Re: [MDADM PATCH 1/1] Array size is wrong when run mdadm -E

[Xiao Ni <xni@redhat.com>]

----- Original Message -----
> From: "Guoqing Jiang" <gqjiang@suse.com>
> To: "Xiao Ni" <xni@redhat.com>
> Cc: linux-raid@vger.kernel.org, ncroxon@redhat.com, "jes sorensen" <jes.sorensen@gmail.com>
> Sent: Wednesday, July 5, 2017 9:56:35 AM
> Subject: Re: [MDADM PATCH 1/1] Array size is wrong when run mdadm -E
> 
> 
> 
> On 07/04/2017 05:55 PM, Xiao Ni wrote:
> >
> > ----- Original Message -----
> >> From: "Guoqing Jiang" <gqjiang@suse.com>
> >> To: "Xiao Ni" <xni@redhat.com>, linux-raid@vger.kernel.org
> >> Cc: ncroxon@redhat.com, "jes sorensen" <jes.sorensen@gmail.com>
> >> Sent: Tuesday, July 4, 2017 5:03:30 PM
> >> Subject: Re: [MDADM PATCH 1/1] Array size is wrong when run mdadm -E
> >>
> >>
> >>
> >> On 07/04/2017 02:50 PM, Xiao Ni wrote:
> >>> The size of array shows wrongly. It shifs wrong number.
> >>>
> >>> Signed-off-by: Xiao Ni <xni@redhat.com>
> >>> ---
> >>>    super1.c | 2 +-
> >>>    1 file changed, 1 insertion(+), 1 deletion(-)
> >>>
> >>> diff --git a/super1.c b/super1.c
> >>> index 86ec850..f3b864d 100644
> >>> --- a/super1.c
> >>> +++ b/super1.c
> >>> @@ -366,7 +366,7 @@ static void examine_super1(struct supertype *st, char
> >>> *homehost)
> >>>    		}
> >>>    		if (ddsks) {
> >>>    			long long asize = __le64_to_cpu(sb->size);
> >>> -			asize = (asize << 9) * ddsks / ddsks_denom;
> >>> +			asize = (asize << 10) * ddsks / ddsks_denom;
> >>>    			printf("     Array Size : %llu%s\n",
> >>>    			       asize >> 10,  human_size(asize));
> >>>    		}
> >> IIRC, sb->size represents the num of sectors, so the shift should be 9.
> > Yes, it should be 9. It should pass bytes to human_size. Thanks for this.
> > It was introduced in d4633e06dfc01723911627fcb104af2ffb6f6a95. So there is
> > the same problem in super0. Is it ok:
> >
> > diff --git a/super0.c b/super0.c
> > index 756cab5..4b4680c 100644
> > --- a/super0.c
> > +++ b/super0.c
> > @@ -131,9 +131,9 @@ static void examine_super0(struct supertype *st, char
> > *homehost)
> >                  }
> >                  if (ddsks) {
> >                          long long asize = sb->size;
> > -                       asize = (asize << 10) * ddsks / ddsks_denom;
> > +                       asize = (asize << 9) * ddsks / ddsks_denom;
> >                          printf("     Array Size : %llu%s\n",
> > -                              asize >> 10,  human_size(asize));
> > +                              asize >> 9,  human_size(asize));
> >                  }
> 
> No, for 0.9 metadata, sb->size means KB not sector.
> 
Thanks for this.

> >          }
> >          printf("   Raid Devices : %d\n", sb->raid_disks);
> > diff --git a/super1.c b/super1.c
> > index 86ec850..2bb1454 100644
> > --- a/super1.c
> > +++ b/super1.c
> > @@ -368,7 +368,7 @@ static void examine_super1(struct supertype *st, char
> > *homehost)
> >                          long long asize = __le64_to_cpu(sb->size);
> >                          asize = (asize << 9) * ddsks / ddsks_denom;
> >                          printf("     Array Size : %llu%s\n",
> > -                              asize >> 10,  human_size(asize));
> > +                              asize >> 9,  human_size(asize));
> >                  }
> >                  if (sb->size != sb->data_size)
> >                          printf("  Used Dev Size : %llu%s\n",
> 
> It is better that you can provide the output of "mdadm -E DISK" etc to prove
> current code is wrong, at least I can see correct output from my side.

[root@dell-per210-01 mdadm]# mdadm -E /dev/loop0 
/dev/loop0:
          Magic : a92b4efc
        Version : 1.2
    Feature Map : 0x0
     Array UUID : 34f18365:ebf46416:ae1f365d:71cc579c
           Name : 0
  Creation Time : Tue Jul  4 02:09:03 2017
     Raid Level : raid5
   Raid Devices : 3

 Avail Dev Size : 8261632 (3.94 GiB 4.23 GB)
     Array Size : 6291456 (6.00 GiB 6.44 GB)
  Used Dev Size : 6291456 (3.00 GiB 3.22 GB)

Ah, I know it wants to show Array Size in KB. It's a raid5 with 3 loop devices. 
I thought they are all sectors subconsciously. But it's better to show sizes with 
the same unit. Right?

Regards
Xiao