fix: Resolve integer overflow in Load API file decoding

This PR addresses the following issue:

When large data is sent to /v2/repository/models/<model_name>/load, the server converts the param_len variable from size_t to const int to pass it to base64_decode_block. This conversion can cause an integer overflow in base64_decode_block, leading to a server crash with a segmentation fault.
A similar issue occurs with /v2/cudasharedmemory/region/{}/register when large data is sent for the raw_handle parameter.

[x] PR title reflects the change and is of format <commit_type>: <Title>
[x] Changes are described in the pull request.
[ ] Related issues are referenced.
[x] Populated github labels field
[x] Added test plan and verified test passes.
[ ] Verified that the PR passes existing CI.
[x] Verified copyright is correct on all changed files.
[ ] Added succinct git squash message before merging ref.
[ ] All template sections are filled out.
[ ] Optional: Additional screenshots for behavior/output changes with before/after.

Check the conventional commit type box here and add the label to the github PR.

triton-inference-server / server