linux-kernel - Re: [PATCH 2/4] nvme-multipath: add the NVME_MULTIPATH

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <Z9hjMtuyd6Y0FJPy@kbusch-mbp.dhcp.thefacebook.com>
Date: Mon, 17 Mar 2025 12:00:18 -0600
From: Keith Busch <kbusch@...nel.org>
To: John Meneghini <jmeneghi@...hat.com>
Cc: Christoph Hellwig <hch@....de>, sagi@...mberg.me, loberman@...hat.com,
	linux-nvme@...ts.infradead.org, linux-kernel@...r.kernel.org,
	emilne@...hat.com, bgurney@...hat.com
Subject: Re: [PATCH 2/4] nvme-multipath: add the NVME_MULTIPATH_PARAM config
 option

On Thu, Mar 13, 2025 at 05:46:25PM -0400, John Meneghini wrote:
> Our RHEL customers do this all of the time to enable DMMP with their NVMe devices. This has caused tremendous
> confusion and turmoil with our RHEL customers who keep on turning nvme-multipath off even when we ship
> RHEL with nvme-multipath on and we tell them turning it off is not supported.
> 
> I finally decided that there is no good way to control this situation w/out removing the
> core.nvme_multipath parameter while CONFIG_NVME_MULTIPATH is enabled.
> 
> This patch enables distros like RHEL simply config out the parameter w/out impacting all of the people out
> there who currently want and use the run time switch when CONFIG_NVME_MULTIPATH is enabled.
> 
> Please let me know what I need to do to make these patches less confusing and get them accepted upstream.

It'd be one thing if it was just me lamenting the loss of the parameter,
but Hannes has customers with a very real and obvious use case for it,
even if Christoph hasn't encountered many (or any?) devices that fall
into that trap. So, from my perspective, I really think pcie hotplug
handling needs to be sorted out before we can just force these to
subscribe to the native mulitpath virtual device hierarchy.

In the meantime, I know RHEL has done out of tree patches for many
reasons, so it shouldn't be a big deal to change the module_param type
from "bool" to "bool_enable_only" in your kernel releases.