Skip to content

kleidiai: fix unsigned overflow bug #15150

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 2 commits into from
Aug 11, 2025
Merged

kleidiai: fix unsigned overflow bug #15150

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
1eb1795
kleidiai: fix unsigned overflow bug
chaxu01 Aug 7, 2025
129e7da
address review comments
chaxu01 Aug 7, 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
25 changes: 17 additions & 8 deletions ggml/src/ggml-cpu/kleidiai/kleidiai.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -259,7 +259,10 @@ class tensor_traits : public ggml::cpu::tensor_traits {
const int64_t m_start = 0;

const int64_t n_step = static_cast<int64_t>(kernel->get_n_step());
const int64_t num_threads = KAI_MIN(n / n_step, nth);
int64_t num_threads = KAI_MIN(n / n_step, nth);
if (num_threads <= 0) {
num_threads = 1;
}

if (ith < num_threads) {
const int64_t num_n_per_thread0 = round_down(n / num_threads, n_step);
Expand Down Expand Up @@ -308,8 +311,9 @@ class tensor_traits : public ggml::cpu::tensor_traits {

GGML_ASSERT(kernel);

const int ith = params->ith;
const int nth = params->nth;
int ith = params->ith;
int nth_raw = params->nth;
int nth = nth_raw > 0 ? nth_raw : 1;

const size_t k = ne00;
const size_t m = ne11;
Expand All @@ -327,9 +331,12 @@ class tensor_traits : public ggml::cpu::tensor_traits {
const size_t num_n_per_thread = kai_roundup(kai_roundup(n, nth) / nth, n_step);
const size_t n_start = ith * num_n_per_thread;

size_t n_to_process = num_n_per_thread;
if ((n_start + n_to_process) > n) {
n_to_process = n - n_start;
size_t n_to_process = 0;
if (n_start < n) {
n_to_process = num_n_per_thread;
if ((n_start + n_to_process) > n) {
n_to_process = n - n_start;
}
}

// Calculate number of columns to be processed per thread
Expand Down Expand Up @@ -361,8 +368,10 @@ class tensor_traits : public ggml::cpu::tensor_traits {
const void* lhs_ptr = (const void*)((const char *)lhs_packed + lhs_packed_offset);
float *dst_ptr = reinterpret_cast<float *>(static_cast<uint8_t *>(dst->data) + dst_offset);

variant_call<void>(kernel->run_kernel, m, n_to_process, k, QK4_0, lhs_ptr, rhs_ptr, dst_ptr, dst_stride,
sizeof(float), -FLT_MAX, FLT_MAX);
if (n_to_process > 0) {
variant_call<void>(kernel->run_kernel, m, n_to_process, k, QK4_0, lhs_ptr, rhs_ptr, dst_ptr, dst_stride,
sizeof(float), -FLT_MAX, FLT_MAX);
}

return true;
}
Expand Down
Loading