Re: [Xen-devel] [PATCH 1/2] tools/xenstore: Do not abort xenstore-ls if a node disappears while iterating

[Ian Jackson <ian.jackson@xxxxxxxxxx>]

Jürgen Groß writes ("Re: [Xen-devel] [PATCH 1/2] tools/xenstore: Do not abort 
xenstore-ls if a node disappears while iterating"):
> On 20.03.20 15:58, David Woodhouse wrote:
> > But it does still happen. And even if I turn the errx() into a warn()
> > to stop it aborting, and add a warn() when the xs_transaction_end()
> > returns EAGAIN... that isn't happening either. I'm just getting
> > inconsistent data, within a transaction.
> 
> Hmm, yes, thinking more about it: a non-transactional write of a node
> which hasn't been written or read by an ongoing transaction is not
> handled in a special way. This could be changed, but would require some
> structural changes.

And making a node visible by XS_DIRECTORY[_PART] doesn't count as
reading it.  But it does count as reading the parent ?
In principle adding or removing a node could be made to count as a
change to the containing directory.  But I don't think doing this as a
response to David's issue is sensible.

Ian.