On Wed, Oct 23, 2013 at 12:11:57PM -0500, Alex Villacís Lasso wrote:
How do I peek into the initramfs memory from within the Linux
system, in order to check the actual contents to see where the
corruption is? Is there any block device to check?
It doesn't work like this. When the kernel boots, it locates the
initrd and then parses it, creating files and directories as it
parses, in a tmpfs-like in-memory filesystem from the (optionally
gzipped-) cpio data. There is no block device backing it.
Because it is using gzip and cpio, the kernel is able to verify the
data looks sane as it goes along, hence the error message.
Note that recent kernels support xz compression. Not sure about your
kernel, but it might help you squeeze a little bit more into the
You could also try to add some debugging to init/initramfs.c (eg. in
the 'do_name', 'do_copy' and 'do_symlink' functions).
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers. Over 100 libraries supported.