Skip to content

GH-127953: Make line number lookup O(1) regardless of the size of the code object #128350

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 7 commits into from
Jan 21, 2025
Merged

GH-127953: Make line number lookup O(1) regardless of the size of the code object #128350

Changes from 2 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
751143f
Make line number lookup O(1) regardless of the size of the code object
markshannon Dec 30, 2024
262abf1
Use co_monitoring line info in PyCode_Addr2Line if it is available
markshannon Dec 30, 2024
9ea88c0
Address review comments
markshannon Jan 3, 2025
c816f7e
Assert that bytes_per_entry is in range(2,6)
markshannon Jan 3, 2025
20917cc
Address review comments
markshannon Jan 20, 2025
a1245e5
Merge branch 'main' into variable-sized-line-cache
markshannon Jan 20, 2025
31601c6
Fix up comments
markshannon Jan 20, 2025
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
44 changes: 16 additions & 28 deletions Python/instrumentation.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -348,20 +348,12 @@ static inline int
get_line_delta(_PyCoLineInstrumentationData *line_data, int index)
{
uint8_t *ptr = &line_data->data[index*line_data->bytes_per_entry+1];
uint32_t value = *ptr;
assert(line_data->bytes_per_entry >= 2);
if (line_data->bytes_per_entry > 2) {
uint32_t value = *ptr;
Copy link
Member

Choose a reason for hiding this comment

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

Could assert here that bytes_per_entry is within range.

Copy link
Member Author

Choose a reason for hiding this comment

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

We can't easily assert anything here, as we don't know the max line number without traversing the entire locations table.
I've added an assert to set_line_delta instead.

Copy link
Member

Choose a reason for hiding this comment

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

I meant that bytes_per_entry is between 1 and 4.

Copy link
Member Author

Choose a reason for hiding this comment

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

(bytes_per_entry is between 2 and 5 because it includes the original opcode)

for (int idx = 2; idx < line_data->bytes_per_entry; idx++) {
ptr++;
value = (value << 8) | *ptr;
if (line_data->bytes_per_entry > 3) {
ptr++;
value = (value << 8) | *ptr;
if (line_data->bytes_per_entry > 4) {
assert(line_data->bytes_per_entry == 5);
ptr++;
value = (value << 8) | *ptr;
}
}
int shift = (idx-1)*8;
value |= ((uint32_t)(*ptr)) << shift;
}
assert(value < INT_MAX);
/* NO_LINE is stored as zero. */
Expand All @@ -375,22 +367,14 @@ set_line_delta(_PyCoLineInstrumentationData *line_data, int index, int line_delt
assert(line_delta >= NO_LINE);
uint32_t adjusted = line_delta - NO_LINE;
uint8_t *ptr = &line_data->data[index*line_data->bytes_per_entry+1];
assert(adjusted < (1ULL << (line_data->bytes_per_entry*8)));
assert(adjusted < (1ULL << ((line_data->bytes_per_entry-1)*8)));
assert(line_data->bytes_per_entry >= 2);
if (line_data->bytes_per_entry > 2) {
if (line_data->bytes_per_entry > 3) {
if (line_data->bytes_per_entry > 4) {
assert(line_data->bytes_per_entry == 5);
*ptr = adjusted >> 24;
ptr++;
}
*ptr = (adjusted >> 16) & 255;
ptr++;
}
*ptr = (adjusted >> 8) & 255;
*ptr = adjusted & 0xff;
for (int idx = 2; idx < line_data->bytes_per_entry; idx++) {
ptr++;
adjusted >>= 8;
*ptr = adjusted & 0xff;
}
*ptr = adjusted & 255;
}

#ifdef INSTRUMENT_DEBUG
Expand Down Expand Up @@ -1541,9 +1525,10 @@ initialize_lines(PyCodeObject *code, int bytes_per_entry)
case END_FOR:
case END_SEND:
case RESUME:
case POP_ITER:
/* END_FOR cannot start a line, as it is skipped by FOR_ITER
* END_SEND cannot start a line, as it is skipped by SEND
* RESUME must not be instrumented with INSTRUMENT_LINE */
* RESUME and POP_ITER must not be instrumented with INSTRUMENT_LINE */
set_original_opcode(line_data, i, 0);
break;
default:
Expand Down Expand Up @@ -1610,7 +1595,10 @@ initialize_lines(PyCodeObject *code, int bytes_per_entry)
}
assert(target >= 0);
if (get_line_delta(line_data, target) != NO_LINE) {
set_original_opcode(line_data, target, _Py_GetBaseCodeUnit(code, target).op.code);
int opcode = _Py_GetBaseCodeUnit(code, target).op.code;
if (opcode != POP_ITER) {
set_original_opcode(line_data, target, opcode);
}
}
}
/* Scan exception table */
Expand Down Expand Up @@ -1733,7 +1721,7 @@ update_instrumentation_data(PyCodeObject *code, PyInterpreterState *interp)
else {
bytes_per_entry = 5;
}
code->_co_monitoring->lines = PyMem_Malloc(1 + code_len *bytes_per_entry);
code->_co_monitoring->lines = PyMem_Malloc(1 + code_len * bytes_per_entry);
if (code->_co_monitoring->lines == NULL) {
PyErr_NoMemory();
return -1;
Expand Down
Loading