Showing posts with label NEC. Show all posts
Showing posts with label NEC. Show all posts

2011-10-03

Flashing a NEC/RENESAS USB 3.0 uPD720200/uPD720200A firmware

It looks like NEC/Renesas are releasing steady releases of their popular USB 3.0 controllers drivers and firmwares. In case you have such a controller, you probably want to check the following page for the latest version.

One of the problems I found with the latest firmware update (4015) is that the GUI Renesas upgrade utility just didn't work for my controller, on Windows 7 x64. Even worse, it created commandline processes ad-infinitum, which, apart from obviously clogging up the system, is the last thing you want to see, as repeated erase/flash cycles can really wear out a flash chip.

The good thing however is that the firmware upgrade comes with a commandline flashing utility (W200FW35.exe), so you might as well use that, and flash the firmware manually. Having a commandline utility may also help with a flashrom implementation in the future, since it should be fairly straightforward to reverse engineer. The one thing you have to be careful of however, when using W200FW35.exe, is that you need to select the right SPI-Flash-ROM device type for flashing, and that the included batch files, which are set to a specific type, may not match the one you have.

Issuing W200FW35.exe /? will provide more information. First, in an elevated command prompt, start by issuing /srom ? to identify your chip and check that the flashing utility can access it:

C:\uPD720200_uPD720200A_FW_Updater>W200FW35.exe /srom ?
Bus:0x04 Device:0x00 Function:0x00
This Device is uPD720200A(Revision 4).
    W25X10BV/20BV/40BV(WINBOND)/EN25F05/10/20/40/(EON)/A25L512/010/020/040(AMIC) Type : 3, PageSize = 0x100, Chip Erase = 0xC7
Then, if you have a uPD720200A based controller, you probably want to issue:
W200FW35.exe /srom 0 /dump backup.mem
W200FW35.exe /srom 0 /write F401502.MEM cfg.ini
/srom 0 above ensures that the flash type will be autoselected. Or, in this case, we could have used /srom 3. If you use an srom parameter other than 0, be mindful that using the wrong type will result in both garbage in and garbage out.
As digitaldiatribe also points out, if you have more than one Renesas chip on your system, you will also need to add the /address switch to select the right one. Update: As pointed out by Eluder, the latest firmwares from Renesas no longer come with the W200FW35.exe commandline utility. It can still be obtained from the "3.0.2.8.0.8 (uPD720200) & 4.0.1.5.0.2 (uPD720200a)" archive available on the link provided above.

2011-08-12

Which controllers support USB 3.0 Debug?

As defined per section 7.6 of the xHCI specifications?

Unfortunately, none of the current ones from Renesas, VIA Labs or Fresco Logic do, and neither of these manufacturers are planning to add the capability on their existing controllers, through a firmware update for instance... Instead, they only have plans to provide it on their next generation controllers, which won't be available for some time (yes, I am aware that the new Renesas chips have been announced since 2011.03, but I have yet to see a PCI-E expansion card with one of those -- If you actually managed to get your hands on one, I'd like to hear from you!  
UPDATE 2011.08.29: Apparently those guys did... but as a world exclusive, since these cards won't be available till later on this year. Darn!).
UPDATE 2012.01.10: At long friggin' last, uPD720201 based PCIE cards, with debug support, are starting to become available, one such being the Buffalo IFC-PCIE4U3S. Only seems to be available in Asia and Australia at the moment however, but even then, orders may not be satisfied before February because of lack of availability...

So this means, if you have a NEC/Renesas uPD720200/uPD720200A, a VIA Labs VL800/801 or a Fresco Logic FL1000/FL1009 based USB 3.0 controller, you're out of luck with regards to USB 3.0 debug.
This doesn't bode too well for USB 3.0 as replacement for legacy RS232 ports...
Oh and the other disappointing part of the current crop of PCIE-USB3 controllers is they don't support USB-3.0 boot either. Quite the letdown...

Hint: If you want to display the Extended Capabilities of your controller, you can do so by modifying the xhci_setup_port_arrays() of drivers/usb/host/xhci-mem.c in () on a recent Linux kernel, as it already has some code to scan the extended caps. Using this method, I was able to find that, as of firmware 4015, the only Extended Capabilities provided by an uPD720200 USB 3.0 controller was "USB Legacy Support" (xHCI specs chpater 7.1) and "xHCI Supported Protocol" (7.2). The same applies for a Fresco Logic FL1000 based controller that I just got my hands on, though this one also sports vendor specific Extended Capabilities, with IDs 192 and 193.

Below are the answers from various manufacturers regarding their planned support of USB 3.0 debug. Being mostly interested in addon cards for existing systems, I haven't asked Intel or AMD.

NEC/Renesas:
"The uPD720200 does not support the debug port capability, and there is no plan to add it. But we have already added debug port support to our newer USB3 host controllers, uPD720201 and uPD720202."
VIA Labs:
"VL800/801 doesn't support debug port capability. As I know, the only one support is Renesas 3rd gen. host controller UPD720201_202, we plan to add this feature with VL805 that will MP next year."
Fresco Logic:
"FL1000/FL1009 didn't support USB debug extended capability.
We are doing it (for the next generation of controllers) and the schedule will be 2012/Q1."