[PATCH v5 1/6] nvme: move ns id info to struct nvme_ns_head
Daniel Wagner
dwagner at suse.de
Thu Dec 14 01:31:28 PST 2023
On Wed, Dec 13, 2023 at 04:38:34PM +0100, Christoph Hellwig wrote:
> On Wed, Dec 13, 2023 at 03:54:25PM +0100, Daniel Wagner wrote:
> > > I think that the whole PI stuff needs to be taken with a bit more
> > > consideration because if not all paths agree on the pi (as we have
> > > hbas with fabrics) we can't just override or do a logical or on
> > > the capabilities/attributes.
> >
> > So should the PI variables stay in nvme_ns at this point? Or should I
> > add some checks which avoid an override and warn in this case?
>
> Didn't we merge the patch from max to require uniform PI setups
> for all controllers that we're using in a multipath setup? I'll
> check the code after finishing a few more things if no one remembers
> offhand.
The newest discussion on this topic I found was this:
https://lore.kernel.org/linux-nvme/ec66d848-5246-529a-7050-afca3d1d981f@nvidia.com/
More information about the Linux-nvme
mailing list