Follow-up from WT-17788, which converted the cross-checkpoint shared disk cache's hash table from TAILQ to LIST because it never used the tail pointer TAILQ carries.

conn->blockhash (block handle hash table, WT_BLOCK/WT_BLOCK_DISAGG in src/include/block.h) and conn->fhhash (file handle hash table, WT_FH in src/include/os.h) share the same pattern: only TAILQ_FOREACH / TAILQ_INSERT_HEAD / TAILQ_REMOVE are used on the "hashq" field, never a tail-only operation. Both arrays are sized by the connection's "hash.buckets" config (default 512, max 65536), so this is a fixed per-bucket cost independent of how many blocks/files are actually open.

Caveat: WT_BLOCK and WT_BLOCK_DISAGG have a static_assert in src/include/verify_build.h requiring their layouts to match up to and including the hashq field, so both structs' hashq fields must be converted from TAILQ_ENTRY to LIST_ENTRY together to keep that assert valid. The in-memory filesystem's own file handle hash table (im_fs->fhhash in src/os_common/os_fs_inmemory.c, sized by the same conn->hash_size) has the identical pattern and should be converted at the same time.

Requires adding the LIST_* macros to src/include/queue.h if not already present from WT-17788.