Files
Sylpheed/tools/ppc-manual/memory/dcbz.md
sim 7f8a81b8f8
All checks were successful
CI / Native — linux (pull_request) Successful in 2h2m37s
CI / WASM — Web (pull_request) Successful in 30m13s
CI / Formatting (pull_request) Successful in 1m23s
docs(ppc-manual): quote Canary and our own decoder, not the retired xenia-rs
The generator had not been able to run correctly since the manual moved into
`tools/ppc-manual/`: it computed the repository root as `HERE.parent.parent`,
which now names `tools/`, so the XML, Canary's emitters and xenia-rs all stopped
resolving — silently, because both scrapers skipped what they could not find.
Every page's references had been pointing at paths that exist nowhere.

What each source contributed, measured on the 350 pages before this change:

  Operation (pseudocode)  251 pages: fixed boilerplate "derives from the xenia-rs
                          interpreter"; 99 carry real hand-written seeds
  C translation           337 pages: the same kind of boilerplate
  xenia-rs snapshot       336 pages: the interpreter arm, pasted in — the only
                          per-instruction semantics on unseeded pages
  links                   xenia-rs opcode/decoder/interpreter + Canary emitter

Now:

  * semantics come from **Xenia Canary**, the reference emulator, read through
    `git show` at a pinned upstream commit (`origin/canary_experimental`,
    f21ebd49e9). Not our checkout: it carries instrumentation and lacked
    upstream's `mcrf` fix, so it would have published probes and a wrong `mcrf`.
    Each page embeds the emitter (`InstrEmit_<mnem>`), and for the 128 pure
    one-line delegations also the helper that holds the semantics.
  * decode references point at `crates/sylpheed-ppc` — the decoder that
    produces `sylpheed.db` — as in-repo relative links.
  * the boilerplate now says what is true, and the C translation guide maps
    Canary's actual HIR calls, checked against `ppc_hir_builder.h` (including
    that `UpdateCR(n, v)` truncates to 32 bits).
  * `rust_scraper.py` -> `decoder_scraper.py` (interpreter half dropped);
    missing sources are now errors, not empty results.

Verified:

  consistency checks        455 XML entries, 350 families, 598 index keys
  hand-written tails        386/386 byte-identical after regeneration
  xenia-rs in generated     0
  pages with a snapshot     349/350 (was 336) — `dcbi` has no Canary emitter at all
  in-repo decoder links     910/910 resolve to a line holding the identifier
  emitter boundaries        brace counter == column-0 `}` rule on 521/521;
                            preprocessor model unit-tested (#if 0/#else/#elif)
  idempotency               re-run: 0 pages updated, 0 working-tree changes

Hand-written notes (outside the generated regions) are not rewritten here:

  * 110 links into `../../xenia-rs/...` were dead; they now point at the file in
    the archived repository (git.mc02.dev/fabi/xenia-rs @ 8401d4d). Line anchors
    were dropped because the notes predate that commit — 0 of 441 old line
    ranges match it — and a precise-looking wrong anchor is worse than none. The
    link text, which carries the author's line numbers, is unchanged.
  * 140 prose claims about xenia-rs's behaviour remain. 23 are verified to hold
    for Canary too (the 32-bit CR0 truncation, OE left unimplemented); the other
    114 need checking one by one, and some invert — e.g. `divdx` notes a correct
    64-bit CR0 update in xenia-rs where Canary's `UpdateCR` truncates. Left for
    a deliberate pass rather than a blind substitution.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-16 20:34:34 +02:00

8.2 KiB
Raw Blame History

dcbz — Data Cache Block Clear to Zero

Category: Memory · Form: DCBZ · Opcode: 0x7c0007ec

Assembler Mnemonics

Mnemonic XML entry Flags Description
dcbz dcbz — Data Cache Block Clear to Zero
dcbz128 dcbz128 — Data Cache Block Clear to Zero 128

Syntax

dcbz [RA0], [RB]
dcbz128 [RA0], [RB]

Encoding

dcbz — form DCBZ

  • Opcode word: 0x7c0007ec
  • Primary opcode (bits 0–5): 31
  • Extended opcode: 1014
  • Synchronising: no
Bits Field Meaning
0–5 OPCD primary opcode (31)
6–10 — reserved
11–15 RA base register (0 ⇒ literal 0)
16–20 RB offset register
21–30 XO extended opcode (1014 for dcbz / 1010 for dcbz128)
31 — reserved

dcbz128 — form DCBZ

  • Opcode word: 0x7c2007ec
  • Primary opcode (bits 0–5): 31
  • Extended opcode: 1014
  • Synchronising: no
Bits Field Meaning
0–5 OPCD primary opcode (31)
6–10 — reserved
11–15 RA base register (0 ⇒ literal 0)
16–20 RB offset register
21–30 XO extended opcode (1014 for dcbz / 1010 for dcbz128)
31 — reserved

Operands

Field Role Description
RA0 dcbz: read; dcbz128: read Source GPR; when the encoded register number is 0 the operand is the literal 64-bit zero, not r0.
RB dcbz: read; dcbz128: read Source GPR.

