Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix dereference behavior on mixed subscript and arrow / dot operators #182

Merged
merged 1 commit into from
Mar 24, 2025
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 1 addition & 3 deletions src/globals.c
Original file line number Diff line number Diff line change
@@ -250,9 +250,7 @@ void hashmap_free(hashmap_t *map)
next = cur->next;
free(cur->key);
free(cur->val);
/* FIXME: Remove this if-clause will cause double free error */
if (cur != map->buckets[0])
free(cur);
free(cur);
cur = next;
}
}
47 changes: 32 additions & 15 deletions src/parser.c
Original file line number Diff line number Diff line change
@@ -750,7 +750,7 @@ void read_lvalue(lvalue_t *lvalue,
var_t *var,
block_t *parent,
basic_block_t **bb,
int eval,
bool eval,
opcode_t op);

/* Maintain a stack of expression values and operators, depending on next
@@ -805,7 +805,7 @@ void read_expr_operand(block_t *parent, basic_block_t **bb)

lex_peek(T_identifier, token);
var_t *var = find_var(token, parent);
read_lvalue(&lvalue, var, parent, bb, 0, OP_generic);
read_lvalue(&lvalue, var, parent, bb, false, OP_generic);

if (!lvalue.is_reference) {
ph1_ir = add_ph1_ir(OP_address_of);
@@ -825,7 +825,7 @@ void read_expr_operand(block_t *parent, basic_block_t **bb)
int open_bracket = lex_accept(T_open_bracket);
lex_peek(T_identifier, token);
var_t *var = find_var(token, parent);
read_lvalue(&lvalue, var, parent, bb, 1, OP_generic);
read_lvalue(&lvalue, var, parent, bb, true, OP_generic);
if (open_bracket)
lex_expect(T_close_bracket);

@@ -962,7 +962,7 @@ void read_expr_operand(block_t *parent, basic_block_t **bb)
} else if (var) {
/* evalue lvalue expression */
lvalue_t lvalue;
read_lvalue(&lvalue, var, parent, bb, 1, prefix_op);
read_lvalue(&lvalue, var, parent, bb, true, prefix_op);

/* is it an indirect call with function pointer? */
if (lex_peek(T_open_bracket, NULL)) {
@@ -1234,13 +1234,13 @@ void read_lvalue(lvalue_t *lvalue,
var_t *var,
block_t *parent,
basic_block_t **bb,
int eval,
bool eval,
opcode_t prefix_op)
{
ph1_ir_t *ph1_ir;
var_t *vd;
int is_address_got = 0;
int is_member = 0;
bool is_address_got = false;
bool is_member = false;

/* already peeked and have the variable */
lex_expect(T_identifier);
@@ -1260,6 +1260,23 @@ void read_lvalue(lvalue_t *lvalue,
while (lex_peek(T_open_square, NULL) || lex_peek(T_arrow, NULL) ||
lex_peek(T_dot, NULL)) {
if (lex_accept(T_open_square)) {
/* if subscripted member's is not yet resolved, dereference to
* resolve base address.
* e.g., dereference of "->" in "data->raw[0]" would be performed
* here.
*/
if (lvalue->is_reference && lvalue->is_ptr && is_member) {
ph1_ir = add_ph1_ir(OP_read);
ph1_ir->src0 = opstack_pop();
vd = require_var(parent);
strcpy(vd->var_name, gen_name());
ph1_ir->dest = vd;
opstack_push(vd);
ph1_ir->size = 4;
add_insn(parent, *bb, OP_read, ph1_ir->dest, ph1_ir->src0, NULL,
ph1_ir->size, NULL);
}

/* var must be either a pointer or an array of some type */
if (var->is_ptr == 0 && var->array_size == 0)
error("Cannot apply square operator to non-pointer");
@@ -1303,8 +1320,8 @@ void read_lvalue(lvalue_t *lvalue,
ph1_ir->src1, 0, NULL);

lex_expect(T_close_square);
is_address_got = 1;
is_member = 1;
is_address_got = true;
is_member = true;
lvalue->is_reference = true;
} else {
char token[MAX_ID_LEN];
@@ -1313,7 +1330,7 @@ void read_lvalue(lvalue_t *lvalue,
/* resolve where the pointer points at from the calculated
* address in a structure.
*/
if (is_member == 1) {
if (is_member) {
ph1_ir = add_ph1_ir(OP_read);
ph1_ir->src0 = opstack_pop();
vd = require_var(parent);
@@ -1327,7 +1344,7 @@ void read_lvalue(lvalue_t *lvalue,
} else {
lex_expect(T_dot);

if (is_address_got == 0) {
if (!is_address_got) {
ph1_ir = add_ph1_ir(OP_address_of);
ph1_ir->src0 = opstack_pop();
vd = require_var(parent);
@@ -1337,7 +1354,7 @@ void read_lvalue(lvalue_t *lvalue,
add_insn(parent, *bb, OP_address_of, ph1_ir->dest,
ph1_ir->src0, NULL, 0, NULL);

is_address_got = 1;
is_address_got = true;
}
}

@@ -1376,8 +1393,8 @@ void read_lvalue(lvalue_t *lvalue,
add_insn(parent, *bb, OP_add, ph1_ir->dest, ph1_ir->src0,
ph1_ir->src1, 0, NULL);

is_address_got = 1;
is_member = 1;
is_address_got = true;
is_member = true;
}
}

@@ -1860,7 +1877,7 @@ bool read_body_assignment(char *token,
int size = 0;

/* has memory address that we want to set */
read_lvalue(&lvalue, var, parent, bb, 0, OP_generic);
read_lvalue(&lvalue, var, parent, bb, false, OP_generic);
size = lvalue.size;

if (lex_accept(T_increment)) {
43 changes: 43 additions & 0 deletions tests/driver.sh
Original file line number Diff line number Diff line change
@@ -433,6 +433,49 @@ int main() {
}
EOF

# Mixed subscript and arrow / dot operators,
# excerpted and modified from issue #165
try_output 0 "DDDDDDMMMEEE1" << EOF
#include <stdlib.h>
#include <string.h>

char a[100];

typedef struct {
char *raw;
} data_t;

int main() {
strcpy(a, "DATA");
data_t *data = malloc(sizeof(data_t));
data->raw = a;
data_t data2;
data2.raw = a;
char *raw = data->raw;
char *raw2 = data2.raw;
/* mixed arrow / dot with subscript operators dereference */
printf("%c", a[0]);
printf("%c", raw[0]);
printf("%c", data->raw[0]);
printf("%c", a[0]);
printf("%c", raw2[0]);
printf("%c", data2.raw[0]);
/* mixed arrow / dot with subscript operators assignment */
data2.raw[0] = 'M';
data->raw[1] = 'E';
printf("%c", a[0]);
printf("%c", raw[0]);
printf("%c", data->raw[0]);
printf("%c", a[1]);
printf("%c", raw2[1]);
printf("%c", data2.raw[1]);
/* their addresses should be same */
printf("%d", &data2.raw[0] == &data->raw[0]);
free(data);
return 0;
}
EOF

# global initialization
try_ 20 << EOF
int a = 5 * 2;