Re: [Xen-devel] [PATCH] dump_p2m_table: For IOMMU

[Santosh Jodh <Santosh.Jodh@xxxxxxxxxx>]

Why do you do this differently than for VT-d here? There you don't check 
next_table_maddr (and I see no reason you would need to). Oh, I see, there's a 
similar check in a different place there. But this needs to be functionally 
similar here then.
Specifically, ...

> +        {
> +            amd_dump_p2m_table_level(
> +                maddr_to_page(next_table_maddr), level - 1, 
> +                address, indent + 1);
> +        }
> +        else

... you'd get into the else's body if next_table_maddr was zero, which is wrong 
afaict. So I think flow like

    if ( next_level )
        print
    else if ( next_table_maddr )
        recurse

would be the preferable way to go if you feel that these zero checks are 
necessary (and if you do then, because this being the case is really a bug, 
this shouldn't go through silently).
[Santosh Jodh] I was basing my code on existing code in the individual files. I 
was just being paranoid as this is debug code and I would not want to crash the 
system. Anyway, I am resending a patch that structures the code in the same way 
for both Intel and AMD.

> +        {
> +            int i;
> +
> +            for ( i = 0; i < indent; i++ )
> +                printk("  ");

printk("%*s...", indent, "", ...);
[Santosh Jodh] Cool - got it.



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel