2.5: system time goes up to 100%

Arador (diegocg@teleline.es)
Sat, 15 Feb 2003 22:22:29 +0100


Hi, i've the following case (more info provided below):

procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
0 1 44268 124776 10380 64736 5 8 93 37 600 281 11 2 85 3
1 0 44268 124160 10584 64736 0 0 196 196 1750 1518 1 41 49 9
1 0 44268 123544 10676 64864 0 0 220 0 1722 1424 2 49 47 2
2 0 44268 123088 10696 64992 0 0 148 0 1777 1681 2 56 41 0
2 0 44268 122304 10712 65136 144 0 304 0 1738 1554 3 51 44 3
1 0 44268 121616 10716 65264 0 0 132 0 1697 1385 2 50 47 0
2 0 44268 120904 10728 65392 0 0 132 376 1791 1425 3 56 42 0
2 0 44268 120408 10728 65520 0 0 128 0 1696 1385 2 50 48 0
2 0 44268 119848 10728 65648 0 0 128 0 1699 1393 2 50 48 0
2 0 44268 119232 10728 65776 0 0 128 0 1698 1375 2 50 47 0
2 0 44268 118728 10728 65904 0 0 128 0 1694 1392 2 50 49 0
2 0 44268 118240 10736 66032 0 0 128 96 1754 2130 6 62 31 0
2 0 44268 117792 10736 66032 0 0 128 0 1739 1664 2 56 43 0
2 0 44268 117232 10736 66160 16 0 16 0 1756 1761 4 60 36 0
3 0 44268 116736 10736 66288 0 0 128 0 1775 1754 2 58 40 0
1 0 44268 116248 10736 66416 0 0 128 0 1343 728 2 56 42 0
1 0 44268 115688 10744 66544 0 0 128 32 1021 106 0 51 48 0
1 0 44268 115072 10744 66672 0 0 128 0 1012 57 1 50 49 0
3 0 44268 114512 10744 66800 0 0 128 0 1036 177 1 51 47 0
7 0 44268 114080 10744 66928 0 0 128 0 1108 455 2 55 43 0
1 0 44268 113464 10752 67056 0 0 128 28 1042 518 4 57 39 0
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
2 0 44268 112904 10752 67184 0 0 128 0 1043 148 1 51 48 0
1 0 44268 112344 10752 67312 0 0 128 0 1012 140 1 52 47 0
1 0 44268 111864 10752 67440 0 0 128 0 1009 45 0 50 49 0
1 0 44268 111248 10752 67568 0 0 128 0 1045 264 2 53 44 0
1 0 44268 110688 10760 67696 0 0 128 20 1108 693 1 55 43 0
1 0 44268 110184 10760 67824 0 0 128 0 1012 370 2 50 48 0
1 0 44268 109584 10760 67952 0 0 128 0 1015 158 1 52 47 0
1 0 44268 109136 10760 67996 0 0 44 0 1023 221 1 53 46 0
1 0 44268 108688 10760 67996 0 0 0 0 1013 224 1 51 47 0
1 0 44268 108408 10772 68028 0 0 36 16 1014 122 4 47 49 1
1 0 44268 103256 10780 68440 4 0 424 0 1084 362 39 4 50 7
1 0 44268 101232 10788 61964 0 0 16 0 1006 522 51 3 46 0

(sorry if it's disordered, it's on a Eterm and i can't cut paste in a decent shape)

system is a p3 2x800, 256 mb ram; kernel is 2.5.61; but i've
experimented it more times in ther 2.5 releases

Disk is a UDMA100, fs is a only ext3 particion (no htree and/or orlov)

It just happens sometimes when i open the mail client sylpheed
in the lkml carpet; 137 MB, like 29000 messages
This mail client saves each message in a file
diego@estel:~/sylpheed-mail/inbox/linux-kernel$ ls -l | wc -l
29096

It takes a lot of time, usually it's opened very fast (as expected
in this great mail client ;)

i took readprofile just after doing this, i hope it helps (BTW shouldn't "total" show more time?):
8209527 total 4,5041
8035501 default_idle 100443,7625
39060 ext3_find_entry 35,9007
19200 serial_in 171,4286
18957 check_poison_obj 131,6458
10930 __constant_c_and_count_memset 68,3125
4057 delay_tsc 126,7812
3843 __preempt_spin_lock 24,0188
2944 __copy_to_user_ll 23,0000
2911 kmem_cache_free 20,2153
2784 do_page_fault 2,3533
2251 page_remove_rmap 4,6896
1955 schedule 1,9092
1940 __constant_memcpy 7,1324
1907 page_add_rmap 5,6756
1469 buffered_rmqueue 3,1659
1410 do_anonymous_page 1,4936
1346 fget 10,5156
1247 zap_pte_range 2,2268
1235 kfree 5,1458
1050 link_path_walk 0,5048
961 ide_outb 60,0625
934 find_get_page 7,2969
898 __d_lookup 4,0089
867 ide_end_request 2,0069
858 cpu_idle 10,7250
836 __wake_up 4,3542
825 release_pages 1,3936
810 free_hot_cold_page 2,6645
782 system_call 17,7727
767 kmem_cache_alloc 4,3580
711 handle_mm_fault 2,9625
708 find_vma 7,3750
701 radix_tree_lookup 6,2589
681 __find_get_block_slow 2,1281
654 do_no_page 0,6928
628 cache_alloc_debugcheck_after 3,0192
617 lookup_bh_lru 4,2847
608 __might_sleep 4,7500
607 __copy_from_user_ll 4,8952
561 do_select 0,8766
519 pte_alloc_map 1,4103
506 __pagevec_lru_add_active 1,1295
470 kmalloc 1,9583
468 __make_request 0,3250
456 unix_poll 2,8500
456 bh_lru_install 1,9000
439 copy_page_range 0,3049
428 do_get_write_access 0,2106
399 poison_obj 6,2344
398 __alloc_pages 0,5293
391 ext3_do_update_inode 0,3703
389 do_mmap_pgoff 0,2152
384 __copy_user_intel 2,1818
364 filemap_nopage 0,4946
360 ide_inb 22,5000
359 ext3_get_block_handle 0,3740
349 fput 10,9062
335 remove_wait_queue 1,9034
328 dnotify_parent 1,4643
326 .text.lock.sched 1,1088
325 serial_out 2,9018
325 add_wait_queue 2,0312
324 handle_IRQ_event 2,8929
321 __pte_chain_free 2,2292
321 journal_dirty_metadata 0,4666
320 pte_chain_alloc 2,0000
316 sys_brk 1,0395
311 do_generic_mapping_read 0,2817
310 __get_page_state 1,7614
300 sock_poll 4,6875
293 ext3_get_inode_loc 0,7043
289 vma_merge 0,3284
289 mark_page_accessed 4,5156
284 prep_new_page 3,5500
258 strnlen_user 3,2250
253 bad_range 2,2589
251 unix_stream_recvmsg 0,2752
245 ide_do_request 0,3828
240 probe_irq_on 0,3000
238 receive_chars 0,3628
237 journal_add_journal_head 0,4003
234 lru_cache_add_active 2,4375
233 start_this_handle 0,2468
230 atomic_dec_and_lock 1,5972
227 ll_rw_block 1,5764
225 __brelse 3,5156
216 ide_intr 0,2935
214 ext3_dirty_inode 0,6080
213 handle_pte_fault 0,7422
212 find_vma_prepare 1,8929
206 sys_select 0,1497
204 free_one_pmd 0,7500
200 journal_stop 0,2778
199 vm_enough_memory 1,0365
198 unlock_buffer 2,0625
196 nr_free_pages 2,4500
196 __constant_c_and_count_memset 1,2250
195 dput 0,6771
191 __pollwait 0,9183
191 path_lookup 0,4263
191 journal_cancel_revoke 0,4263
187 arch_get_unmapped_area 0,9947
184 serial8250_interrupt 0,4259
179 kmem_flagcheck 2,7969
178 journal_get_write_access 1,1125
175 ext3_read_inode 0,1854
173 do_brk 0,3180
172 unix_stream_sendmsg 0,2108
172 __constant_c_and_count_memset 1,0750
171 ext3_check_dir_entry 0,7125
167 find_revoke_record 1,4911
167 __find_get_block 1,3047
164 unlock_nd 0,9318
162 __mark_inode_dirty 0,4821
162 get_unused_fd 0,3375
156 vfs_read 0,5132
155 journal_unlock_journal_head 0,6458
153 may_open 0,3187
153 __do_page_cache_readahead 0,3741
152 __set_page_dirty_buffers 0,2714
147 snd_pcm_lib_write1 0,0875
147 skb_dequeue 0,7656
145 find_next_usable_block 0,1849
145 dentry_open 0,3125
144 do_lookup 0,3750
142 old_mmap 0,3859
139 generic_fillattr 1,0859
137 strncpy_from_user 1,2232
137 __journal_file_buffer 0,2141
136 current_kernel_time 1,7000
135 wake_up_buffer 2,1094
131 getname 0,6298
130 __fput 0,4062
129 ext3_get_branch 0,5375
129 dget_locked 2,6875
125 __constant_c_and_count_memset 0,7812
124 add_timer 0,4844
123 remove_shared_vm_struct 0,7688
122 get_empty_filp 0,2007
120 do_wp_page 0,1136
117 __generic_file_aio_read 0,2216
[...]
67 syscall_exit 6,0909
(i left this, just because it shows too many time compared to others in its zone)

Diego Calleja
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/