docs: add a faq on EROFS image sizes

It's frequently asked..

Signed-off-by: Gao Xiang <xiang@kernel.org>

src/conf.py

@@ -22,6 +22,7 @@
 ]
 
 myst_enable_extensions = ["colon_fence"]
+myst_heading_anchors = 3
 
 templates_path = ['_templates']
 

src/faq.md

@@ -1,3 +1,60 @@
 # 🙋 Frequently Asked Questions
 
-Under construction..
+## Why are images packaged in EROFS larger than those with SquashFS?
+
+First of all, the initial target use cases of EROFS are _high-performance
+embedded scenarios, such as smartphones powered by Android_.  Runtime
+performance is always the top priority for EROFS (or, systems and applications
+will be lagged), even if it means sacrificing some ultra-space savings to avoid
+significant performance regressions against uncompressed approaches.
+
+However, EROFS has landed **several new ondisk features** to narrow the slight
+size difference with SquashFS.  When comparing, please ensure the same
+configuration is used:
+
+ - **Compression algorithm (if data is compressed)**: EROFS uses *LZ4* by
+   default due to lowest decompression latencies among popular open-source
+   algorithms, while SquashFS uses *GZIP* instead;
+
+ - **Compressed extent size**: Almost all filesystems that natively support
+   compression typically cut data into compressed extents for random access.
+   EROFS focuses on smaller physical clusters to maximize random performance and
+   use *block-sized physical clusters* by default (usually 4 KiB), whileas
+   SquashFS uses *128 KiB*.  It can be adjusted using the `-C` option with
+   `mkfs.erofs`;
+
+ - **Compression level**: For example, EROFS uses *LZ4HC_CLEVEL_DEFAULT (level 9)*
+   for LZ4 HC, whereas SquashFS often uses
+   *[LZ4HC_CLEVEL_MAX (level 12)](https://github.com/plougher/squashfs-tools/blob/4.6.1/squashfs-tools/lz4_wrapper.c#L52)*;
+
+ - **Advanced features**: Enable the `-Efragments` option for EROFS when
+   comparing with SquashFS.
+
+In addition, EROFS may produce larger images due to the following differences:
+ - **Inode size**: EROFS uses 64-byte ondisk inodes (extended inodes) to support
+   nanosecond timestamps, whereas SquashFS often uses
+   [32-byte inodes with only 4-byte timestamps](https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/fs/squashfs/squashfs_fs.h?h=v6.12#n334)
+   for regular files;  Consider switching to 32-byte EROFS compact inodes (e.g.,
+   by using `-T`) if per-file timestamps are not a strong requirement;
+
+ - **Lack of metadata compression**: If your testsets contain a large number of
+   files, EROFS may result in larger images compared to SquashFS because of
+   metadata compression is not supported.  Again, it isn't considered at first
+   due to bad impacts to random metadata performance but it may be implemented
+   in the future.
+
+ - **File-based deduplication**: SquashFS deduplicates files with identical data
+   by default (it can be disabled with `-no-duplicates`), whereas EROFS does not
+   (except for hardlinks).  However, EROFS offers finer-grained data
+   deduplication using `-Ededupe`.
+
+However, note that SquashFS doesn't always outperform EROFS in image size
+either.  EROFS images are often significantly smaller (while still offering
+better runtime performance) when compressing files in small compressed extent
+sizes.
+
+Additionally, EROFS has supported CDC-like [compressed data deduplication](design.md#data-deduplication)
+since Linux 6.1, which gives extra space saving too.  Please make sure that
+the options `-Ededupe` and `-Eall-fragments` are specified with `mkfs.erofs`.
+
+## 🚧 Under construction..