git-etawen-dev-earthonion/garlic-worker
Watch
1
0
Fork
You've already forked garlic-worker
 0
Code Issues Pull requests Releases 1 Activity

ps5: handle PS4 saves natively (single PS5 takes both queues) #1

Merged
earthonion merged 0 commits from ps5-handle-ps4-saves into main 2026-05-13 23:12:56 -05:00
Conversation 0 Commits - Files changed -
earthonion commented 2026-05-13 22:46:21 -05:00 (Migrated from git.etawen.dev)

Summary

Makes the PS5 worker process both PS5 and PS4 jobs natively. A single jailbroken PS5 can now drain both garlicsaves.com queues at once.

Why this works

garlic-savemgr already proves the PS5 can decrypt PS4 sealed keys via the same /dev/pfsmgr ioctl 0xc0845302 the PS5 worker uses. The only difference is where the sealed key lives:

  • PS5 save image → pfsSKKey magic embedded at offset 0x800 of the image
  • PS4 save image → 0x60 sealed-key bytes in a companion sdimg_<save>.bin file

After decryption, sceFsMountSaveData() mounts both formats identically.

Changes

src/ps5/savedata.c (+163 / -50)

  • detect_is_ps4() reads byte 0 of the save image (0x01=PS4, 0x02=PS5)
  • save_mount() branches: PS5 path is unchanged; PS4 path copies the companion .bin and loads the 0x60 sealed-key bytes
  • Same ioctl 0xc0845302 + sceFsMountSaveData() flow afterward

src/ps5/worker.c (+46 / -7)

  • Poll loop alternates ?platform=ps5 and ?platform=ps4 each tick; sleeps only after both return empty
  • After a job is found, resets to ps5-first so PS5 queue keeps priority
  • Dispatch declines PS4 encrypt / createsave (PS5 worker only produces PS5-format PFS images via sceFsCreatePprPfsSaveDataImage) — HTOS keeps those queued for a real PS4 worker
  • PS4 decrypt / resign / reregion / keyset all work via the new mount path

No server-side changes

HTOS-web /api/worker/next already accepts ?platform=ps4 and filters by params.platform (see routes/api.py:190). No server patch required.

Test plan

  • Build: make ps5 → garlic-worker-ps5.elf (verified locally, 172KB, no warnings)
  • Run on real PS5 (on supported firmware via P2JB)
  • Pull a PS5 decrypt job → confirm still works
  • Pull a PS4 decrypt job → confirm sdimg_* + .bin mount succeeds
  • Pull a PS4 resign job → confirm sealed-key roundtrip
  • Pull a PS4 encrypt/createsave job → confirm worker declines and HTOS re-queues
  • Monitor last_platform in worker_keys — should flip between ps4/ps5 based on whichever queue had the job

Followups (not in this PR)

  • PS4 encrypt / createsave support on PS5 (would need PS4 PFS creation, which garlic-savemgr does not do either)
  • Same dual-poll behavior in TCP direct-connect mode (currently TCP auth advertises only ps5)
## Summary Makes the PS5 worker process **both PS5 and PS4 jobs** natively. A single jailbroken PS5 can now drain both garlicsaves.com queues at once. ## Why this works `garlic-savemgr` already proves the PS5 can decrypt PS4 sealed keys via the same `/dev/pfsmgr` ioctl `0xc0845302` the PS5 worker uses. The only difference is *where the sealed key lives*: - PS5 save image → `pfsSKKey` magic embedded at offset `0x800` of the image - PS4 save image → 0x60 sealed-key bytes in a companion `sdimg_<save>.bin` file After decryption, `sceFsMountSaveData()` mounts both formats identically. ## Changes ### `src/ps5/savedata.c` (+163 / -50) - `detect_is_ps4()` reads byte 0 of the save image (`0x01`=PS4, `0x02`=PS5) - `save_mount()` branches: PS5 path is unchanged; PS4 path copies the companion `.bin` and loads the 0x60 sealed-key bytes - Same `ioctl 0xc0845302` + `sceFsMountSaveData()` flow afterward ### `src/ps5/worker.c` (+46 / -7) - Poll loop alternates `?platform=ps5` and `?platform=ps4` each tick; sleeps only after both return empty - After a job is found, resets to ps5-first so PS5 queue keeps priority - Dispatch declines PS4 `encrypt` / `createsave` (PS5 worker only produces PS5-format PFS images via `sceFsCreatePprPfsSaveDataImage`) — HTOS keeps those queued for a real PS4 worker - PS4 `decrypt` / `resign` / `reregion` / `keyset` all work via the new mount path ## No server-side changes HTOS-web `/api/worker/next` already accepts `?platform=ps4` and filters by `params.platform` (see `routes/api.py:190`). No server patch required. ## Test plan - [ ] Build: `make ps5` → `garlic-worker-ps5.elf` (verified locally, 172KB, no warnings) - [ ] Run on real PS5 (on supported firmware via P2JB) - [ ] Pull a PS5 decrypt job → confirm still works - [ ] Pull a PS4 decrypt job → confirm `sdimg_*` + `.bin` mount succeeds - [ ] Pull a PS4 resign job → confirm sealed-key roundtrip - [ ] Pull a PS4 encrypt/createsave job → confirm worker declines and HTOS re-queues - [ ] Monitor `last_platform` in `worker_keys` — should flip between `ps4`/`ps5` based on whichever queue had the job ## Followups (not in this PR) - PS4 `encrypt` / `createsave` support on PS5 (would need PS4 PFS creation, which `garlic-savemgr` does not do either) - Same dual-poll behavior in TCP direct-connect mode (currently TCP `auth` advertises only ps5)
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels
No items
No labels
Milestone
Clear milestone
No items
No milestone
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
git-etawen-dev-earthonion/garlic-worker!1
No description provided.