From: Olaf Hering initramfs can not be used in current 2.6 kernels, the files will never be executed because prepare_namespace doesn't care about them. The only way to workaround that limitation is a root=0:0 cmdline option to force rootfs as root filesystem. This will break further booting because rootfs is not the final root filesystem. This patch checks for the presence of /init which comes from the cpio archive (and thats the only way to store files into the rootfs). This binary/script has to do all the work of prepare_namespace(). --- 25-akpm/Documentation/early-userspace/README | 26 ++++++++++++++++++++++++++ 25-akpm/init/main.c | 7 +++++++ 2 files changed, 33 insertions(+) diff -puN Documentation/early-userspace/README~initramfs-search-for-init-orig Documentation/early-userspace/README --- 25/Documentation/early-userspace/README~initramfs-search-for-init-orig 2004-03-22 00:38:02.064936960 -0800 +++ 25-akpm/Documentation/early-userspace/README 2004-03-22 00:38:02.067936504 -0800 @@ -71,5 +71,31 @@ custom initramfs images that meet your n For questions and help, you can sign up for the early userspace mailing list at http://www.zytor.com/mailman/listinfo/klibc +How does it work? +================= + +The kernel has currently 3 ways to mount the root filesystem: + +a) all required device and filesystem drivers compiled into the kernel, no + initrd. init/main.c:init() will call prepare_namespace() to mount the + final root filesystem, based on the root= option and optional init= to run + some other init binary than listed at the end of init/main.c:init(). + +b) some device and filesystem drivers built as modules and stored in an + initrd. The initrd must contain a binary '/linuxrc' which is supposed to + load these driver modules. It is also possible to mount the final root + filesystem via linuxrc and use the pivot_root syscall. The initrd is + mounted and executed via prepare_namespace(). + +c) using initramfs. The call to prepare_namespace() must be skipped. + This means that a binary must do all the work. Said binary can be stored + into initramfs either via modifying usr/gen_init_cpio.c or via the new + initrd format, an cpio archive. It must be called "/init". This binary + is responsible to do all the things prepare_namespace() would do. + + To remain backwards compatibility, the /init binary will only run if it + comes via an initramfs cpio archive. If this is not the case, + init/main.c:init() will run prepare_namespace() to mount the final root + and exec one of the predefined init binaries. Bryan O'Sullivan diff -puN init/main.c~initramfs-search-for-init-orig init/main.c --- 25/init/main.c~initramfs-search-for-init-orig 2004-03-22 00:38:02.065936808 -0800 +++ 25-akpm/init/main.c 2004-03-22 00:38:02.068936352 -0800 @@ -604,6 +604,13 @@ static int init(void * unused) smp_init(); do_basic_setup(); + /* + * check if there is an early userspace init, if yes + * let it do all the work + */ + if (sys_access("/init", 0) == 0) + execute_command = "/init"; + else prepare_namespace(); /* _