[linus:master] [mm/secretmem] 97d34aa65c: stress-ng.secretmem.ops_per_sec 99.5% regression

[linus:master] [mm/secretmem] 97d34aa65c: stress-ng.secretmem.ops_per_sec 99.5% regression

[kernel test robot]

Hello,

kernel test robot noticed a 99.5% regression of stress-ng.secretmem.ops_per_sec on:

commit: 97d34aa65c29cca85e3e9050f4c936389b38a054 ("mm/secretmem: properly account locked pages") https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master

testcase: stress-ng
config: x86_64-rhel-9.4
compiler: gcc-14
test machine: 224 threads 2 sockets Intel(R) Xeon(R) Platinum 8480CTDX (Sapphire Rapids) with 256G memory
parameters:

	nr_threads: 100%
	testtime: 60s
	test: secretmem
	cpufreq_governor: performance

If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <yi1.lai@intel.com>
| Closes: https://lore.kernel.org/oe-lkp/202609230824.7a9189aa-lkp@intel.com


Details are as below:
-------------------------------------------------------------------------------------------------->

The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20260923/202609230824.7a9189aa-lkp@intel.com

=========================================================================================
compiler/cpufreq_governor/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime:
  gcc-14/performance/x86_64-rhel-9.4/100%/debian-13-x86_64-20250902.cgz/lkp-spr-r02/secretmem/stress-ng/60s

commit: 
  35b0fb391b ("mm/mremap: reset unfaulted VMA page offset for MREMAP_DONTUNMAP")
  97d34aa65c ("mm/secretmem: properly account locked pages")

35b0fb391b0df573 97d34aa65c29cca85e3e9050f4c
---------------- --------------------------- 
         %stddev     %change         %stddev
             \          |                \  
   6080503           -99.6%      23092 ±  3%  stress-ng.secretmem.ops
     85584           -99.5%     385.28 ±  3%  stress-ng.secretmem.ops_per_sec
     71.48           -16.0%      60.07        stress-ng.time.elapsed_time
     71.48           -16.0%      60.07        stress-ng.time.elapsed_time.max
     97435 ±  2%   +1276.4%    1341103        stress-ng.time.involuntary_context_switches
    230308           -96.1%       9096        stress-ng.time.maximum_resident_set_size
  18401243         +3349.6%  6.348e+08        stress-ng.time.minor_page_faults
     20417            -8.7%      18639        stress-ng.time.percent_of_cpu_this_job_got
     14581           -25.0%      10937        stress-ng.time.system_time
     13.34         +1858.5%     261.33        stress-ng.time.user_time
      6898 ±  2%  +96037.7%    6631576 ±  3%  stress-ng.time.voluntary_context_switches

Disclaimer:
Results have been estimated based on internal Intel analysis and are provided
for informational purposes only. Any difference in system hardware or software
design or configuration may affect actual performance.


-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

Re: [linus:master] [mm/secretmem] 97d34aa65c: stress-ng.secretmem.ops_per_sec 99.5% regression

[Vlastimil Babka (SUSE)]

On 9/23/26 03:00, kernel test robot wrote:
> 
> 
> Hello,
> 
> kernel test robot noticed a 99.5% regression of stress-ng.secretmem.ops_per_sec on:
> 
> commit: 97d34aa65c29cca85e3e9050f4c936389b38a054 ("mm/secretmem: properly account locked pages") https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master

Well maybe there is some inefficiency that can be fixed (haven't checked)
but if it's the cost for proper accounting then so be it? I doubt creating
secretmem pages is something userspace would have to do intensively enough
to matter.

> testcase: stress-ng
> config: x86_64-rhel-9.4
> compiler: gcc-14
> test machine: 224 threads 2 sockets Intel(R) Xeon(R) Platinum 8480CTDX (Sapphire Rapids) with 256G memory
> parameters:
> 
> 	nr_threads: 100%
> 	testtime: 60s
> 	test: secretmem
> 	cpufreq_governor: performance
> 
> If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
> | Reported-by: kernel test robot <yi1.lai@intel.com>
> | Closes: https://lore.kernel.org/oe-lkp/202609230824.7a9189aa-lkp@intel.com
> 
> 
> Details are as below:
> -------------------------------------------------------------------------------------------------->
> 
> The kernel config and materials to reproduce are available at:
> https://download.01.org/0day-ci/archive/20260923/202609230824.7a9189aa-lkp@intel.com
> 
> =========================================================================================
> compiler/cpufreq_governor/kconfig/nr_threads/rootfs/tbox_group/test/testcase/testtime:
>   gcc-14/performance/x86_64-rhel-9.4/100%/debian-13-x86_64-20250902.cgz/lkp-spr-r02/secretmem/stress-ng/60s
> 
> commit: 
>   35b0fb391b ("mm/mremap: reset unfaulted VMA page offset for MREMAP_DONTUNMAP")
>   97d34aa65c ("mm/secretmem: properly account locked pages")
> 
> 35b0fb391b0df573 97d34aa65c29cca85e3e9050f4c
> ---------------- --------------------------- 
>          %stddev     %change         %stddev
>              \          |                \  
>    6080503           -99.6%      23092 ±  3%  stress-ng.secretmem.ops
>      85584           -99.5%     385.28 ±  3%  stress-ng.secretmem.ops_per_sec
>      71.48           -16.0%      60.07        stress-ng.time.elapsed_time
>      71.48           -16.0%      60.07        stress-ng.time.elapsed_time.max
>      97435 ±  2%   +1276.4%    1341103        stress-ng.time.involuntary_context_switches
>     230308           -96.1%       9096        stress-ng.time.maximum_resident_set_size
>   18401243         +3349.6%  6.348e+08        stress-ng.time.minor_page_faults
>      20417            -8.7%      18639        stress-ng.time.percent_of_cpu_this_job_got
>      14581           -25.0%      10937        stress-ng.time.system_time
>      13.34         +1858.5%     261.33        stress-ng.time.user_time
>       6898 ±  2%  +96037.7%    6631576 ±  3%  stress-ng.time.voluntary_context_switches
> 
> Disclaimer:
> Results have been estimated based on internal Intel analysis and are provided
> for informational purposes only. Any difference in system hardware or software
> design or configuration may affect actual performance.
> 
> 

Re: [linus:master] [mm/secretmem] 97d34aa65c: stress-ng.secretmem.ops_per_sec 99.5% regression

[Lorenzo Stoakes (ARM)]

On Wed, Sep 23, 2026 at 08:35:47AM +0200, Vlastimil Babka (SUSE) wrote:
> On 9/23/26 03:00, kernel test robot wrote:
> >
> >
> > Hello,
> >
> > kernel test robot noticed a 99.5% regression of stress-ng.secretmem.ops_per_sec on:
> >
> > commit: 97d34aa65c29cca85e3e9050f4c936389b38a054 ("mm/secretmem: properly account locked pages") https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
>
> Well maybe there is some inefficiency that can be fixed (haven't checked)
> but if it's the cost for proper accounting then so be it? I doubt creating
> secretmem pages is something userspace would have to do intensively enough
> to matter.

Yeah, exactly.

This isn't something that should be benchmarked, this is possibly the most
ludicrious benchmark I've ever seen :)

If establishing a secretmem region is on your hot path you have engineered your
software wrong.

This is effectively a security flaw so I'm afraid must remain.

--
Cheers, Lorenzo