Re: [PATCH] usb: xhci: handle uPD720201 and uPD720202 w/o ROM

[Christian Lamparter]

Hello,

On Tuesday, June 21, 2016 05:56:58 AM Yoshihiro Shimoda wrote:
> > From: Christian Lamparter
> > Sent: Tuesday, June 21, 2016 12:32 AM
> > 
> > On Wednesday, June 08, 2016 12:14:57 AM Christian Lamparter wrote:
> > > This patch adds a firmware check for the uPD720201K8-711-BAC-A
> > > and uPD720202K8-711-BAA-A variant. Both of these chips are listed
> > > in Renesas' R19UH0078EJ0500 Rev.5.00 "User's Manual: Hardware" as
> > > devices which need a firmware in order to work as they do not have
> > > support to load the firmware from an external ROM.
> > >
> > > Currently, the xhci-pci driver is unable to initialize the hcd in
> > > this case. Instead it will wait for 30 seconds and cause a timeout
> > > in xhci_handshake() and fails.
> > >
> > > [    5.116990] xhci_hcd 0000:45:00.0: new USB bus registered ...
> > > [   32.335215] xhci_hcd 0000:45:00.0: can't setup: -110
> > > [   32.340179] xhci_hcd 0000:45:00.0: USB bus 2 deregistered
> > > [   32.345587] xhci_hcd 0000:45:00.0: init 0000:45:00.0 fail, -110
> > > [   32.351496] xhci_hcd: probe of 0000:45:00.0 failed with error -110
> > >
> > > Cc: Yoshihiro Shimoda <yoshihiro.shimoda.uh@xxxxxxxxxxx>
> > > Signed-off-by: Christian Lamparter <chunkeey@xxxxxxxxx>
> > Hello?
> > 
> > Are there any news on this? Or is there anything else that I'm missing
> > which blocks this patch? If it's because the device won't be working
> > either with this patch, then please let me know.
> 
> Thank you for the patch with CC me.
> However, I'm afraid but I don't know the detail of the Renesas xHCI PCI controller.
> (My job is for R-Car environment for now.)
Oh. I see. Still, I would like to thank you for your input. I looked at
the rcar loader and firmware: The firmware header (aa55 + firmware version
pointer at 0x14) are carbon-copies of the pci parts (or vice versa?). I guess
I shouldn't be surprised, I bet the rcar and uPD both have a similar SuperH.
The firmware upload mechanism itself seems to be a bit different though, the
spec for the pci devices lists much more stuff to check and verify. Also, the
pci variant uploades 2 x 32bit packages at a time (via data0 and data1 in the
pci config space). However, the procedure is still somewhat identical.

Anyway, if you want to take a look: I went ahead and wrote a firmware loader
for the uPD720201 and uPD720202: I have attached in the patch below. The device
now works for this APM82181 (PowerPC 464 - Big Endian).

xhci_hcd 0000:45:00.0: xHCI Host Controller
xhci_hcd 0000:45:00.0: new USB bus registered, assigned bus number 2
xhci_hcd 0000:45:00.0: hcc params 0x014051cf hci version 0x100 quirks 0x01000090
xhci_hcd 0000:45:00.0: xHCI Host Controller
xhci_hcd 0000:45:00.0: new USB bus registered, assigned bus number 3
usb usb3: We don't know the algorithms for LPM for this host, disabling LPM.
hub 3-0:1.0: USB hub found
hub 3-0:1.0: 2 ports detected
usb 3-1: new SuperSpeed USB device number 2 using xhci_hcd
usb-storage 3-1:1.0: USB Mass Storage device detected
scsi host1: usb-storage 3-1:1.0
scsi 1:0:0:0: Direct-Access     Intenso  Slim Line        PMAP PQ: 0 ANSI: 6
sd 1:0:0:0: [sdb] 15466496 512-byte logical blocks: (7.92 GB/7.38 GiB)
...

> By the way, the issue seems similar with R-Car environment though :)
> http://thread.gmane.org/gmane.linux.kernel.stable/175457/focus=140699
> 
> and fix patch for it:
> http://thread.gmane.org/gmane.linux.kernel.stable/177524
Yes :). Same thing here. Load the firmware and the device is happy.
I know this is a big ask: Do you know or can you give me a lead
to whom would be willing to help/support the uPD720201/2? It would
be great if the firmware could be added to linux-firmware.git.

Note: This is one way to do it. But there are many more... I think adding
a pci_enable quirk to drivers/usb/host/pci-quirk.c is the least invasive
way... unless of course, someone knows a even better method... So please:
let me know!

Regards,
Christian
---