Skip to content

Fix: Potential out-of-bounds read in non-transposed [SD]GEMV kernels for A64FX and Neoverse V1. #5240

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
martin-frbg merged 1 commit into from
Apr 22, 2025
Merged

Fix: Potential out-of-bounds read in non-transposed [SD]GEMV kernels for A64FX and Neoverse V1. #5240

Show file tree
Hide file tree
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
6 changes: 3 additions & 3 deletions kernel/arm64/gemv_n_sve_v1x3.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -77,9 +77,9 @@ int CNAME(BLASLONG m, BLASLONG n, BLASLONG dummy1, FLOAT alpha, FLOAT *a,
svbool_t pg01 = ((j + width * 1) < n) ? SV_TRUE() : svpfalse();
svbool_t pg02 = ((j + width * 2) < n) ? SV_TRUE() : svpfalse();

SV_TYPE temp0_vec = SV_DUP(alpha * x0_ptr[ix]);
SV_TYPE temp1_vec = SV_DUP(alpha * x1_ptr[ix]);
SV_TYPE temp2_vec = SV_DUP(alpha * x2_ptr[ix]);
SV_TYPE temp0_vec = ((j + width * 0) < n) ? SV_DUP(alpha * x0_ptr[ix]) : SV_DUP(0.0);
SV_TYPE temp1_vec = ((j + width * 1) < n) ? SV_DUP(alpha * x1_ptr[ix]) : SV_DUP(0.0);
SV_TYPE temp2_vec = ((j + width * 2) < n) ? SV_DUP(alpha * x2_ptr[ix]) : SV_DUP(0.0);
i = 0;
BLASLONG sve_size = SV_COUNT();
while ((i + sve_size * 1 - 1) < m) {
Expand Down
6 changes: 3 additions & 3 deletions kernel/arm64/gemv_n_sve_v4x3.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -86,9 +86,9 @@ int CNAME(BLASLONG m, BLASLONG n, BLASLONG dummy1, FLOAT alpha, FLOAT *a,
svbool_t pg22 = ((j + width * 2) < n) ? SV_TRUE() : svpfalse();
svbool_t pg32 = ((j + width * 2) < n) ? SV_TRUE() : svpfalse();

SV_TYPE temp0_vec = SV_DUP(alpha * x0_ptr[ix]);
SV_TYPE temp1_vec = SV_DUP(alpha * x1_ptr[ix]);
SV_TYPE temp2_vec = SV_DUP(alpha * x2_ptr[ix]);
SV_TYPE temp0_vec = ((j + width * 0) < n) ? SV_DUP(alpha * x0_ptr[ix]) : SV_DUP(0.0);
SV_TYPE temp1_vec = ((j + width * 1) < n) ? SV_DUP(alpha * x1_ptr[ix]) : SV_DUP(0.0);
SV_TYPE temp2_vec = ((j + width * 2) < n) ? SV_DUP(alpha * x2_ptr[ix]) : SV_DUP(0.0);
i = 0;
BLASLONG sve_size = SV_COUNT();
while ((i + sve_size * 4 - 1) < m) {
Expand Down
Loading