On Tue, 15 Nov 2011 11:59:12 +0200
Luiz Augusto von Dentz <luiz.dentz@xxxxxxxxx> wrote:
> Hi Antonio,
>
> On Tue, Nov 15, 2011 at 1:37 AM, Antonio Ospite
> <ospite@xxxxxxxxxxxxxxxxx> wrote:
> > Hi,
> >
> > with master (a267bc2) I am getting this in dmesg for Sixaxis
> > controllers:
> >
> > input: PLAYSTATION(R)3 Controller as /devices/pci0000:00/0000:00:02.0/...
> > generic-bluetooth 0005:0000:0000.0018: input,hidraw1: BLUETOOTH HID v0.00 ...
> > ^^^^^^^^^^^^^^^^^ ^^^^ ^^^^
> >
> > generic driver is used because vendor_id and product_id are 0000!
> > while I would expect this:
> >
> > input: PLAYSTATION(R)3 Controller as /devices/pci0000:00/0000:00:02.0/...
> > sony 0005:054C:0268.0017: input,hidraw1: BLUETOOTH HID v0.00 ...
> > ^^^^ ^^^^ ^^^^
> >
> > Having vendor_id and product_id set to 0000 was making the device not
> > working, the story is briefly like that:
> > - the hid-sony kernel driver makes the controller operational;
> > - devices are bound to hid drivers using vendor_id and product_id;
> > - with those set to 0000, hid-sony ends up not being used at all.
>
> That is probably a bug in the code, does the device properties got the
> right values?
>
Yes they do right after read_device_id_from_did(), I have investigated
a little more.
The 'did' file for the device contains:
00:1B:FB:XX:XX:XX FFFF 054C 0268 0000
Only now I realize that the source field (the one now set to FFFF)
is supposed to be either 0x0001 ore 0x0002 like said in [1, section
5.6] about VendorIDSource, so I was doing it wrong. Setting the value to
0x0002 (as the vendor is taken from the USB device) make the device
working.
But _maybe_ this error on my side exposed something inaccurate in BLueZ
too, please keep reading.
This is what was happening in device_cerate(): read_device_id() failed
and this block was never called.
if (read_device_id(srcaddr, address, NULL, &vendor, &product,
&version) == 0) {
device_set_vendor(device, vendor);
device_set_product(device, product);
device_set_version(device, version);
}
However, the values were taken correctly in read_device_id_from_did(),
but this snippet makes read_device_id() return with an error
regardeless:
err = read_device_id_from_did(srcaddr, dstaddr, &lsource,
vendor, product,version);
if (!err) {
if (lsource == 0xffff)
err = -ENOENT;
return err;
}
Now, the reference to the special value FFFF I can find in [1] is
about VendorID [1, section 5.2], NOT VendorIDSource, so my wrong value
in the 'did' file might have exposed a spec misinterpretation here.
Should the check be:
if (*vendor == 0xffff)
err = -ENOENT;
? Changing the other assignment of this value in the code too.
And what about issuing a warning if the 'source' field stored by
store_device_id() is not 0x0001 or 0x0002?
Thanks,
Antonio
[1]
https://developer.bluetooth.org/KnowledgeCenter/TechnologyOverview/Documents/DeviceID_SPEC.pdf
--
Antonio Ospite
http://ao2.it
PGP public key ID: 0x4553B001
A: Because it messes up the order in which people normally read text.
See http://en.wikipedia.org/wiki/Posting_style
Q: Why is top-posting such a bad thing?
Attachment:
pgpVisZVsGOIS.pgp
Description: PGP signature
