[Nut-upsuser] Atlantis-Land UPS with Cypress CY7C63723 chip (cypress_m8 driver) support

[ngpost1]

Thank to Charles and Arnaud for your replies.

At this point, I am posting (with permission) my mail exchange with 
Lonnie Mendez, along with the lsusb output.


 Basically, as Lonnie put it, the chip works as a HID device (and I can 
confirm on that), so it should be supported under NUT.

 I am absolutely no expert on the matter of USB. The only thing I can do 
(and most happily!) is test whatever is available, since I have this UPS 
that desperately needs to be interfaced.

Thanks in advance for your attention and help.


--Max

============================================================================= 

Hello!

I have an UPS made by Ablerex, it's a Vesta 1500  (see  
http://www.uisabler.com.tw/HTM/product-lineinteractive.htm ) , rebranded 
as Atlantis-Land (model A03-S1501, see 
http://www.atlantis-land.com/prodotti.php?catID=22 ).

Inside the unit (I opened it) there is a Cypress CY7C63723-PC chip 
which, according the Documentation/usb/usb-serial.txt, should be 
supported just fine by your driver. Also, a search on the Cypress site 
confirms that the 7C63723 belongs to the CY4601 family.

Unfortunately, when I connect the UPS to the USB port (tried on several 
PC's and notebooks) I get a

drivers/usb/input/hid-core.c: couldn't find an input interrupt endpoint
usbhid: probe of 3-2:1.0 failed with error -5


error from dmesg. This guy over here has exactly the same problems I am 
experiencing:

http://www.mail-archive.com/linux-usb-users@lists.sourceforge.net/msg14649.html 


I understand if all goes well the UPS's port should show up as 
/dev/ttyUSB0 or something, but this is not my case. I tried  searching 
online without results. I also tried reading the driver source, but I am 
all but a kernel hacker and got lost pretty soon.

I am quite lost; moreover, I cannot understand if the problem is kernel 
related or driver related (therefore I have no idea where I should post 
for help). I am posting the output of dmesg below.

Thanks in advance for your attention and best regards.

P.S.: These come from a machine running Fedora Core 2 (old, I know); but 
I got the same results with the latest Mepis and Knoppix 3.9 as well.


This is from dmesg:

Sep 27 15:23:27 fc2 kernel: usb 1-2: new low speed USB device using 
uhci_hcd and address 3
Sep 27 15:23:27 fc2 kernel: drivers/usb/input/hid-core.c: couldn't find 
an input interrupt endpoint
Sep 27 15:23:27 fc2 kernel: usbhid: probe of 1-2:1.0 failed with error -5




This is an lsusb -v:   [please see attached file 
Atlantis-land_lsusb.txt.gz]



------------------------------------------------------------------------
On Tuesday 27 September 2005 8:33 am, you wrote:
> > Hello!
> > >    I have an UPS made by Ablerex, it's a Vesta 1500  (see  > 
> http://www.uisabler.com.tw/HTM/product-lineinteractive.htm ) , 
> rebranded > as Atlantis-Land (model A03-S1501, see > 
> http://www.atlantis-land.com/prodotti.php?catID=22 ).
> > >    Inside the unit (I opened it) there is a Cypress CY7C63723-PC 
> chip > which, according the Documentation/usb/usb-serial.txt, should 
> be > supported just fine by your driver. Also, a search on the Cypress 
> site > confirms that the 7C63723 belongs to the CY4601 family.
> >   

  Yes, the CY7C63723 is in the family.  However in this instance, the 
device seems
to go by the standards defined in the HID documentation at usb.org 
(notice the several
usage definitions given by the hid report descriptor).  So the 
cypress_m8 driver will most
definitely NOT work with this device (ie: it doesn't go by the CY4601 
specifications, which
needs 2 interrupt endpoints).


> >    Unfortunately, when I connect the UPS to the USB port (tried on > 
> several PC's and notebooks) I get a
> > > drivers/usb/input/hid-core.c: couldn't find an input interrupt 
> endpoint
> > usbhid: probe of 3-2:1.0 failed with error -5
> >
>   

  Sounds about right.  The hid-core needs at least an interrupt input 
endpoint to function.
This device seems to do all communication through endpoint zero, which 
would be
control trasactions using the HID specification for UPS devices at 
usb.org.  The device
on unix will most likely being doing coummication through NUT, which 
would use libusb.
See further down.

> > >    error from dmesg. This guy over here has exactly the same 
> problems I > am experiencing:
> > > 
> http://www.mail-archive.com/linux-usb-users@lists.sourceforge.net/msg14649.html 
>
> > >    I understand if all goes well the UPS's port should show up as 
> > /dev/ttyUSB0 or something, but this is not my case. I tried  
> searching > online without results. I also tried reading the driver 
> source, but I am > all but a kernel hacker and got lost pretty soon.
> > >    I am quite lost; moreover, I cannot understand if the problem 
> is > kernel related or driver related (therefore I have no idea where 
> I > should post for help). I am posting the output of dmesg below.
>   

  I'd email the developers over at networkupstools.org (the guys that 
made NUT) the entire lsusb -v output you gave me
from the BusID all the way down to the last usage in the hid report 
descriptor.  They can most likely
use this information to add support for this device, as it seems to go 
by some standards... at least under
the usb.org HID specification for UPS devices.  However, the idVendor 
and idProduct values looks very funny.
0001 and 0000.  idVendor might be invalid (ie: they didn't want to pay 
usb.org for a valid number id) and you could potentially
run into conflicts with devices which are valid for idVendor being 0001.

 -Lonnie


-------------- next part --------------
A non-text attachment was scrubbed...
Name: Atlantis-land_lsusb.txt.gz
Type: application/gzip
Size: 3175 bytes
Desc: not available
Url : http://lists.alioth.debian.org/pipermail/nut-upsuser/attachments/20051209/e5b836fd/Atlantis-land_lsusb.txt.bin