DAOS-0000 placement: Introduce O(1) fast path for massive GX objects

[wangshilong]

Problem:
The traditional jump hash placement algorithm suffers severe performance degradation when generating layouts for Object Classes with a massive number of groups (e.g., GX class). For instance, in a 500-node cluster using EC16P3GX (~16K targets), the group count (grp_nr) easily exceeds
800. During layout generation, the first ~30 groups quickly exhaust all domain usage bitsets (dom_used). The remaining ~770 groups consistently collide in d_hash_jump(), forcing the algorithm into slow O(D) fallback loops (dom_isset_2ranges). This generates tens of millions of inner-loop bitmap checks (O(GSD)), severely blocking CPU and causing unacceptable latency spikes during object creation and rebuild layout mapping.

Solution:
Introduce a fast path optimization specifically designed for objects with heavy group counts, guarded by a new pool layout version to guarantee strict backward compatibility.

1. Bump DAOS_POOL_OBJ_VERSION to 3 (DAOS_POOL_OBJ_VERSION_3).
2. In get_object_layout(), detect if layout_ver >= 3 and group counts are sufficiently large (jmop_grp_nr >= jmop_dom_nr * 1.5).
3. Under these conditions, bypass the standard non-leaf hash-collision routines by heavily leaning into get_object_layout_gx_fast().
4. The GX fast path uses an OID-seeded Fisher-Yates algorithm to pre-shuffle domains, dealing out targets to shards sequentially.

By reducing traditional collisions, the time complexity drops significantly from O(G * S * D) down to near O(D + G * S), eliminating the massive CPU stalls and providing an estimated 500x-3000x speedup for GX placement on large-scale clusters. Legacy pools (layout <= v2) naturally bypass this avoiding any unexpected layout shifts.

Steps for the author:

• Commit message follows the guidelines.
• Appropriate Features or Test-tag pragmas were used.
• Appropriate Functional Test Stages were run.
• At least two positive code reviews including at least one code owner from each category referenced in the PR.
• Testing is complete. If necessary, forced-landing label added and a reason added in a comment.

After all prior steps are complete:

• Gatekeeper requested (daos-gatekeeper added as a reviewer).

[github-actions]

Errors are Unable to load ticket data
https://daosio.atlassian.net/browse/DAOS-0000

[daosbuild3]

Test stage Functional on EL 8.8 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17667/1/execution/node/1073/log