The Huge Disaster Within The Linux 2.6.35 Kernel (phoronix.com)

Some performance benchmarks have dropped significantly, for a few days, recently, due to a few commits, in a codebase that is under version control (git), in an unreleased branch, and so presumeably can be backed out, cherry-picked, etc.

Calling this a disaster I think is hyperbole. Because unless there's some incredible benefit to the commits in question, distinct from the negative performance impact, that outweighs the performance impact, I'm sure those changes will either be backed out or reimplemented.

Also, the OA's choice of title specifically feels like hyped-up link bait.

There's an identified regression in a kernel weeks before even the first release candidate and months before it is shipped to any actual users. A "huge disaster" indeed.

What a bs article, the kernel is not even released, the whole point of having release candidates is to weed out stuff like this.

Previous kernels have had similar issues pre-release and during RC1 versions, all of those were mitigated by the time the kernel was actually released.

Using this for production right now or claiming that the released version will have these faults is just dumb.

Why don't they file a bug report or send in a diff instead of writing articles like this.

Looks more like a Huge Disaster for the credibility of Phoronix and Michael Larabel to me.

Oh no, another phoronix ridiculous benchmark. Won't you keep quiet sometimes? :)

What a great set of measurements. Without being able to test the performance of the kernel, such a problem might really develop into a disaster.

Thanks for the effort you have obviously invested to get these results and best of luck to the kernel developers to narrow down the issue.