config: deepin_loongarch_desktop_defconfig: disbale ZERO_CALL_USED_REG

[opsiff]

It show -10% performance overhead in loongarch,
as other arch configs do, disable it in loongarch now. If we sure need security enhance, we can bring it back.

Unixbench micro test:
1 parallel: 1167.0 1288.6 10.4%
8 parallel: 4301.6 4791.3 11.3%

Log:
1 parallel:
before:
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 40211977.7 3445.8
Double-Precision Whetstone 55.0 5025.3 913.7
Execl Throughput 43.0 3602.6 837.8
File Copy 1024 bufsize 2000 maxblocks 3960.0 541638.0 1367.8
File Copy 256 bufsize 500 maxblocks 1655.0 141748.0 856.5
File Copy 4096 bufsize 8000 maxblocks 5800.0 1656091.0 2855.3
Pipe Throughput 12440.0 787709.6 633.2
Pipe-based Context Switching 4000.0 82120.6 205.3
Process Creation 126.0 7469.1 592.8
Shell Scripts (1 concurrent) 42.4 11985.3 2826.7
Shell Scripts (8 concurrent) 6.0 4240.7 7067.9
System Call Overhead 15000.0 704474.0 469.6
========
System Benchmarks Index Score 1167.0

after disable config:

Dhrystone 2 using register variables 116700.0 41690728.5 3572.5
Double-Precision Whetstone 55.0 5026.8 914.0
Execl Throughput 43.0 4006.3 931.7
File Copy 1024 bufsize 2000 maxblocks 3960.0 637314.0 1609.4
File Copy 256 bufsize 500 maxblocks 1655.0 170265.0 1028.8
File Copy 4096 bufsize 8000 maxblocks 5800.0 1896428.0 3269.7
Pipe Throughput 12440.0 867146.6 697.1
Pipe-based Context Switching 4000.0 94776.9 236.9
Process Creation 126.0 7901.1 627.1
Shell Scripts (1 concurrent) 42.4 12825.9 3025.0
Shell Scripts (8 concurrent) 6.0 4622.5 7704.2
System Call Overhead 15000.0 790876.8 527.3
========
System Benchmarks Index Score 1288.6

8 parallel:
before:
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 221465392.9 18977.3
Double-Precision Whetstone 55.0 36704.9 6673.6
Execl Throughput 43.0 19620.7 4563.0
File Copy 1024 bufsize 2000 maxblocks 3960.0 1007204.0 2543.4
File Copy 256 bufsize 500 maxblocks 1655.0 259253.0 1566.5
File Copy 4096 bufsize 8000 maxblocks 5800.0 3193191.0 5505.5
Pipe Throughput 12440.0 4382384.7 3522.8
Pipe-based Context Switching 4000.0 645803.9 1614.5
Process Creation 126.0 36849.0 2924.5
Shell Scripts (1 concurrent) 42.4 40227.1 9487.5
Shell Scripts (8 concurrent) 6.0 5220.6 8701.0
System Call Overhead 15000.0 3459103.6 2306.1
========
System Benchmarks Index Score 4301.6

after disable config:

Dhrystone 2 using register variables 116700.0 221216371.2 18956.0
Double-Precision Whetstone 55.0 36708.4 6674.3
Execl Throughput 43.0 21867.6 5085.5
File Copy 1024 bufsize 2000 maxblocks 3960.0 1188570.0 3001.4
File Copy 256 bufsize 500 maxblocks 1655.0 327965.0 1981.7
File Copy 4096 bufsize 8000 maxblocks 5800.0 3632452.0 6262.8
Pipe Throughput 12440.0 4855414.6 3903.1
Pipe-based Context Switching 4000.0 759442.0 1898.6
Process Creation 126.0 40174.7 3188.5
Shell Scripts (1 concurrent) 42.4 43971.7 10370.7
Shell Scripts (8 concurrent) 6.0 5664.2 9440.3
System Call Overhead 15000.0 3959683.3 2639.8
========
System Benchmarks Index Score 4791.3

[deepin-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please ask for approval from opsiff. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• deepin/OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment