Re: [PATCH] Add retain_initrd boot option

[Randy Dunlap <randy.dunlap@oracle.com>]

Michael Neuling wrote:
> Add retain_initrd option to control freeing of initrd memory after
> extraction.  By default, free memory as previously. 
> 
> The first boot will need to hold a copy of the in memory fs for the
> second boot.  This image can be large (much larger than the kernel),
> hence we can save time when the memory loader is slow.  Also, it reduces
> the memory footprint while extracting the first boot since you don't
> need another copy of the fs.
> 
> Signed-off-by: Michael Neuling <mikey@neuling.org>
> ---
> Removed unnecessary init of do_retain_initrd as suggested by Randy
> Dunlap.
> 
>  Documentation/kernel-parameters.txt |    2 ++
>  init/initramfs.c                    |   18 ++++++++++++++++--
>  2 files changed, 18 insertions(+), 2 deletions(-)
> 
> Index: linux-2.6-ozlabs/Documentation/kernel-parameters.txt
> ===================================================================
> --- linux-2.6-ozlabs.orig/Documentation/kernel-parameters.txt
> +++ linux-2.6-ozlabs/Documentation/kernel-parameters.txt
> @@ -1366,6 +1366,8 @@ and is between 256 and 4096 characters. 
>  	resume=		[SWSUSP]
>  			Specify the partition device for software suspend
>  
> +	retain_initrd	[RAM] Keep initrd memory after extraction
> +
>  	rhash_entries=	[KNL,NET]
>  			Set number of hash buckets for route cache
>  
> Index: linux-2.6-ozlabs/init/initramfs.c
> ===================================================================
> --- linux-2.6-ozlabs.orig/init/initramfs.c
> +++ linux-2.6-ozlabs/init/initramfs.c
> @@ -487,6 +487,17 @@ static char * __init unpack_to_rootfs(ch
>  	return message;
>  }
>  
> +static int do_retain_initrd;
> +
> +static int __init retain_initrd_param(char *str)
> +{
> +	if (*str)
> +		return 0;
> +	do_retain_initrd = 1;
> +	return 1;
> +}
> +__setup("retain_initrd", retain_initrd_param);
> +
>  extern char __initramfs_start[], __initramfs_end[];
>  #ifdef CONFIG_BLK_DEV_INITRD
>  #include <linux/initrd.h>
> @@ -494,10 +505,13 @@ extern char __initramfs_start[], __initr
>  
>  static void __init free_initrd(void)
>  {
> -#ifdef CONFIG_KEXEC
>  	unsigned long crashk_start = (unsigned long)__va(crashk_res.start);
>  	unsigned long crashk_end   = (unsigned long)__va(crashk_res.end);

I'm still not seeing how using crashk_res is valid here when
CONFIG_KEXEC=n.  Can you explain that, please?
You did test this with KEXEC=y and KEXEC=n, right?

> +	if (do_retain_initrd)
> +		goto skip;
> +
> +#ifdef CONFIG_KEXEC
>  	/*
>  	 * If the initrd region is overlapped with crashkernel reserved region,
>  	 * free only memory that is not part of crashkernel region.
> @@ -515,7 +529,7 @@ static void __init free_initrd(void)
>  	} else
>  #endif
>  		free_initrd_mem(initrd_start, initrd_end);
> -
> +skip:
>  	initrd_start = 0;
>  	initrd_end = 0;
>  }


-- 
~Randy