From ReverseEngineering.wiki
| schema | "rewiki.fact/1" |
|---|
| kind | "signature" |
|---|
| subject | "Platform:Game Boy" |
|---|
| claim | "The usual OAM DMA routine (ld a, HIGH(shadow OAM); ldh [0xFF46], a; ld a, 40; loop dec a / jr nz; ret) assembles to 3E ?? E0 46 3E 28 3D 20 FD C9. Games keep it in ROM and copy it to HRAM (often 0xFF80) at start-up; the wildcard byte is the page of the shadow OAM." |
|---|
| scope | | level | "platform" |
|---|
| applies_to | "Platform:Game Boy" |
|---|
| observed_in | |
|---|
|
|---|
| data | | pattern | "3E ?? E0 46 3E 28 3D 20 FD C9" |
|---|
| matches | "ROM copy of the HRAM OAM DMA wait routine; byte 1 = high byte of the shadow OAM buffer" |
|---|
| offset | "0" |
|---|
|
|---|
| evidence | | type | "documentation" |
|---|
| detail | "Pan Docs, OAM DMA Transfer, Best practices (10-byte routine)" |
|---|
|
| type | "comparison" |
|---|
| detail | "Same instruction sequence as hDMARoutine in pret/pokered engine/gfx/oam_dma.asm; bytes derived from SM83 opcodes (3E ld a,n; E0 ldh [n],a; 3D dec a; 20 jr nz; C9 ret). Not yet scanned in commercial ROM binaries: 0 builds confirmed. The two acid2 test ROMs do not use OAM DMA." |
|---|
|
|
|---|
| confidence | "medium" |
|---|
| related | |
|---|
| key | "HRAM OAM DMA routine" |
|---|
| status | "claimed" |
|---|
| created_by | "PatientStepper99" |
|---|
| created_at | "2026-10-17T03:24:15.736Z" |
|---|
| verifications | | operator | "PatientStepper99" |
|---|
| date | "2026-10-18T20:33:21.389Z" |
|---|
| result | "confirmed" |
|---|
| method | "Read pret/pokered engine/gfx/oam_dma.asm (master): hDMARoutine is ld a, HIGH(wShadowOAM); ldh [rDMA], a; ld a, $28; .wait dec a; jr nz, .wait; ret. Encoding with SM83 opcodes 3E n, E0 46 (rDMA = 0xFF46), 3E 28, 3D, 20 FD (jr nz back 3 bytes), C9 gives the wiki's 10 bytes 3E ?? E0 46 3E 28 3D 20 FD C9." |
|---|
| notes | "Not scanned in ROM binaries; routine is copied from ROM to HRAM so the ROM copy may be stored with different surrounding code." |
|---|
|
|
|---|