xhci problems.
Chunfeng yun is currently busy because of debugging clock-issue for mt7629.
Any idea about emmc-issue on r64? I’ve wrote to weijie gao and chaotian jing,but no response yet
they’re all busy. Maybe loop this guys along with wenbin.mei and cc me
looks like your change to the mt7622_compat-struct fixes the problem, i do not get any error on mount. i only see some errors in log while booting mt7622_mmctest.log (29,8 KB)
[ 1.367124] mtk-msdc 11230000.mmc: msdc_cmd_done: cmd=52 arg=00000C00; rsp 00000000; cmd_error=-110
[ 1.376165] mtk-msdc 11230000.mmc: msdc_track_cmd_data: cmd=52 arg=00000C00; host->error=0x00000002
same for 11240000
#define ETIMEDOUT 110 /* Connection timed out */
so i guess the timeout happens here:
Just to link original error-message: Bpi-r64 quick start (boot from eMMC)
The previous patch was workaround and this error is normal. We need to figure out the root cause, so can you apply this kernel patch to print more details https://pastebin.com/va8s8hr0
I need two full uart logs and email to me - (old uboot + kernel) and (upstream uboot + kernel)
looks like the new patch causes bootloop…booting now for 200s…maybe i need to include initrd? i guess reads from sdcard are also recorded…
boot with initrd (buildroot) takes 8 seconds 
i need to revert the prior patch, is this right? or should leave the compat-struct values?
mt7622_mmctest.log (186,1 KB)
only apply this patch.
legacy uboot (uboot 2014 (not upstream version you’re using)) + kernel
upstream uboot + kernel
it looks like for the emmc-mount issue there is a working Patch: https://github.com/frank-w/BPI-R2-4.14/commit/1ab1bcdb71b2ba02dce30c826827cc015f4bb4f5 after rebasing (to split debug+dts) it becomes this: https://github.com/frank-w/BPI-R2-4.14/commit/f644b6766b28c72fb6ccef5832201032808dc2ca
basicly the patches are applied to this branch:
https://github.com/frank-w/BPI-R2-4.14/commits/5.4-mmctest
in my quick test i do not get a error on mounting, can anybody else test this too?
got a new patch which seems fine too and should be the final solution which will be posted to patchwork, added it to my 5.4-main branch to include it into kernel-images built by travis-ci in my repo
There is also an emmc-issue in linux caused by missing pullup/pulldown pinctrl-definitions. This seems to affect only v1.1 r2-boards…i cannot reproduce issue on my v1.0 board. Maybe it’s caused by different vendor of emmc-chip (v1.0 is samsung, v1.1 is possibly sec)
Just for reference…patch is posted to mailinglist waiting for acknowledge from mtk
https://patchwork.ozlabs.org/project/uboot/patch/[email protected]/
on BPI-R2 we noticed, that pcie is not enumerable in uboot if no card is inserted in pcie-slot…so sata is also not recognized -reported to mtk they will look. I have also some patches for mt7622-pcie that seem to work,but i have not got sata working…i see an sata controller,but not drive connected
Edit: I got sata working for mt7622 based on pcie-patches i got from mtk. I hope they get posted to mainline soon, to post my ahci driver.
If anyone wants to try patches here is my tree:
https://github.com/frank-w/u-boot/commits/2020-07-bpi-r64-pciev2
Build it (bpi-r64,32bit mode) and install it or load via tftp (e.g. “run askubootnet”)
After bootup with sata drive connected, enter uboot console and enter
scsi scan
Your drive should be found and you can show content of it with
ls scsi 0:1
For first partition (1) if it is a partition with buildin filesystem-support.
BPI-R64> scsi scan
scanning bus for devices...
Target spinup took 0 ms.
