Re: [PATCH net 2/2] dpll: report dumps interrupted by topology changes

[Ivan Vecera]

On 11/10/26 23:02, Jakub Kicinski wrote:
> On Thu,  5 Nov 2026 14:11:45 +0100 Ivan Vecera wrote:
> > Pin and device dumps release dpll_lock between parts and resume from the
> > id saved by the previous part. An object registered in between is missed
> > if its id is below that one, as is the case for an object allocated
> > before the dump started. User space gets an incomplete dump and has no
> > way to tell.
>
> This is not really a problem, just something AI loves to complain about.
> IMO dumps must be correct only wrt unmodified objects. If we have
> objects A B C, and A and C either get registered or removed during the
> dump - whether they appear in the dump does not matter. Only B must
> appear. If the IDs are from Xarray they are usually stable, the problems
> with dump consistency mostly come when dumping a list or a hash table.
>
> Please repost with a better commit msg if there's some list / hash table
> here that I'm not spotting. I'll take patch 1 as is.

There is none, both dumps walk an xarray, and the objects they can miss
are the ones registered or unregistered during the dump. I'll drop it.

Thanks for taking patch 1.

Ivan