cwt22 5.12.0-1 fails to boot off NVME

[stronnag]

• Latest release packages installed
• default fdtflle set to starfive/jh7110-starfive-visionfive-2-v1.3b.dtb
• No nmve* device nodes under /dev

Boot log attached. cwt22-5.12.0-1-bootfail.txt

[cwt]

@stronnag Did you use the new u-boot from here: https://github.com/cwt-vf2/u-boot-starfive-vf2/releases/tag/2024.04-1 ?

[stronnag]

I did not (did not even know about it). I'll try that later this evening.

Thanks.

[cwt]

@stronnag yeah, sorry, I didn't have a time to write a proper document or update README yet.

The u-boot images will be installed to /usr/share/u-boot-starfive-vf2/

Which you will need flashcp to flash your board or dd them to SD partition 1 and 2.

[stronnag]

No difference, same stack trace and no nvme device nodes.

[cwt]

@stronnag might be a regression in 6.6 against your nvme controller? I don't know yet, below is my dmesg, on 1.2A board booted from SD:

[user@ArchVF2 ~]$ dmesg |grep -ie nvme
[   12.353204] nvme nvme0: pci function 0001:01:00.0
[   12.368635] nvme 0001:01:00.0: enabling device (0000 -> 0002)
[   12.450213] nvme nvme0: allocated 32 MiB host memory buffer.
[   12.466212] nvme nvme0: 4/0/0 default/read/poll queues
[   12.503786]  nvme0n1: p1 p2 p3 p4
[   12.615814] BTRFS: device label VF2 devid 1 transid 164714 /dev/nvme0n1p4 scanned by (udev-worker) (315)

[stronnag]

I will wait for your updated readme, maybe I have missed something.

[cwt]

can you show me your dmesg |grep -ie pcie?

user@ArchVF2 ~]$ dmesg |grep -ie pcie
[   10.618937] pcie-starfive 940000000.pcie: host bridge /soc/pcie@940000000 ranges:
[   10.633405] pcie-starfive 940000000.pcie:      MEM 0x0030000000..0x0037ffffff -> 0x0030000000
[   10.651402] pcie-starfive 940000000.pcie:      MEM 0x0900000000..0x093fffffff -> 0x0900000000
[   10.878381] pcie-starfive 940000000.pcie: port link up
[   11.037114] pcie-starfive 940000000.pcie: PCI host bridge to bus 0000:00
[   11.880502] pcie-starfive 9c0000000.pcie: host bridge /soc/pcie@9c0000000 ranges:
[   11.893526] pcie-starfive 9c0000000.pcie:      MEM 0x0038000000..0x003fffffff -> 0x0038000000
[   11.909391] pcie-starfive 9c0000000.pcie:      MEM 0x0980000000..0x09bfffffff -> 0x0980000000
[   12.133474] pcie-starfive 9c0000000.pcie: port link up
[   12.141236] pcie-starfive 9c0000000.pcie: PCI host bridge to bus 0001:00
[   12.268413] pci 0001:01:00.0: 4.000 Gb/s available PCIe bandwidth, limited by 5.0 GT/s PCIe x1 link at 0001:00:00.0 (capable of 31.504 Gb/s with 8.0 GT/s PCIe x4 link)

[stronnag]

The machine is just backing up the SSD now (sorry), however:

In the 6.6 console boot log, the only mention of pcie (not good) is:

Waiting 10 seconds for device /dev/disk/by-label/jagular.btrfs ...
[    4.501158] mmc0: Failed to initialize a non-removable card
[   13.650031] platform 940000000.pcie: deferred probe pending
[   13.661930] platform 9c0000000.pcie: deferred probe pending

From the V5 kernel:

$ dmesg | grep -ie pcie
[    1.564795] pcie_plda 2b000000.pcie: Failed to get power-gpio, but maybe it's always on.
[    1.564982] pcie_plda 2b000000.pcie: host bridge /soc/pcie@2B000000 ranges:
[    1.565023] pcie_plda 2b000000.pcie:      MEM 0x0030000000..0x0037ffffff -> 0x0030000000
[    1.565047] pcie_plda 2b000000.pcie:      MEM 0x0900000000..0x093fffffff -> 0x0900000000
[    1.697045] pcie_plda 2b000000.pcie: Port link up.
[    1.803173] pcie_plda 2b000000.pcie: PCI host bridge to bus 0000:00
[    1.811958] pcie_plda 2b000000.pcie: msi#0 address_hi 0x0 address_lo 0x190
[    1.817365] pcie_plda 2c000000.pcie: Failed to get power-gpio, but maybe it's always on.
[    1.817519] pcie_plda 2c000000.pcie: host bridge /soc/pcie@2C000000 ranges:
[    1.817559] pcie_plda 2c000000.pcie:      MEM 0x0038000000..0x003fffffff -> 0x0038000000
[    1.817582] pcie_plda 2c000000.pcie:      MEM 0x0980000000..0x09bfffffff -> 0x0980000000
[    1.959604] pcie_plda 2c000000.pcie: Port link up.
[    2.067124] pcie_plda 2c000000.pcie: PCI host bridge to bus 0001:00
[    2.071392] pci 0001:01:00.0: 4.000 Gb/s available PCIe bandwidth, limited by 5.0 GT/s PCIe x1 link at 0001:00:00.0 (capable of 31.504 Gb/s with 8.0 GT/s PCIe x4 link)
[    2.075941] pcie_plda 2c000000.pcie: msi#0 address_hi 0x0 address_lo 0x190
[    2.344483] pcie_plda 2c000000.pcie: msi#0 address_hi 0x0 address_lo 0x190
[    2.344501] pcie_plda 2c000000.pcie: msi#1 address_hi 0x0 address_lo 0x190
[    2.344515] pcie_plda 2c000000.pcie: msi#2 address_hi 0x0 address_lo 0x190
[    2.344528] pcie_plda 2c000000.pcie: msi#3 address_hi 0x0 address_lo 0x190
[    2.344541] pcie_plda 2c000000.pcie: msi#4 address_hi 0x0 address_lo 0x190

[stronnag]

And for nvme

The working kernel v5.

$ dmesg | grep -ie nvme
[    2.075756] nvme nvme0: pci function 0001:01:00.0
[    2.075817] nvme 0001:01:00.0: enabling device (0000 -> 0002)
[    2.309977] nvme nvme0: allocated 64 MiB host memory buffer.
[    2.348212] nvme nvme0: 4/0/0 default/read/poll queues
[    2.376807]  nvme0n1: p1 p2 p3 p4 p5 p6
[    5.716972] BTRFS: device label jagular.btrfs devid 1 transid 154495 /dev/nvme0n1p4 scanned by (udev-worker) (221)

and nothing for kernel v6.6

[stronnag]

As discussed in the RV forum, missing module / builtin. Fixed by

MODULES=(phy-jh7110-pcie)

Closing, as fix identified and new kernel in progress.