Hojun-Cho/ww
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

selfhost+test: strlit-inline str-def value-load shape (#12)

Browse Source 
Class A silent miscompile. wwstage cgenexpr.ww cgident's bare-ident
deflookup→true branch and cgdot's module-qualified leaf branch
emitted `MOVQ <mod>.<name>(SB), AX` for a `def MSG: str = "..."`
value reference — a load from a SB symbol that emit_data never
writes. Str defs are not laid out at SB; they live as interned
strlits the .ptr/.len fold (post-#4c) and value-load consume.
Cstage already strlit-inlines via Sdef walks #1 (case N_IDENT
non-local) and #2 (case N_DOT untyped-lhs); wwstage now matches
the (LEAQ _S_<n>(SB), MOVQ $<len>, BX) emit shape per rule 10.

Surfaced by reviewer-def during #4c R3 while attempting option (B)
for the cstage Sdef walks #1/#2 prefer-pass — both walks'
cs-vs-ws byte-id sentinel rows could not pass while wwstage
emitted the bogus DATAW shape. Filed as #12 and deferred until
the wwstage emit shape was fixed. Unblocks #11 + #13 (cstage
prefer-pass graduations).

Latent: no in-tree corpus referenced a str def as a value (only
as .ptr/.len via cgdot field-fold) prior to lib/strings c3 —
same corpus-coverage-blind shape as the #4a-#4e graduations.

746_strdef_inline pins both sites with 2 rows: bare ident +
mod-qualified. Each row asserts `LEAQ _S_` + `MOVQ $<strlit_len>,`
inside the caller TEXT before RET, anti-checks the pre-fix
`<mod>.<name>(SB)` symbol-load, and cs-vs-ws byte-id per row.

120/120 ok. ww2 == ww3 == ww4 byte-id holds.
This commit is contained in:
Hojun-Cho
2026-05-18 23:56:58 +09:00
parent 049ebc14a1
commit d985622cb1
5 changed files with 354 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
selfhost/cmd/wcc/cgenexpr.ww
View File

@@ -552,7 +552,26 @@ fn cgident(c: *cgen, n: *node) void = {
return;
};
// Top-level `def` constant — load from its DATA symbol.
// Str defs (rhs N_STRLIT) aren't laid out at a SB symbol; the
// MOVQ symname(SB) fallback below would emit a bogus reference
// (e.g. `alpha.MSG(SB)`, never DATAW-defined). Strlit-inline
// the (LEAQ ptr, MOVQ $len) pair instead, mirroring cstage
// Sdef walk #1 N_IDENT bare-load (cmd/w6c/cgen.c). Filed #12.
if (deflookup(c, nm)) {
let drhs: *node = deflookuprhs(c, nm);
if (drhs != nil) {
if (drhs.kind == nkind.N_STRLIT) {
let bytes: str = drhs.str;
let lab: str = internstrlit(c, bytes);
emitline("\tLEAQ\t");
os.write(1, lab.ptr, lab.len: u64);
emitline("(SB), AX\n");
emitline("\tMOVQ\t$");
emitint(bytes.len: i64);
emitline(", BX\n");
return;
};
};
emitline("\tMOVQ\t");
emitsymname(c, nm);
emitline("(SB), AX\n");
@@ -1721,6 +1740,25 @@ fn cgdot(c: *cgen, n: *node) void = {
emitline("(SB), AX\n");
return;
};
// `mod.MSG` where MSG is `def MSG: str = "..."` —
// strlit-inline matches cstage Sdef walk #2 in
// cmd/w6c/cgen.c N_DOT mod-qualified. Without this
// the MOVQ leaf(SB) fallback emits a bogus ref
// (`alpha.MSG(SB)`, never DATAW-defined). Filed #12.
let drhs: *node = deflookuprhs(c, fld);
if (drhs != nil) {
if (drhs.kind == nkind.N_STRLIT) {
let bytes: str = drhs.str;
let lab: str = internstrlit(c, bytes);
emitline("\tLEAQ\t");
os.write(1, lab.ptr, lab.len: u64);
emitline("(SB), AX\n");
emitline("\tMOVQ\t$");
emitint(bytes.len: i64);
emitline(", BX\n");
return;
};
};
let mqop: str = localloadop(c, letvartnode(c, fld));
if (streq(mqop, "MOVQ")) {
emitline("\tMOVQ\t");