News for hackers. Stuff that matters.

A comment on Avoid speculative indirect calls in kernel.

Re:Avoid speculative indirect calls in kernel by dexwiz on Wednesday January 03, 2018 @08:18PM (#16067157)
So is there no way to disable the performance crippling fix? If you ran an air gapped or highly controlled environment wouldn’t you want th option of disabling this to keep performance? How many Giga Terra or Exo FLops will be lost to this fix in super computers? Or is the way to disable it just not merge this patch into your custom kernel code?
[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by Decabytes on Wednesday January 03, 2018 @08:56PM (#16067360)
Correct me if I'm wrong, but I don't think super computers are updating there kernels at all.
[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by striking on Wednesday January 03, 2018 @11:48PM (#16068117)
New software is written with newer kernels in mind. To continue running newer software, you will need to update your kernels.

Additionally: you might update infrequently when it comes to feature releases, but you'll definitely need security patches.

[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by ams6110 on Wednesday January 03, 2018 @09:41PM (#16067560)
They do, but not often. Maybe once a year or so.
[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by zik on Wednesday January 03, 2018 @08:25PM (#16067205)
It can be controlled by a boot option in linux.
[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by userbinator on Wednesday January 03, 2018 @08:28PM (#16067214)
Not this one, which basically penalises every indirect call by preventing them from being speculatively executed.
[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by wilun on Wednesday January 03, 2018 @10:09PM (#16067718)
I don't see why the text could not be fixed at boot time. It has been done before.
[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by jtc331 on Wednesday January 03, 2018 @11:37PM (#16068072)
Because this actually changes how every indirect call is compiled.
[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by wilun on Thursday January 04, 2018 @08:34AM (#16070196)
So what? IIRC we used to have (and maybe still have) a feature that patched out spinlocks in binary code when it detects it is run on monoprocessor systems. You "just" have to identify all the call sites, and patch the code there during the boot process.
[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by revelation on Wednesday January 03, 2018 @08:25PM (#16067203)
This isn't the fix with the 5% quoted performance loss (that one clears all of TLB on every syscall), I don't think this is as costly. But it's certainly not free and it's certainly not clean.

I'm disappointed Intel didn't bother to test the performance impact of this and report it in the patchset. Benchmarks are ordinarily expected with something like this (very hot paths they are changing), and they did have 6 months..

[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by PhantomGremlin on Wednesday January 03, 2018 @10:30PM (#16067823)
I'm disappointed Intel didn't bother to test the performance impact of this and report it in the patchset.

You should be disappointed that they didn't report it. But I'd be willing to bet every last dollar of my net worth that Intel has done very extensive performance testing.

I'd venture that the results are far too embarrassing to report.

[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by barkingcat on Thursday January 04, 2018 @02:52PM (#16073446)
Yup this for sure. They probably ran the performance test bench thousands of times, and only then released it with the PR department fully briefed about how bad it was going to be.
[ Parent | on HN ]
Re:Avoid speculative indirect calls in kernel by userbinator on Wednesday January 03, 2018 @09:31PM (#16067513)
This is worse, far worse --- it basically disables prediction for indirect jumps and calls, plus adds another few extra clock cycles to each one of those for the "retoline" instruction sequence.
[ Parent | on HN ]