Register Effects

dcbz

  • Reads (always): RA0, RB
  • Reads (conditional): none
  • Writes (always): none
  • Writes (conditional): none

dcbz128

  • Reads (always): RA0, RB
  • Reads (conditional): none
  • Writes (always): none
  • Writes (conditional): none

Status-Register Effects

No condition-register or status-register effects.

Operation (pseudocode)

; No hand-written pseudocode for this instruction yet.
; The authoritative semantics are the Canary emitter snapshot under
; Implementation References; about half of Canary's emitters open
; with the PPC-style definition as a comment (`RD <- (RA) + (RB)`).
; Every side effect is also enumerated in the Register Effects and
; Status-Register Effects tables above.

C Translation Example

/* No hand-written C yet. Translate the Canary emitter snapshot   */
/* under Implementation References; its HIR maps directly:        */
/*   f.LoadGPR(n) / f.StoreGPR(n, v)  -> r[n] / r[n] = v          */
/*   f.LoadFPR / StoreFPR, f.LoadVR / StoreVR -> f[n], v[n]        */
/*   f.Load(ea, T), f.Store(ea, v) -> raw read / write; emitters   */
/*     wrap them in f.ByteSwap for the big-endian guest value      */
/*   f.UpdateCR(n, v)  -> CR field n from v's LOW 32 BITS vs 0     */
/*   f.LoadCA / f.StoreCA -> xer.CA;  f.StoreSAT -> vscr.SAT       */
/*   i.XO.RA, i.D.DS, ... -> the bit-fields listed under Operands  */
/* The Register Effects and Status-Register Effects tables above  */
/* enumerate every side effect a faithful translation must emit.  */

Implementation References

dcbz

Canary emitter (frozen snapshot @ f21ebd49e9)
int InstrEmit_dcbz(PPCHIRBuilder& f, const InstrData& i) {
  // EA <- (RA) + (RB)
  // memset(EA & ~31, 0, 32)
  // On Xbox360 there is no short cache line. Normal dcbz always clears 128
  // bytes.
  return InstrEmit_dcbz128(f, i);
}

// ── delegates to (src/xenia/cpu/ppc/ppc_emit_memory.cc:1159) ──
int InstrEmit_dcbz128(PPCHIRBuilder& f, const InstrData& i) {
  // EA <- (RA) + (RB)
  // memset(EA & ~31, 0, 32)
  Value* ea = CalculateEA_0(f, i.X.RA, i.X.RB);
  // dcbz128 - 128 byte set
  int block_size = 128;
  int address_mask = ~127;
  f.Memset(f.And(ea, f.LoadConstantInt64(address_mask)), f.LoadZeroInt8(),
           f.LoadConstantInt64(block_size));
  return 0;
}

dcbz128

Canary emitter (frozen snapshot @ f21ebd49e9)
int InstrEmit_dcbz128(PPCHIRBuilder& f, const InstrData& i) {
  // EA <- (RA) + (RB)
  // memset(EA & ~31, 0, 32)
  Value* ea = CalculateEA_0(f, i.X.RA, i.X.RB);
  // dcbz128 - 128 byte set
  int block_size = 128;
  int address_mask = ~127;
  f.Memset(f.And(ea, f.LoadConstantInt64(address_mask)), f.LoadZeroInt8(),
           f.LoadConstantInt64(block_size));
  return 0;
}

Special Cases & Edge Conditions

  • Cache-line size mismatch. Stock PowerPC dcbz zeroes one architectural cache line — 32 bytes on classic POWER, but the Xenon's L1 line is 128 bytes. Microsoft added dcbz128 (encoded with bit-9 set so RT field reads as 1) to clear a true Xenon line in one instruction. Most Xbox 360 code therefore emits dcbz128; a stray dcbz only zeroes 32 bytes and silently leaves the rest of the line uncleared.
  • Alignment is forced via mask. The effective address is masked by ~31 (dcbz) or ~127 (dcbz128) before writing — the low bits are dropped, not validated. Calling dcbz r0, r3 with r3 = 0x10037 writes zeros to 0x10000..0x1007F, not 0x10037..0x100B6.
  • No memory read; pure write. Real hardware allocates the line in cache and may skip a read-from-memory fill ("cache-line zero" optimisation). Xenia simulates the architectural effect — 32 (or 128) bytes of zero in target memory — without modelling cache state.
  • RA0 semantics. RA = 0 selects literal zero as the base, so dcbz128 0, RB zeros the line containing address RB. The update form does not exist for cache-control instructions.
  • Block-fill idiom. Compilers and hand-written copy loops pair dcbz128 with stvx / stw sequences to avoid the cache-line read-allocate that a cold store would trigger. Skipping the read is the entire point.
  • Privilege. dcbz is unprivileged (problem-state); does not require supervisor mode. It can fault on protection or unmapped memory like an ordinary store.
  • Sequencing. Not synchronising. Pair with sync / lwsync when the zeros must be visible before subsequent loads on another thread.
  • dcbf — flush a line back to memory.
  • dcbst — store-through (write-back without invalidate).
  • dcbi — invalidate (privileged on most cores).
  • dcbt, dcbtst — touch / touch-for-store hints.
  • icbi — instruction-cache invalidate (companion to data-cache control).
  • stvx, stw — typical pair-mates in block-fill loops.

IBM Reference