AHCI 0001.0300 32 slots 1 ports 6 Gbps 0x1 impl SATA mode
flags: ncq stag pm clo only pmp pio slum part ccc apst
Device 0: (0:0) Vendor: ATA Prod.: ST750LM022 HN-M7 Rev: 2AR1
Type: Hard Disk
Capacity: 715404.8 MB = 698.6 GB (1465149168 x 512)
BPI-R64> ls scsi 0:1
EFI/
512 BOOTSECT.BAK
1 file(s), 1 dir(s)
BPI-R64>
also added my current state of xHCI for both devices there (dts-nodes + missing clock-definitions). both devices reporting xHCI 0.96 which is currently not supported by driver (hope it will do in near future)
BPI-R64> usb start
starting USB...
Bus usb@1a0c0000: hcd: 0x1a0c0000, ippc: 0x1a0c4700
u2p:2, u3p:1
Register 300010f NbrPorts 3
Starting the controller
USB XHCI 0.96
scanning bus usb@1a0c0000 for devices... 1 USB Device(s) found
BPI-R2> usb start
starting USB...
Bus usb@1a1c0000: hcd: 0x1a1c0000, ippc: 0x1a1c4700
u2p:1, u3p:1
Register 200010f NbrPorts 2
Starting the controller
USB XHCI 0.96
Bus usb@1a240000: hcd: 0x1a240000, ippc: 0x1a244700
u2p:1, u3p:1
Register 200010f NbrPorts 2
Starting the controller
USB XHCI 0.96
scanning bus usb@1a1c0000 for devices... 1 USB Device(s) found
scanning bus usb@1a240000 for devices... 1 USB Device(s) found
scanning usb for storage devices... 0 Storage Device(s) found
good news: USB is now also working on both devices R2&R64 
BPI-R2> usb start
starting USB...
Bus usb@1a1c0000: hcd: 0x1a1c0000, ippc: 0x1a1c4700
u2p:1, u3p:1
Register 200010f NbrPorts 2
Starting the controller
USB XHCI 0.96
Bus usb@1a240000: hcd: 0x1a240000, ippc: 0x1a244700
u2p:1, u3p:1
Register 200010f NbrPorts 2
Starting the controller
USB XHCI 0.96
scanning bus usb@1a1c0000 for devices... 1 USB Device(s) found
scanning bus usb@1a240000 for devices... 2 USB Device(s) found
scanning usb for storage devices... 1 Storage Device(s) found
BPI-R2> usb tree
USB device tree:
1 Hub (5 Gb/s, 0mA)
U-Boot XHCI Host Controller
1 Hub (5 Gb/s, 0mA)
| U-Boot XHCI Host Controller
|
+-2 Mass Storage (480 Mb/s, 200mA)
USB Flash Disk 906B030002F4
BPI-R2> ls usb 0:1
efi/
4767728 kernel
1 file(s), 1 dir(s)
=========================================================
BPI-R64> usb start
starting USB...
Bus usb@1a0c0000: hcd: 0x1a0c0000, ippc: 0x1a0c4700
u2p:2, u3p:1
Register 300010f NbrPorts 3
Starting the controller
USB XHCI 0.96
scanning bus usb@1a0c0000 for devices... 2 USB Device(s) found
scanning usb for storage devices... 1 Storage Device(s) found
BPI-R64> ls usb 0:1
91 mcurom.md5
System Volume Information/
1 file(s), 1 dir(s)
BPI-R64>
https://github.com/frank-w/u-boot/commits/2020-10-bpi
So at current state we only need to fix the pcie issue on r2 (watchdog reset if no card inserted).
Edit: latest patches for usb (except xhci 0.96 support) are now also applied to upstream
for r2 issue we have now a workaround…anybody who wants to use SATA on r2 without pcie-card inserted can disable pcie0 in dts (arch/arm/dts/mt7623n-bananapi-bpi-r2.dts)
--- a/arch/arm/dts/mt7623n-bananapi-bpi-r2.dts
+++ b/arch/arm/dts/mt7623n-bananapi-bpi-r2.dts
@@ -215,7 +215,7 @@
status = "okay";
pcie@0,0 {
- status = "okay";
+ status = "disabled";
};
pcie@1,0 {
now there is no hang on enabling port0 and scsi scan can be used