[lapack][rocSOLVER] Implement gerqf and getri

[zjin-lcf]

Summary

• implement non-batched gerqf and getri for buffer and USM APIs
• retain checked 32-bit pivot conversion for getri, which has no 64-bit rocSOLVER entry point
• make pivot-conversion, solve, and temporary-storage cleanup ordering explicit
• add the corresponding scratchpad queries

Extracted from #767 for focused review. This branch is stacked on #778; until that PR merges, it intentionally includes the shared foundation commit.

Test plan

• clang-format 19.1 check
• git diff --check
• rocSOLVER backend and RT/CT test binaries build for gfx90a
• focused gerqf tests on MI210 with ONEAPI_DEVICE_SELECTOR=hip:*: 24 passed across RT/CT
• local getri execution is blocked by an assertion in the custom DPC++ runtime's device-kernel registration before the rocSOLVER call; the same assertion reproduces in the existing getrs pivot-conversion path, and CI should exercise this path with the supported toolchain