Skip to content

gh-126024: unicodeobject: optimize find_first_nonascii #127790

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.

Sign up for GitHub

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

Merged
merged 1 commit into from
Dec 13, 2024
Merged

gh-126024: unicodeobject: optimize find_first_nonascii #127790

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
23 changes: 14 additions & 9 deletions Objects/unicodeobject.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5077,28 +5077,32 @@ load_unaligned(const unsigned char *p, size_t size)
static Py_ssize_t
find_first_nonascii(const unsigned char *start, const unsigned char *end)
{
// The search is done in `size_t` chunks.
// The start and end might not be aligned at `size_t` boundaries,
// so they're handled specially.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@encukou I copied your comments.

const unsigned char *p = start;

if (end - start >= SIZEOF_SIZE_T) {
const unsigned char *p2 = _Py_ALIGN_UP(p, SIZEOF_SIZE_T);
// Avoid unaligned read.
#if PY_LITTLE_ENDIAN && HAVE_CTZ
if (p < p2) {
size_t u;
memcpy(&u, p, sizeof(size_t));
u &= ASCII_CHAR_MASK;
if (u) {
return (ctz(u) - 7) / 8;
}
p = p2;
size_t u;
memcpy(&u, p, sizeof(size_t));
u &= ASCII_CHAR_MASK;
if (u) {
return (ctz(u) - 7) / 8;
}
p = _Py_ALIGN_DOWN(p + SIZEOF_SIZE_T, SIZEOF_SIZE_T);
#else /* PY_LITTLE_ENDIAN && HAVE_CTZ */
const unsigned char *p2 = _Py_ALIGN_UP(p, SIZEOF_SIZE_T);
while (p < p2) {
if (*p & 0x80) {
return p - start;
}
p++;
}
#endif

const unsigned char *e = end - SIZEOF_SIZE_T;
while (p <= e) {
size_t u = (*(const size_t *)p) & ASCII_CHAR_MASK;
Expand All @@ -5115,6 +5119,7 @@ find_first_nonascii(const unsigned char *start, const unsigned char *end)
}
}
#if PY_LITTLE_ENDIAN && HAVE_CTZ
assert((end - p) < SIZEOF_SIZE_T);
// we can not use *(const size_t*)p to avoid buffer overrun.
size_t u = load_unaligned(p, end - p) & ASCII_CHAR_MASK;
if (u) {
Expand Down
Loading