[lvm-devel] Re: [PATCH]: unify logical volume structure initialization

Mikulas Patocka mpatocka at redhat.com
Mon Sep 28 18:23:20 UTC 2009


On Mon, 28 Sep 2009, Alasdair G Kergon wrote:

> > > Already initialised AFAICT.
> > No, it isn't ... if you look up at this function (import_pool_lvs), 
> > there's: lv = dm_pool_zalloc(mem, sizeof(*lv)
>  
> After applying the patch I got 2 x lv_init() on the same lv and so I
> removed the second one.  The pool code began as a copy of the format1 code
> so I'm not sure how the duplicate initialisation crept in.

OK, I see it. Fine to remove the second one.

Mikulas

> > The requirement is this: there must be a central point where the lv 
> > entries are initialized. So that if someone adds new entries, he writes 
> > the initialization only in one place. If you want to unify allocation and 
> > initialization, it is possible, just do it (but it requires somehow bigger 
> > code change than this patch).
>  
> Indeed - that's what lv_create_empty() was meant to be (and what new code
> should be trying to use).
> 
> Alasdair
> 




More information about the lvm-devel mailing list