Re: [Libusb-devel] Getting firmware from a device/driver

[elvez <el...@gm...>]

ti...@pr... wrote:
> On Thu, Nov 23, 2006 at 10:44:15AM +0100, Luca wrote:
>   
>> On 11/22/06, Tim Roberts <ti...@pr...> wrote:
>>     
>>> I guess that depends on how you define "totally different".  EZUSB was
>>> the first generation, created by AnchorChips, which was bought by
>>> Cypress.  It does USB 1.1 only.  The FX2 was the second generation.  It
>>> handles USB 2.0, has a faster processor, and uses less power.  They both
>>> have an 8051, they have a similar philosophy, and the command sets are
>>> similar, but not identical.
>>>       
>> OK, but what troubles me is why I need to specify the "-t fx2" option
>> to download most of the firmware, even though the device is an ez-usb?
>>     
>
> Don't know.  Can you pop the top and check the label on the chip?  The
> descriptors you posted were clearly for USB 1.1.  It's possible to 
> configure an FX2 to be USB 1.1, but it would be unusual, unless you 
> are plugging this into a USB 1.1 hub.
>   
Yes, my PC is not USB 2, but if you say that CY7C64713 is USB 1.1, too, 
I think it's improbable for an older model to
be USB 2. Anyway, I'll try to check the chip inside the device, just to 
make sure.
>   
>> I tried downloading again the (partial) firmware via fxload, rebooted,
>>     
>
> You shouldn't have to reboot.  The last step of the download process is
> to reset the 8051, which forces the re-enumeration.
>   
I know, it was just one of the latest desperate attempts I've made...
>   
>> Today, I spoke with a developer from the italian division of the
>> company developing the device, and although he didn't know how the
>> internals of the device works, he said he'll try to send me some
>> helpful material.
>>
>> In the meantime I'll try to direct my efforts over a newer version of
>> the device which is based on the CY7C64713 processor. Do you think
>> it'll be more difficult to work on this device?
>>     
>
> CY7C64713 is the chip number of the EZUSB FX1.  It's still USB 1.1.
>
>   

Wow, that's not a good news!!! So I guess problems I'll face will be the 
same with the newer device. In fact, there is a difference: to interface 
with the CY7C64713-based device in Windows, I use the so-called 
CyAPI.h/CyAPI.lib combination. Do you think this additional layer will 
add to the problems, or is it just an optional layer to let programmers 
not worry much about low-level errors and unexpected events?