NEC/Renesas PCI USB3 Cards (uPD720201) Host halt failed, -19

[Fabian]

Hello,

Using Ubuntu installation media (ubuntu-18.04.2-server-amd64.iso / ubuntu-18.04.[1-2]-live-server-amd64.iso / Ubuntu Live 18.04.2 iso / Ubuntu Mate Live Ubuntu Live 18.04.[1-2] iso)

Startup is hanging with :
[ 0.214796] pci 0000:0d:00.0: [1912:0014] type 00 class 0x0c0330
[ 0.215620] pci 0000:0d:00.0: reg 0x10: [mem 0xfdbf0000-0xfdbf1fff 64bit]
[ 0.219160] pci 0000:0d:00.0: PME# supported from D0 D3hot D3cold
[ 79.272644] xhci_hcd 0000:0d:00.0: xHCI Host Controller
[ 79.272710] xhci_hcd 0000:0d:00.0: new USB bus registered, assigned bus number 7
[ 79.535236] xhci_hcd 0000:0d:00.0: Host halt failed, -19
[ 79.535291] xhci_hcd 0000:0d:00.0: can't setup: -19
[ 79.535347] xhci_hcd 0000:0d:00.0: USB bus 7 deregistered
[ 79.555578] xhci_hcd 0000:0d:00.0: init 0000:0d:00.0 fail, -19

I'm facing exactly the same bug as these one:
https://<email address hidden>/msg102138.html
https://bugs.gentoo.org/show_bug.cgi?id=612704

I installed Ubuntu Server 18.04.2 without the PCI card without trouble and then reinserted the PCI Card .... computer never start always hanging at the same steps as above.

I managed to start only once time with the PCI card plugged by adding "pci=nomsi" to the kernel command line using an Ubuntu Mate Live media with persistence and had this output:

[ 0.214796] pci 0000:0d:00.0: [1912:0014] type 00 class 0x0c0330
[ 0.215620] pci 0000:0d:00.0: reg 0x10: [mem 0xfdbf0000-0xfdbf1fff 64bit]
[ 0.219160] pci 0000:0d:00.0: PME# supported from D0 D3hot D3cold
[ 8.041967] xhci_hcd 0000:0d:00.0: Resetting
[ 9.062082] xhci_hcd 0000:0d:00.0: timed out waiting for pending transaction; performing function level reset anyway
[ 10.406071] xhci_hcd 0000:0d:00.0: not ready 1123ms after FLR; waiting
[ 11.462075] xhci_hcd 0000:0d:00.0: not ready 2147ms after FLR; waiting
[ 13.542072] xhci_hcd 0000:0d:00.0: not ready 4195ms after FLR; waiting
[ 17.894085] xhci_hcd 0000:0d:00.0: not ready 8291ms after FLR; waiting
[ 26.342086] xhci_hcd 0000:0d:00.0: not ready 16483ms after FLR; waiting
[ 44.006072] xhci_hcd 0000:0d:00.0: not ready 32867ms after FLR; waiting
[ 78.822072] xhci_hcd 0000:0d:00.0: not ready 65635ms after FLR; giving up
[ 79.222070] xhci_hcd 0000:0d:00.0: Refused to change power state, currently in D3
[ 79.272644] xhci_hcd 0000:0d:00.0: xHCI Host Controller
[ 79.272710] xhci_hcd 0000:0d:00.0: new USB bus registered, assigned bus number 7
[ 79.535236] xhci_hcd 0000:0d:00.0: Host halt failed, -19
[ 79.535291] xhci_hcd 0000:0d:00.0: can't setup: -19
[ 79.535347] xhci_hcd 0000:0d:00.0: USB bus 7 deregistered
[ 79.555578] xhci_hcd 0000:0d:00.0: init 0000:0d:00.0 fail, -19

It was lucky because it never worked again.
PCI card was not working but computer started.

Here you can find included lspci -vv, lshw and full dmesg taken from during this unique startup with the PCI Card.

