... path^2.1

/usr/share/man/man8 on Debian

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... supported^3.1

Working with fdisk is out of scope of this document. See man fdisk for full usage documentation

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... dirtied^3.2

Why LFS must know about dirtied pages is described in Section 5.1.3

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... automatically^3.3

Autoloading must be enabled in kernel.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... operations^4.1

Directory operations are file system operations that affect multiple inodes and either all or no changes must take place.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...emergency^5.1

See message LFS_GC_INFO_WORK in the section 5.10.1.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... empty^5.2

See section 5.8.3

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... it^5.3

Either an element of the i_blocks array in inode or of an indirect block.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...mmap^5.4

Even though it is illegal in Linux to mmap a block behind the end of file, a user process can write data to a hole inside a file and thus create a block nevertheless.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... operation^5.5

This operation is provided by David Howell's patch.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... way^5.6

Of course, the block I/O scheduler may queue our BIO requests.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... not^5.7

At the moment there is only one block per page and so the mask always contains 1. However, it is going to be necessary to support file systems with smaller block sizes.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... writing^5.8

Even though this is not a very clean approach, it has been discussed in the Linux-fsdevel mailing list and has been generally regarded as tricky but viable. Our experiments seem to have confirmed this.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... info^5.9

The superblock info contains a pointer to this structure

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...work_queue^5.10

This queue used to contain collect requests as well, hence its name. However, currently it stores outgoing messages only.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...LFS_INODE_UNUSED ^5.11

this macro, as defined in include/linux/lfs_fs.h, is equal to 0 since the glibc library is biased to EXT2 and expects unused inodes to be set to 0.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... function^5.12

copied from EXT3

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... seek^5.13

Seeking is limited to usage of telldir(), seekdir(), rewinddir() and seek() with SEEK_SET only.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... writer^5.14

Access to a directory is exclusive, but readdir() calls can be interleaved with calls by other applications that change content of the directory.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.