[ofa-general] Re: [PATCH] IB/core user memory registrations

[Eli Cohen]

On Wed, 2007-05-09 at 20:00 -0700, Roland Dreier wrote:
> > @@ -1020,7 +1020,7 @@ static struct ib_mr *mthca_reg_user_mr(s
>  >  	int shift, n, len;
>  >  	int i, j, k;
>  >  	int err = 0;
>  > -	int write_mtt_size;
>  > +	int write_mtt_size = mthca_write_mtt_size(dev);
>  >  
>  >  	mr = kmalloc(sizeof *mr, GFP_KERNEL);
>  >  	if (!mr)
> 
> Not sure I understand what this is fixing... can you be more explicit?
> As far as I can see, the first use of write_mtt_size in mthca_reg_user_mr()
> is the line
> 
> 	write_mtt_size = min(mthca_write_mtt_size(dev), (int) (PAGE_SIZE / sizeof *pages));
> 
> so I'm not sure why we need another initialization?  (I'm looking at
> Linus's latest tree, which contains the mlx4 merge)
> 
>  - R.

This initialization was not in the tree I was working on. Now I see it
is fixed in the updated tree. Thanks.