Skip to content

[GPU] Fix: Prevent out-of-bounds read in dequantize merge #32302

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

Open
wants to merge 5 commits into
base: master
Choose a base branch
from
+5 −1
Open

[GPU] Fix: Prevent out-of-bounds read in dequantize merge #32302

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
303ae28
Fix: Prevent out-of-bounds read in dequantize merge
sshekhar563 Oct 6, 2025
0d8161b
Merge branch 'master' into fix-dequantize-segfault
sshekhar563 Oct 6, 2025
6181432
Update prepare_quantization.cpp
sshekhar563 Oct 6, 2025
44ad611
Update prepare_quantization.cpp
sshekhar563 Oct 6, 2025
f1244ce
Update prepare_quantization.cpp
sshekhar563 Oct 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
6 changes: 5 additions & 1 deletion src/plugins/intel_gpu/src/graph/graph_optimizer/prepare_quantization.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -372,6 +372,11 @@ void prepare_quantization::prepare_dequantize_merge(program& p, eltwise_node& el
for (size_t i = 1; i < eltwise_node.get_dependencies().size(); i++) {
auto mem0 = get_scale_shift_mem(eltwise_dep, i);
auto mem1 = get_scale_shift_mem(eltwise_node, i);

if (mem0->get_layout().bytes_count() != mem1->get_layout().bytes_count()) {
same_params = false;
break;
}

mem_lock<uint8_t, mem_lock_type::read> mem0_lock{mem0, stream};
mem_lock<uint8_t, mem_lock_type::read> mem1_lock{mem1, stream};
Expand All @@ -384,7 +389,6 @@ void prepare_quantization::prepare_dequantize_merge(program& p, eltwise_node& el
break;
}
}

// Avoid mem0 and mem1's memory are inplace, but they have different layout.
if (!mem0->get_layout().get_partial_shape().compatible(mem1->get_layout().get_partial_shape())) {
same_params = false;
Expand Down
Loading