Hi,

I've just been asked to check out an SDP problem, which occurred with
a Motorola A1000 phone.  It seems like the phone doesn't return any
Service Class ID's for its Handsfree Audio Gateway profile.

This turns out to be caused by sdp_get_service_classes() failing.  I
added some debug code to this function (see sdp-dtd-debug.patch, and
compile with -DSDP_DEBUG), to help diagnose the problem.

For some reason, spd_get_service_classes thinks there's a SDP_UINT16
(0x09) instead of a SDP_UUID16 (0x19) in the Service Class ID list of
the HF AG record, but here hcidump and sdptool disagree.

sdptool says (see sdptool-arm for full dump):

Attribute Identifier : 0x0 - ServiceRecordHandle
  Integer : 0x10003
Attribute Identifier : 0x1 - ServiceClassIDList
  Data Sequence
    Integer : 0x656e
    Integer : 0x6a
    Integer : 0x100

while hcidump says (see hcidump-arm.raw):

          record #4
              aid 0x0000 (SrvRecHndl)
                 uint 0x10003
              aid 0x0001 (SrvClassIDList)
                 < uuid-16 0x111f (Handsfree AG) uuid-16 0x1203 (Audio) >

However, this is all on an ARM platform, while on an x86 platform it
DOES seem to work correctly!

sdptool says on x86 (see sdptool-x86):

Attribute Identifier : 0x0 - ServiceRecordHandle
  Integer : 0x10003
Attribute Identifier : 0x1 - ServiceClassIDList
  Data Sequence
    UUID16 : 0x111f - HandsfreeAudioGateway
    UUID16 : 0x1203 - GenericAudio

and hcidump says (see hcidump-x86.raw):

          record #3
              aid 0x0000 (SrvRecHndl)
                 uint 0x10003
              aid 0x0001 (SrvClassIDList)
                 < uuid-16 0x111f (Handsfree AG) uuid-16 0x1203 (Audio) >

So it seems to be another nice alignment problem somewhere in the SDP
library stuff! (Hi Marcel :P)  If anyone has an idea where this might
be, please let me know.