Re: [patch 00/11] rseq: Optimize exit to user space

[Jens Axboe]

On 8/13/25 3:32 PM, Thomas Gleixner wrote:
> On Wed, Aug 13 2025 at 11:45, Jens Axboe wrote:
>> On 8/13/25 10:29 AM, Thomas Gleixner wrote:
>>>
>>> suffered badly from the stale eventmask bits and the cleaned up version
>>> brought a whopping 25+% performance gain.
>>
>> Thanks for doing this work, it's been on my list to take a look at rseq
>> as it's quite the pig currently and enabled by default (with what I
>> assume is from a newer libc).
> 
> Yes, recent glibc uses it.
> 
> Could you give it a test ride to see whether this makes a difference in
> your environment?

Yep, I'll give a spin. Won't be before Monday as I'm out of town thu/fri
this week, just as a heads-up.

If I recall correct, I'd see rseq at 2-3% of overall CPU time last I
tested. I'll try and do some pre/post numbers for some of the stuff I
usually run.

-- 
Jens Axboe