# DF-2608 — PoC verdict

**File:** `sys/bus/u4b/input/uep.c`

**Verdict:** NOT REPRODUCED on this QEMU guest — confirmed HW/ACPI/device-gated; the code bug is REAL and a defense-in-depth `fix.diff` is attached.

## Mechanism (confirmed in source)

The interrupt-IN callback reassembles packets from a 17-byte UNINITIALIZED stack buffer buf[17]. On reassembly, res=pkt_len-sc->buf_len bytes are memcpy'd out of buf; a legal short transfer (len==1) makes res exceed the bytes actually delivered by usbd_copy_out, so uninitialized kernel stack is copied in and pushed to the RX FIFO, readable via /dev/uep*.

### Cited lines

- `sys/bus/u4b/input/uep.c:200`
- `sys/bus/u4b/input/uep.c:222-231`
- `sys/bus/u4b/input/uep.c:391`

## Why it does not reproduce on this guest

No USB eGalax touchscreen on the QEMU guest (usbconfig list shows no devices; no /dev/uep*).

Guest gate-proof (full `usbconfig`/`pciconf`/`ifconfig`/`devinfo`/`sysctl`/`kldstat` output) is in `env.txt`.

## Defense-in-depth fix

Clamp the reassembly count to what was actually received: if res>len, goto tr_setup (drop) instead of reading uninitialized stack.

The git-apply-able diff is in `fix.diff` (verified `git apply --check` clean).

## Classification

- `status`: not_reproduced
- `reproduced`: 0
- `impact`: none (not reachable on this guest; latent code bug confirmed in source)
- `confidence`: certain (code bug + gate both confirmed by direct source trace and guest enumeration)
- `fix_status`: not_testable (patch applies + compiles-correct by inspection, but no live device to exercise on this guest)