aboutsummaryrefslogtreecommitdiffstats
Commit message (Expand)AuthorAge
...
* fuzz_asd: stop neutering asd_validate_ptr; put buffers where it accepts themuser2026-09-22
* fuzz_sos_ring: add a coverage gate, calibrated against a positive controluser2026-09-22
* Record that the repaired fuzz_sysdrv rediscovers a known latent defectuser2026-09-22
* fuzz_sysdrv: map the fabric aperture so benign register reads stop being find...user2026-09-22
* fuzz_sysdrv: actually run SYS_DRV init; it was fuzzing a closed gateuser2026-09-22
* ASD has no init phase; drop the redundant asd_feature_enabled patchuser2026-09-22
* Run the TA's real init and real session open instead of patching globalsuser2026-09-22
* Add a validated checker for un-propagated retractionsuser2026-09-22
* Fix fuzz_pss_differential's calling convention; parser is bounded and stricteruser2026-09-22
* Wire the Bus: HDCP's ASD RPCs now run in the real ASD dispatcheruser2026-09-22
* Record two failed attempts at an explicit ASD load status codeuser2026-09-22
* ASD loads, verifies and runs (harness/load_asd.py)user2026-09-22
* Measured: the memref subtree is dead code in this configurationuser2026-09-22
* RETRACTION: "management ring" naming was wrong; call graph redone properlyuser2026-09-22
* Ring-1 hypothesis tested, not confirmed; mgmt dispatcher still unreacheduser2026-09-22
* Found it: the harness has been driving the wrong dispatcheruser2026-09-22
* Refute my own next step: ring cmd 5 does not unlock the memref paramuser2026-09-22
* Ring TA-command path cannot satisfy the HDCP gate at all (structural)user2026-09-22
* HDCP invoke ABI decoded: command lives in the shared buffer, not the ringuser2026-09-22
* fuzz_ta_invoke: reached 0/80 TA invocations; make it say souser2026-09-22
* Stop the boot harnesses running long after the work is doneuser2026-09-22
* Track the firmware images and keep the symbol maps alongside themPeter Geissler2026-09-22
* Initial import: AMD PSP reverse engineering, emulator and vulnerability researchPeter Geissler2026-09-22