Hi Hannes, Thanks for reviewing. On 12/14/2016 05:04 AM, Hannes Reinecke wrote:
So wouldn't it make more sense to move filter_devnode() into pathinfo(), too, to avoid further inconsistencies between _filter_path() and pathinfo()?
Agree.
Especially as it looks that if we need to call filter_devnode() in get_refwwid(), too; starting with line 976 we're just calling 'store_pathinfo' for the device node, with no check for blacklisted devnode at all. (Which also goes to explain why I have this mysterious bug where blacklisting by device node doesn't properly work ...). So moving filter_devnode() in pathinfo would be more sensible and clean up the overall programming model.
I see. Ok, changed/submitted v4. Hope it helps w/ that mysterious bug. -- Mauricio Faria de Oliveira IBM Linux Technology Center -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
