On Thu, Nov 29, 2007 at 11:49:17AM -0600, bo y wrote:
> I am very confused about the way kernel mount a disk-based root like
> root=/dev/sda1. Does this special file "/dev/sda1" exist inside the
> very very first filesystem, rootfs? How does the kernel get the major,
> minor number out of this line "root=/dev/sda1"? Otherwise, there is no
> way to reference the actual device, right ? I have browsed the whole
> source tree and found no place where command_line (or cmdline) is
> parsed to convert "root=/dev/xxx" to something like kdev_t or dev_t.
Please, read:
http://www.ibm.com/developerworks/library/l-linuxboot/index.html
Your kernel has care about initrd image only. This RAM disk image is
the initial temporary root filesystem. The real root filesystem is
mounted and pivoted (pivot_root syscall) by a script from the initrd
image.
For example Fedora/RHEL uses nash interpreter in initrd image. This
interpreter parses the root= option and mounts the real root
filesystem.
Karel
--
Karel Zak <kzak@xxxxxxxxxx>
--
To unsubscribe from this list: send an email with
"unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx
Please read the FAQ at http://kernelnewbies.org/FAQ
