[dm-devel] [PATCH] fix extended partition mapping

[Christophe Varoqui]

On sam., 2013-01-05 at 21:57 -0500, Phillip Susi wrote:
> The linux kernel maps the extended partition only
> so that LILO can be installed there.  The length is always set
> to two sectors to allow this, and most tools know to ignore the
> device.  kpartx was mapping the entire extended partition, then
> stacking the logical partitions on top of it.  This presented
> a device that looked like an entirely separate disk that
> contains only the logical partitions.  This patch fixes kpartx
> to conform with the normal Linux behavior.

Applied.
Thanks.

> ---
>  kpartx/dos.c    |    2 ++
>  kpartx/kpartx.c |    9 +++------
>  2 files changed, 5 insertions(+), 6 deletions(-)
> 
> diff --git a/kpartx/dos.c b/kpartx/dos.c
> index 1691105..a1a9961 100644
> --- a/kpartx/dos.c
> +++ b/kpartx/dos.c
> @@ -98,6 +98,8 @@ read_dos_pt(int fd, struct slice all, struct slice *sp, int ns) {
>  			break;
>  		}
>  		if (is_extended(p.sys_type)) {
> +			sp[i].size = 2; /* extended partitions only get two
> +					   sectors mapped for LILO to install */
>  			n += read_extended_partition(fd, &p, i, sp+n, ns-n);
>  		}
>  	}
> diff --git a/kpartx/kpartx.c b/kpartx/kpartx.c
> index b5e0a32..f359afe 100644
> --- a/kpartx/kpartx.c
> +++ b/kpartx/kpartx.c
> @@ -515,7 +515,6 @@ main(int argc, char **argv){
>  			d = c;
>  			while (c) {
>  				for (j = 0; j < n; j++) {
> -					uint64_t start;
>  					int k = slices[j].container - 1;
>  
>  					if (slices[j].size == 0)
> @@ -541,11 +540,9 @@ main(int argc, char **argv){
>  					}
>  					strip_slash(partname);
>  
> -					start = slices[j].start - slices[k].start;
> -					if (safe_sprintf(params, "%d:%d %" PRIu64,
> -							 slices[k].major,
> -							 slices[k].minor,
> -							 start)) {
> +					if (safe_sprintf(params, "%s %" PRIu64,
> +							 device,
> +							 slices[j].start)) {
>  						fprintf(stderr, "params too small\n");
>  						exit(1);
>  					}