docs: improve clarity in xattrs and chunked format documentation

[SKM2227229725]

This PR improves clarity in xattr and chunk-based format documentation.

Changes:

• Added note about reserved e_name_index values (0 and 5)
• Standardized "Bloom filter" capitalization
• Improved wording for long xattr prefix structure
• Clarified chunk size formula expression to avoid ambiguity

These changes are based on discussion in #28.

[Copilot]

Pull request overview

This PR expands and refines the EROFS on-disk format documentation by adding dedicated pages for extended attributes (xattrs) and the chunk-based file layout, and by improving phrasing/consistency in related docs.

Changes:

• Add new xattrs on-disk format documentation, including reserved e_name_index notes and Bloom filter wording.
• Add new chunk-based file format documentation, including clearer chunk size expression and multi-device addressing details.
• Update the on-disk docs index to link these optional-feature pages; minor wording fix in core superblock table.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
src/ondisk/xattrs.md	New xattr format page (superblock/inode fields, inline/shared xattrs, long prefixes, Bloom filter, image-share xattrs).
src/ondisk/chunked_format.md	New chunk-based inode format page (chunk info record, entry formats, device table, address resolution).
src/ondisk/index.md	Adds narrative + toctree entries to surface the new optional-feature docs.
src/ondisk/core_ondisk.md	Improves wording (“48-bit layout specific”) in the superblock table.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Add a blank line before the (shared_xattr_area)= label. As written, the label is immediately adjacent to the preceding paragraph, which can cause MyST/Sphinx to treat it as part of the paragraph instead of a standalone label and may break cross-referencing.

Suggested change

	All other `e_name_index` values (including `0` and `5`) are reserved and must not be used unless defined by a future format extension.
	All other `e_name_index` values (including `0` and `5`) are reserved and must not be used unless defined by a future format extension.

[Copilot]

The e_name_index mapping table is labeled as a "Prefix", but entries system.posix_acl_access and system.posix_acl_default are full xattr names rather than prefixes (they don't end with a dot). Consider renaming the column/wording (e.g. "Prefix or full name") and/or explicitly stating that for these indexes the stored name suffix length (e_name_len) is 0.

Suggested change

	| Value | Prefix |
	|-------|--------|
	| 1 | `user.` |
	| 2 | `system.posix_acl_access` |
	| 3 | `system.posix_acl_default` |
	| 4 | `trusted.` |
	| 6 | `security.` |
	| Value | Prefix or full name |
	|-------|----------------------|
	| 1 | `user.` |
	| 2 | `system.posix_acl_access` |
	| 3 | `system.posix_acl_default` |
	| 4 | `trusted.` |
	| 6 | `security.` |
	For indices `2` and `3`, the string shown above is the entire xattr name (no additional suffix bytes are stored), so `e_name_len` must be `0`. For the other indices, the string is a prefix and the remaining name bytes are stored as the suffix of length `e_name_len`.