Turbo Boost Max 3.0 support

Latest response

How long does it typically take RHEL to incorporate new kernel fixes into their release?

I'm not sure why Intel only appear to support TBM 3.0 in windows!

We have seen some interesting issues with i9 performance on RHEL.

Support for TBM 3.0 seems to be enabled in kernel 4.10 upwards.

I've read that more fixes are in kernel 4.15:
https://www.phoronix.com/scan.php?page=news_item&px=Linux-4.15-TBM3-Skylake-Fix

Any ideas how long RHEL will take to support?

Matt McNally's picture

Responses

If you need functionality from the upstream kernel my suggestion would be to raise a support ticket requesting that the features be backported to the current RHEL kernel.

It depends on the particular feature. Some things we proactively backport because we believe they are useful, or because another feature or product needs them. Some things we will not backport unless someone asks for it.

For this particular feature, you're in luck! We have already have pulled it in for RHEL 7.5 via Red Hat Private Bug 1465349 - [Intel 7.5 Feat] Favored Cores / Asymmetric Core Turbo / Intel Turbo Boost Max Technology 3.0 so you'll see it in our kernel in Q2 2018.

Jamie,

Interested in the process here. If it's already pulled in for 7.5 (which is 6+ months away), is there any possibility it will be pulled in to the 7.4 release before the 7.5 release? Does the fact that the work has been done increase the possibility of it being included in 7.4 if it is specifically requested by a customer?

Again it depends on the feature :)

We usually only do feature enhancement and hardware enablement on synchronous errata, meaning minor releases like RHEL 7.4, RHEL 7.5, and so on.

Asynchronous errata, meaning the package updates between those minor releases, are usually for bugfix errata and security errata only. It's really rare to see new features added here, and definitely no large features.

This is described on the product lifecycle page at "Production Phases - Production 1 Phase" underneath the red and blue table:

If available, new or improved hardware enablement and select enhanced software functionality may be provided at the discretion of Red Hat, generally in minor releases. Hardware enablement that does not require substantial software changes may be provided independent from minor releases at Red Hat's discretion.

The above CPU feature bug has a giant laundry list of individual patches which I guess would mean "substantial software changes" so I don't think this particular TBM feature would meet criteria for 7.4.z release.

To give a comparison for something minor, if a new piece of hardware just needs an additional PCI ID defined in the driver and no other changes, that may be considered for asynchronous errata but we'd want a good business case as to why that's necessary to do right now. Specifically for drivers, there are other means to deliver things there such as the Driver Update Program.

Was this helpful?

We appreciate your feedback. Leave a comment if you would like to provide more detail.
It looks like we have some work to do. Leave a comment to let us know how we could improve.
Close

Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.