Re: [PATCH] Fix initramfs size calculation

[Sam Ravnborg <sam@ravnborg.org>]

On Wed, Aug 25, 2010 at 05:57:12PM +0200, Michael Holzheu wrote:
> The size of a built-in initramfs is calculated in init/initramfs.c by 
> "__initramfs_end - __initramfs_start".  Those symbols are defined in the
> linker script include/asm-generic/vmlinux.lds.h:
> 
> #define INIT_RAM_FS                                                     \
>         . = ALIGN(PAGE_SIZE);                                           \
>         VMLINUX_SYMBOL(__initramfs_start) = .;                          \
>         *(.init.ramfs)                                                  \
>         VMLINUX_SYMBOL(__initramfs_end) = .;
> 
> If the initramfs file has an odd number of bytes, the "__initramfs_end"
> symbol points to an odd address, for example, the symbols in the System.map
> might look like:
> 
>     0000000000572000 T __initramfs_start
>     00000000005bcd05 T __initramfs_end	  <-- odd address
> 
> At least on s390 this is a problem:

Another way to fix this could be to align . to an even
address like this:
> #define INIT_RAM_FS                                                     \
>         . = ALIGN(PAGE_SIZE);                                           \
>         VMLINUX_SYMBOL(__initramfs_start) = .;                          \
>         *(.init.ramfs)                                                  \
          . = ALIGN(32);                                                  \
>         VMLINUX_SYMBOL(__initramfs_end) = .;
> 

32 was selected as this is what we will introduce as the default
alignment in linker scripts anyway.

This I guess is a problem we have had some time and a minimal fix is
easier to have backported by the stable team.

> The patch also restructures the "usr/initramfs_data.xxx.S" files to use a
> common macro that includes the (compressed) initramfs file and calculates
> the __initramfs_size.
I like this anyway. Could you do this as a separate patch?

	Sam