I tried the same PCI Card with two computers and always had the same results.
Kernel is hanging with "xhci_hcd 0000:0d:00.0: init 0000:0d:00.0 fail, -19" and nothing more.

I tried this PCI Card with a CentOS 7.3 on same computer and kernel (3.10) is starting well and PCI card is fully working.

Many thanks for your assistance.
I apologize if it's an already reported bug.

Kind Regards

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is missing log files that will aid in diagnosing the problem. While running an Ubuntu kernel (not a mainline or third-party kernel) please enter the following command in a terminal window:

apport-collect 1820363

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Kai-Heng Feng (kaihengfeng)]

Since it works in v4.8 but not v4.9, please do a kernel bisection.

[Mike (michael-ubu-pan)]

What Firmware does Your card have?
There are many versions online and some of them are for Motherboards only (has different config for chips mounted on MB embedded chipset than for expansion card PCI/PCI_E).

[Fabian (moocan2112)]

Hello,

I can't produce more log because kernel hangs at startup. Can't start computer without the PCI card and insert when powered on.

@Mike, The card has firmware 2.0.2.6 which seems to the last version.
She was provided with 2.0.2.0 and tried too with 2.0.2.4.
I updated to the last version using Windows on an other computer which has PCI-X slots.
I tested this PCI card on a Gigabyte mainboard and an HP Proliant mainboard and each time kernel hangs with the same error "Host halt failed, -19".

If you have some links I can try to update firmware of the card with other releases.

I have too the PCIe version based on the same chip "Renesas Technology Corp. uPD720201" and kernel is not hanging with the PCIe version. It seems to be related to the PCI/PCI-X.

I never done a kernel bisection and will try to follow Ubuntu documentation that I found.

Kind Regards

[Mike (michael-ubu-pan)]

@Fabien

Some FWs over the net are for 2 port and some for 4-port. I had situation after Windows Application FW upgrade, that card was programmed wrongly as 2 channel 4 port instead of 1 channel 2 port card resulted in yellow Windows mark after every reboot. It was with 2.0.2.6 FW. Please rollback to 2.0.2.0 - it is confirmed the best stable for PCI-PCIX. NEC had lack of good programmers and newer FW was newer but not better/stable and brake other things. Beleve me I tried everything with upd720200(old gen v3) and upd7202202(new gen). The best for Windows is EJ168 - works very smooth under every version and until kernel v4.9 also. After that kernel masters also beleved that they can handle every USB 3.0 without problems, which results in boot hang.

[Fabian (moocan2112)]

@Mike

Many Thanks for your answer.
The PCI-X is a 4 ports with 2 channels.
The PCIe is 7 ports with an unknown number of channels.

The PCI-X was intended to be used with a HP Proliant for external Backup purpose which is running Ubuntu Server 18.04 LTS. No more PCIe slots are available because they are used by the Smart Array RAID Controller.

I will try to revert back this weekend to 2.0.2.0 using the last survival Windows.

But if I understand well what you mean I have no chance to get the PCI-X adapter working with any kernel with a version > 4.9 ?

I really appreciate your help.
Thank you Mike.

Kr

[Mike (michael-ubu-pan)]

@Fabien

I confirm from 4.10 broken support and fighting for repairing USB 3.0 cards on other Bug thread. I tried to bitsect kernel 4.9a to 4.10a but they are so different that it cannot be done by git automatically. Somene who redesigned 4.10 PCI USB 3.0 mess should take care of it. The hardware is ready to test.

I order You to use W201FW21.exe for updatng firmware to upd720201/upd720202 or W200FW35.exe for updatng firmware to upd720200/upd720200A. The integrated GUI all in one updater starting with Fxxxxxx.exe cannot downgrade firmware at all. You will find them on the net easily by the NEC firmware. The gods of linux flashrom package didn't want to take care of upd720200a datasheet and PCI/PCIE host to support FW flashing on linux.