| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Fri, 21 Nov 2008 13:51:39 +0100
From: Eric Dumazet <dada1@...mosbay.com>
To: David Miller <davem@...emloft.net>
CC: mingo@...e.hu, cl@...ux-foundation.org, rjw@...k.pl,
linux-kernel@...r.kernel.org, kernel-testers@...r.kernel.org,
efault@....de, a.p.zijlstra@...llo.nl
Subject: Re: [Bug #11308] tbench regression on each kernel release from 2.6.22
-> 2.6.28
David Miller a écrit :
> From: Eric Dumazet <dada1@...mosbay.com>
> Date: Fri, 21 Nov 2008 09:51:32 +0100
>
>> Now, I wish sockets and pipes not going through dcache, not tbench affair
>> of course but real workloads...
>>
>> running 8 processes on a 8 way machine doing a
>>
>> for (;;)
>> close(socket(AF_INET, SOCK_STREAM, 0));
>>
>> is slow as hell, we hit so many contended cache lines ...
>>
>> ticket spin locks are slower in this case (dcache_lock for example
>> is taken twice when we allocate a socket(), once in d_alloc(), another one
>> in d_instantiate())
>
> As you of course know, this used to be a ton worse. At least now
> these things are unhashed. :)
Well, this is dust compared to what we currently have.
To allocate a socket we :
0) Do the usual file manipulation (pretty scalable these days)
(but recent drop_file_write_access() and co slow down a bit)
1) allocate an inode with new_inode()
This function :
- locks inode_lock,
- dirties nr_inodes counter
- dirties inode_in_use list (for sockets, I doubt it is usefull)
- dirties superblock s_inodes.
- dirties last_ino counter
All these are in different cache lines of course.
2) allocate a dentry
d_alloc() takes dcache_lock,
insert dentry on its parent list (dirtying sock_mnt->mnt_sb->s_root)
dirties nr_dentry
3) d_instantiate() dentry (dcache_lock taken again)
4) init_file() -> atomic_inc on sock_mnt->refcount (in case we want to umount this vfs ...)
At close() time, we must undo the things. Its even more expensive because
of the _atomic_dec_and_lock() that stress a lot, and because of two cache
lines that are touched when an element is deleted from a list.
for (i = 0; i < 1000*1000; i++)
close(socket(socket(AF_INET, SOCK_STREAM, 0));
Cost if run one one cpu :
real 0m1.561s
user 0m0.092s
sys 0m1.469s
If run on 8 CPUS :
real 0m27.496s
user 0m0.657s
sys 3m39.092s
CPU: Core 2, speed 3000.11 MHz (estimated)
Counted CPU_CLK_UNHALTED events (Clock cycles when not halted) with a unit mask of 0x00 (Unhalted core cycles) count 100
000
samples cum. samples % cum. % symbol name
164211 164211 10.9678 10.9678 init_file
155663 319874 10.3969 21.3647 d_alloc
147596 467470 9.8581 31.2228 _atomic_dec_and_lock
92993 560463 6.2111 37.4339 inet_create
73495 633958 4.9088 42.3427 kmem_cache_alloc
46353 680311 3.0960 45.4387 dentry_iput
46042 726353 3.0752 48.5139 tcp_close
42784 769137 2.8576 51.3715 kmem_cache_free
37074 806211 2.4762 53.8477 wake_up_inode
36375 842586 2.4295 56.2772 tcp_v4_init_sock
35212 877798 2.3518 58.6291 inotify_d_instantiate
33199 910997 2.2174 60.8465 sysenter_past_esp
31161 942158 2.0813 62.9277 d_instantiate
31000 973158 2.0705 64.9983 generic_forget_inode
28020 1001178 1.8715 66.8698 vfs_dq_drop
19007 1020185 1.2695 68.1393 __copy_from_user_ll
17513 1037698 1.1697 69.3090 new_inode
16957 1054655 1.1326 70.4415 __init_timer
16897 1071552 1.1286 71.5701 discard_slab
16115 1087667 1.0763 72.6464 d_kill
15542 1103209 1.0381 73.6845 __percpu_counter_add
13562 1116771 0.9058 74.5903 __slab_free
13276 1130047 0.8867 75.4771 __fput
12423 1142470 0.8297 76.3068 new_slab
11976 1154446 0.7999 77.1067 tcp_v4_destroy_sock
10889 1165335 0.7273 77.8340 inet_csk_destroy_sock
10516 1175851 0.7024 78.5364 alloc_inode
9979 1185830 0.6665 79.2029 sock_attach_fd
7980 1193810 0.5330 79.7359 drop_file_write_access
7609 1201419 0.5082 80.2441 alloc_fd
7584 1209003 0.5065 80.7506 sock_init_data
7164 1216167 0.4785 81.2291 add_partial
7107 1223274 0.4747 81.7038 sys_close
6997 1230271 0.4673 82.1711 mwait_idle
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists