Bug 2054443 - NVMe/TCP remove Tech Preview designation
Summary: NVMe/TCP remove Tech Preview designation
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Red Hat Enterprise Linux 8
Classification: Red Hat
Component: kernel
Version: 8.6
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: rc
: ---
Assignee: John Meneghini
QA Contact: Marco Patalano
Gabriela Nečasová
URL:
Whiteboard:
Depends On: 2057517 2078806
Blocks:
TreeView+ depends on / blocked
 
Reported: 2022-02-15 00:29 UTC by Chris Leech
Modified: 2023-01-18 14:06 UTC (History)
7 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of: 2054441
Environment:
Last Closed: 2022-03-14 20:08:49 UTC
Type: Bug
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Red Hat Issue Tracker RHELPLAN-112264 0 None None None 2022-02-15 00:35:51 UTC

Description Chris Leech 2022-02-15 00:29:35 UTC
+++ This bug was initially created as a clone of Bug #2054441 +++

Description of problem:

We would like to consider moving the nvme-tcp driver out of tech preview and into full support.

Additional info:

Testing confidence of base nvme-tcp functionality within Red Hat has increased greatly over the past few month.  We now have partner storage targets being tested against, and have fixed the known issues found with our current level of testing that would prevent moving to full support.

We would like to postpone this decision to after dev freeze as an exception based on partner testing and feedback.  That way partners can test with the first rc build after we finish integrating final bug fixes, and provide feedback based on that.  If no blocker issues are found with partner testing, and we have a storage partner ready to move this technology into full support with us, then we will remove the tech preview status from the nvme-tcp driver.

This is for the current NVMe/TCP host basic feature set.  Additional functionality is being targeted for later releases.  If no storage partners are willing to fully support this technology as is, then we can remain in a tech preview state until such a time as both functionality and stability meet the requirements for production use.

--- Additional comment from RHEL Program Management on 2022-02-15 00:25:22 UTC ---

A request has been made to complete this BZ as an exception (that is, after the deadline).  To facilitate review, the Product Owner (or a delegate) must provide an impact statement.  This is done by clicking the [reply] link on this comment, and then replying in-line to this message. 
 This must be done to facilitate review, even if the case appears to be obvious, or is already covered elsewhere in the BZ. The exception will not be reviewed until this is done.

======= Impact Statement ======= 

What is the benefit of making this change after the deadline? What is the impact on customer satisfaction, and on the business?

a. For bugfixes, there must be support from someone in Customer Support, a Partner Manager, Product Manager, and/or a Business Unit rep. (e.g., potentially from a layered product BU). 
b. It is unusual for new features to be allowed in the release after the deadline. For this reason, additional justification must be obtained from a stakeholder outside of the SST, to ensure an objective review of the benefit to the product. 

What is the risk to the release schedule, quality, and the impact of diverting resources from other efforts? Will there be enough time to do the necessary large-scale, cross-function, regression, stress, or fault-insertion testing that may be required to verify this change? 

============================= 

In addition to the impact statement above, check the BZ to confirm the following: 

1. The Internal Target Release (ITR) field reflects the correct release.
2. Set the Internal Target Milestone (ITM) field to indicate by when the work will be done. The approval for this exception will expire on that milestone. 
3. Ensure qa_ack+ and devel_ack+ are set, and that those acks are not leftover from before the exception request.
4. Prepare a RHEL rpm scratch build and have this change validated on the latest RHEL milestone compose by someone other than the developer. A comment must be added to the Bugzilla indicating the validation is successful and there were no observed regressions.

All of these steps must be complete before this change is reviewed as an exception.

Comment 5 John Meneghini 2022-03-03 16:15:59 UTC
Removing Tech Preview in RHEL 8.6 will depend upon the outcome of BZ2057517

Comment 7 John Meneghini 2022-03-09 14:37:13 UTC
Note: this BZ should be acked+ and tested by QE, but the MR is currently NACKed and will not merge until BZ 2057517 is resolved.

Comment 8 Marco Patalano 2022-03-09 14:41:09 UTC
With kernel-4.18.0-371.el8:

# modprobe nvme_tcp
Mar  9 09:38:40 storageqe-28 kernel: TECH PREVIEW: NVMe/TCP may not be fully supported.#012Please review provided documentation for limitations.

I then verified the message is no longer present with the test kernel from #6

Sanity check completed:

# git diff linux-4.18.0-371.el8/drivers/nvme/host/tcp.c linux-4.18.0-371.el8.mr2286_220302_0636/drivers/nvme/host/tcp.c
diff --git a/linux-4.18.0-371.el8/drivers/nvme/host/tcp.c b/linux-4.18.0-371.el8.mr2286_220302_0636/drivers/nvme/host/tcp.c
index c24fedc..6553722 100644
--- a/linux-4.18.0-371.el8/drivers/nvme/host/tcp.c
+++ b/linux-4.18.0-371.el8.mr2286_220302_0636/drivers/nvme/host/tcp.c
@@ -2623,8 +2623,6 @@ static int __init nvme_tcp_init_module(void)
                return -ENOMEM;
 
        nvmf_register_transport(&nvme_tcp_transport);
-
-       mark_tech_preview("NVMe/TCP", THIS_MODULE);
        return 0;
 }

Comment 9 John Meneghini 2022-03-14 20:08:49 UTC
After speaking with each of or Storage partners we've decided that we are not going to remove the NVMe/TCP Tech Preview designation in RHEL 8.6.

So this BZ is not needed.  I am closing this BZ.

Comment 10 Mitul 2023-01-18 14:06:01 UTC
Hi,

Found kernel taint 65536 for rhel 8.7

---------------------------------------------
kernel tainted with below value:
65536
---------------------------------------------
---------------------------------------------
kernel version:
4.18.0-425.10.1.el8_7.x86_64
---------------------------------------------
---------------------------------------------
Red Hat release:
redhat-release-8.7-0.3.el8.x86_64
---------------------------------------------
---------------------------------------------
Printing the BIOS information:
Getting SMBIOS data from sysfs.
SMBIOS 2.8 present.
---------------------------------------------
--------------------------------------------
Warning found in sosreport:
--------------------------------------------
---------------------------------------------
Error/Failure/Bug found in sosreport:
[    0.282004] acpi PNP0A03:00: _OSC: not requesting OS control; OS requires [ExtendedConfig ASPM ClockPM MSI]
[    0.284011] acpi PNP0A03:00: fail to add MMCONFIG information, can't access extended PCI configuration space under this bridge.
[    0.286279] acpiphp: Slot [3] registered
--
[   11.310583] Installing knfsd (copyright (C) 1996 okir.de).
[   11.354684] svc: failed to register nfsdv3 RPC service (errno 111).
[   11.355911] svc: failed to register nfsaclv3 RPC service (errno 111).
[   76.898378] systemd[1]: Stopping Journal Service...
---------------------------------------------
---------------------------------------------
Tech Preview found in sosreport:
[    2.910683] TECH PREVIEW: NVMe/TCP may not be fully supported.
---------------------------------------------

X 	
With the TAINT_TECH_PREVIEW mask value being removed upstream and within RHEL 9 the TAINT_AUX bit is used instead of TAINT_TECH_PREVIEW within 8.6 and later kernels.


Note You need to log in before you can comment on or make changes to this bug.