Some USB devices uses several report IDs

[gyurco]

See

https://www.atari-forum.com/viewtopic.php?p=470071#p470071

[gyurco]

Meanwhile I discovered in vebxenon's mouse that several report id's have report_size, but not report_count, so the last value is re-used, but not what's defined in the previous, skipped report:

HIDP: REPORT_ID(18)
HIDP:   - total bit count: 40 (5 bytes, 0 bits)
HIDP:   - report 16 is usable
HIDP: USAGE(72/0x48)
HIDP:  -> UNSUPPORTED USAGE
HIDP: REPORT_COUNT(1)
HIDP: REPORT_SIZE(2)
HIDP: LOGICAL_MINIMUM(0/0)
HIDP: LOGICAL_MAXIMUM(1)
HIDP: PHYSICAL_MINIMUM(1/1)
HIDP: PHYSICAL_MAXIMUM(1)
HIDP: FEATURE(2)

HIDP: REPORT_ID(16)
HIDP:   - total bit count: 40 (5 bytes, 0 bits)
HIDP:   - report 16 is usable
HIDP: USAGE(56/0x38)
HIDP:  -> axis usage
HIDP: MOUSE: found wheel @ 1
HIDP: PHYSICAL_MINIMUM(0/0)
HIDP: PHYSICAL_MAXIMUM(0)
HIDP: LOGICAL_MINIMUM(129/129)
HIDP: LOGICAL_MAXIMUM(127)
HIDP: REPORT_SIZE(8)
HIDP:   (Z-AXIS @ 48 (byte 6, bit 0))
HIDP: INPUT(6)
HIDP: END_COLLECTION(0)
HIDP:   -> phys/log end
HIDP: COLLECTION(2)
HIDP:   -> logical
HIDP: REPORT_ID(18)
HIDP:   - total bit count: 56 (7 bytes, 0 bits)

Here bit count incremented by 2x8, however I guess it should be 1x8 only, so you're right about not skipping global items. No need to worry about min-max values, they're defined if changed.
That way even my patch ends up at 56 bits.
Need to dig through that HID specs...

[gyurco]

Also noticed the usage_count difference with the wheel. Probably your version is the good one, as 'found wheel @ 0' looks right instead of '@ 1'.