Skip to content

Commit

Permalink
This should address "strict-aliasing". Cosmetic change to a wasm file…
Browse files Browse the repository at this point in the history 
…. Also addresses CodeQL.
  • Loading branch information
@aous72
aous72 committed Sep 20, 2024
1 parent 7b0f5c8 commit b6ddad9
Show file tree
Hide file tree
Showing 2 changed files with 47 additions and 37 deletions.
80 changes: 45 additions & 35 deletions src/apps/others/ojph_img_io.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -593,11 +593,11 @@ namespace ojph {
start_of_data = ojph_ftell(fh);

// alloc. linebuffer to hold a line of image data, if more than 1 comp.
if (temp_buf_byte_size < num_comps * width * sizeof(float))
if (temp_buf_byte_size < num_comps * (size_t)width * sizeof(float))
{
if (alloc_p == NULL)
{
temp_buf_byte_size = num_comps * width * sizeof(float);
temp_buf_byte_size = num_comps * (size_t)width * sizeof(float);
void* t = temp_buf;
if (temp_buf)
temp_buf = (float*)realloc(temp_buf, temp_buf_byte_size);
Expand All @@ -611,7 +611,7 @@ namespace ojph {
else
{
assert(temp_buf_byte_size == 0); //cannot reallocate the buffer
temp_buf_byte_size = num_comps * width * sizeof(float);
temp_buf_byte_size = num_comps * (size_t)width * sizeof(float);
alloc_p->pre_alloc_data<float>(temp_buf_byte_size, 0);
}
}
Expand All @@ -623,7 +623,7 @@ namespace ojph {
{
if (alloc_p == NULL)
return;
temp_buf = alloc_p->post_alloc_data<float>(num_comps * width, 0);
temp_buf = alloc_p->post_alloc_data<float>(num_comps * (size_t)width, 0);
}

/////////////////////////////////////////////////////////////////////////////
Expand All @@ -636,13 +636,14 @@ namespace ojph {
if (comp_num == 0)
{
si64 loc = start_of_data;
loc += (height - 1 - cur_line) * num_comps * width * sizeof(float);
loc += (size_t)(height-1 - cur_line) * num_comps * width * sizeof(float);
if (ojph_fseek(fh, loc, SEEK_SET) != 0)
{
close();
OJPH_ERROR(0x03000061, "Error seeking in file %s", fname);
}
size_t result = fread(temp_buf, sizeof(float), num_comps * width, fh);
size_t result =
fread(temp_buf, sizeof(float), num_comps * (size_t)width, fh);
if (result != num_comps * width)

Check failure

Code scanning / CodeQL

Multiplication result converted to larger type High

Multiplication result may overflow 'unsigned int' before it is converted to 'unsigned long'.
{
close();
Expand All @@ -652,38 +653,42 @@ namespace ojph {
cur_line = 0;
}

union {
si32* s;
ui32* u;
float* f;
} sp, dp;

if (little_endian)
{
ui32 shift = 32 - bit_depth[comp_num];
const float* sp = temp_buf + comp_num;
float* dp = line->f32;
sp.f = temp_buf + comp_num;
dp.f = line->f32;
if (shift)
for (ui32 i = width; i > 0; --i, sp += num_comps)
for (ui32 i = width; i > 0; --i, sp.f += num_comps)
{
si32 v = *(si32*)sp;
v >>= shift;
*dp++ = *(float*)&v;
si32 s = *sp.s;
s >>= shift;
*dp.s++ = s;
}
else
for (ui32 i = width; i > 0; --i, sp += num_comps)
*dp++ = *sp;
for (ui32 i = width; i > 0; --i, sp.f += num_comps)
*dp.f++ = *sp.f;
}
else {
ui32 shift = 32 - bit_depth[comp_num];
const float* sp = temp_buf + comp_num;
float* dp = line->f32;
sp.f = temp_buf + comp_num;
dp.f = line->f32;
if (shift)
for (ui32 i = width; i > 0; --i, sp += num_comps) {
ui32 v = be2le(*(ui32*)sp);
si32 u = *(si32*)&v;
u >>= shift;
*dp++ = *(float*)&u;
for (ui32 i = width; i > 0; --i, sp.f += num_comps) {
ui32 u = be2le(*sp.u);
si32 s = *(si32*)&u;
s >>= shift;
*dp.s++ = s;
}
else
for (ui32 i = width; i > 0; --i, sp += num_comps) {
ui32 v = be2le(*(ui32*)sp);
*dp++ = *(float*)&v;
}
for (ui32 i = width; i > 0; --i, sp.f += num_comps)
*dp.u++ = be2le(*sp.u);
}

return width;
Expand All @@ -710,7 +715,7 @@ namespace ojph {
num_components > 1 ? 'F' : 'f', width, height, scale);
if (result == 0)
OJPH_ERROR(0x03000072, "error writing to file %s", filename);
buffer_size = width * num_components * sizeof(float);
buffer_size = (size_t)width * num_components * sizeof(float);
buffer = (float*)malloc(buffer_size);
fname = filename;
cur_line = 0;
Expand Down Expand Up @@ -739,23 +744,28 @@ namespace ojph {
assert(fh);

ui32 shift = 32 - bit_depth[comp_num];
float* dp = buffer + comp_num;
const float* sp = line->f32;
union {
ui32* u;
float* f;
} sp, dp;

dp.f = buffer + comp_num;
sp.f = line->f32;

if (shift)
for (ui32 i = width; i > 0; --i, dp += num_components, ++sp)
for (ui32 i = width; i > 0; --i, dp.f += num_components, ++sp.f)
{
ui32 v = *(ui32*)sp;
v <<= shift;
*dp = *(float*)&v;
ui32 u = *sp.u;
u <<= shift;
*dp.u = u;
}
else
for (ui32 i = width; i > 0; --i, dp += num_components)
*dp = *sp++;
for (ui32 i = width; i > 0; --i, dp.f += num_components)
*dp.f = *sp.f++;

if (comp_num == num_components - 1)
{
size_t samples_per_line = num_components * width;
size_t samples_per_line = num_components * (size_t)width;
si64 loc = start_of_data;
loc += (height - 1 - cur_line)* samples_per_line * sizeof(float);
if (ojph_fseek(fh, loc, SEEK_SET) != 0)
Expand Down
4 changes: 2 additions & 2 deletions src/core/transform/ojph_colour_wasm.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -133,8 +133,8 @@ namespace ojph {
v128_t s = wasm_v128_load(sp);
v128_t c = wasm_i32x4_lt(s, zero); // 0xFFFFFFFF for -ve value
v128_t v_m_sh = wasm_i32x4_sub(sh, s); // - shift - value
v_m_sh = wasm_v128_and(v_m_sh, c); // keep only - shift - value
s = wasm_v128_andnot(s, c); // keep only +ve or 0
v_m_sh = wasm_v128_and(c, v_m_sh); // keep only - shift - value
s = wasm_v128_andnot(c, s); // keep only +ve or 0
s = wasm_v128_or(s, v_m_sh); // combine
wasm_v128_store(dp, s);
}
Expand Down

0 comments on commit b6ddad9

Please sign in to comment.