lkml.org 
[lkml]   [2003]   [Aug]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectStrange (scheduler/vm?) starvations in latest 2.6 tree
From
Date
Hi!

Some days ago after updating the kernel source from bkcvs my system
started behaving very strange. When compiling something (without make
-j) the mouse cursor sometimes got *very* jumpy and even hangs for some
time, also xmms likes to skip. It seems like the whole system freezes
for some time.

I've got Ingo's and Con's scheduler tweaks applied and they worked great
so far so I don't think they are the problem.

I've found that pdflush somehow seems to be involved. When these hangs
appear, top shows some processes in uninterruptible sleep (including
pdflush), and pdflush takes a lot of cpu time, like 15%.

The system doesn't swap, it has a lot of memory and is also quite fast.

vmstat -n 1 output is added below (the system didn't swap in or out a
single page so I've removed the columns that were entirey zero).

The hangs always show up when the bo number gets big. The only processes
running are X, xmms and the compiler, the problems show up without doing
anything special.

The kernel is compiled with x86/athlon support, preemptible kernel,
local io-apic and acpi and runs on a via kt400 chipset.

procs-----memory---------- -----io---- --system-- ----cpu----
r b free buff cache bi bo in cs us sy id wa
3 1 48072 75336 228300 84 205 1276 1508 26 4 64 7
2 0 48136 75372 228392 52 52 1150 1647 90 10 0 0
6 2 59784 75376 228384 0 4476 1580 1103 60 40 0 0
0 2 59784 75376 228384 0 1040 1368 726 2 1 0 97
3 2 43464 75376 228404 16 652 1409 1624 43 4 0 53
1 2 42248 75380 228548 68 668 1341 1477 87 8 0 5
0 2 27224 75380 228552 4 328 1255 723 77 5 0 18
0 3 59480 75380 228592 0 2452 1325 723 11 6 0 83
0 3 59480 75380 228592 0 472 1267 1669 7 2 0 91
4 1 49808 75388 228732 128 1608 1301 1430 43 12 0 45
3 2 52624 75392 229220 444 0 1409 2726 89 11 0 0
2 1 48400 75396 229556 276 0 1421 2619 85 15 0 0
2 1 44936 75412 230344 468 0 1291 3516 76 17 0 6
4 1 41072 75424 230900 512 0 1426 3178 90 10 0 0
6 3 47216 75424 230944 40 3708 1569 1434 80 20 0 0
1 4 34480 75424 230976 0 808 1309 1222 92 8 0 0
0 5 33312 75424 231428 0 728 1619 1660 61 4 0 35
3 0 37472 75428 231912 44 244 1251 2515 87 12 0 1
3 0 40136 75432 232140 124 0 1431 2431 86 13 0 1
2 1 33352 75440 232392 128 0 1632 3621 84 16 0 0
4 0 38272 75440 232472 144 0 1369 2812 87 12 0 1
4 1 38976 75444 233128 340 0 1724 2990 80 16 0 4
6 2 30016 75448 233208 48 2888 1506 1607 91 9 0 0
2 2 27712 75448 233212 8 992 1624 2378 95 5 0 0
4 0 38336 75448 233264 8 172 1255 1778 93 7 0 0
2 0 30016 75452 233428 136 0 1276 2325 91 9 0 0
1 0 37216 75456 233472 28 0 1129 1926 88 12 0 0
1 0 34016 75456 233636 16 0 1396 1774 92 8 0 0
1 0 33000 75460 233832 100 0 1129 1694 86 11 0 3
4 1 36200 75464 233912 64 2532 1392 1356 90 10 0 0
1 1 32664 75464 233912 0 736 1545 1856 23 4 0 72
2 0 21208 75464 232352 0 252 1324 1989 93 7 0 0

procs-----memory---------- -----io---- --system-- ----cpu----
r b free buff cache bi bo in cs us sy id wa
2 0 7024 76040 271172 85 240 1289 1533 27 4 61 7
0 0 6960 76040 271300 0 0 1235 1557 0 0 100 0
0 0 6960 76040 271300 0 0 1119 1551 1 0 99 0
0 0 6992 76040 271300 0 0 1115 1539 0 0 100 0
0 0 6992 76040 271300 0 0 1416 3066 9 2 89 0
0 0 6992 76040 271300 0 108 1523 3338 12 4 84 0
0 0 6992 76040 271300 0 0 1541 2773 7 2 91 0
0 0 6864 76040 271428 0 0 1282 1639 1 0 99 0
0 0 6864 76040 271428 0 0 1124 1581 0 0 100 0
0 0 6864 76040 271428 0 0 1123 1633 1 0 99 0
0 0 6864 76040 271428 0 44 1140 1843 0 1 98 1
0 0 6856 76040 271428 0 0 1133 1696 1 1 98 0
0 0 6728 76040 271556 0 0 1224 1676 1 0 99 0
0 0 8456 76040 269868 8 0 1121 1633 1 1 98 0
0 0 8456 76040 269868 0 0 1130 1725 1 0 99 0
0 0 8464 76040 269868 0 100 1142 1736 1 1 98 0
5 0 9160 76048 269096 4 3951 1676 1475 3 2 91 3
0 0 9160 76048 269096 0 72 1209 1283 1 0 97 1
0 0 9032 76048 269224 0 0 1253 1771 2 1 97 0
0 0 9032 76048 269224 0 0 1136 1669 1 0 99 0
0 0 9032 76048 269224 0 60 1132 1603 0 1 99 0
2 0 6808 76052 268776 40 0 1125 1764 41 11 47 0
1 0 6744 76056 268820 0 0 1115 1576 86 14 0 0
2 0 13400 74948 264776 8 0 1216 1699 83 17 0 0
1 0 13272 74948 264832 0 0 1115 1649 85 15 0 0
0 2 13080 74952 264900 0 3976 1378 658 34 20 0 46
0 2 12952 74952 264932 0 916 1316 1654 3 2 0 95
1 1 11928 74960 265028 0 896 1322 1800 79 18 0 3
2 0 12696 74960 265108 0 0 1245 1727 81 19 0 0
1 0 12240 74964 260876 0 0 1211 1756 81 19 0 0
1 0 6096 74964 264596 0 0 1115 1961 71 29 0 0
1 0 11408 74964 264552 4 0 1265 2390 80 20 0 0
1 0 13392 74976 264596 0 0 1128 2456 91 9 0 0
0 3 12752 74976 265268 0 2680 1319 1098 76 6 0 18
0 2 12880 74976 265328 0 420 1382 1586 5 1 0 94
2 1 9744 75028 268256 8 2813 1649 1617 20 35 0 45
0 1 9296 75548 268324 2 833 1498 2829 2 19 0 79
1 0 8720 76064 268388 3 818 1499 2807 1 17 0 82
1 2 8720 76268 267936 1 6467 1503 1578 8 18 0 74
5 3 8608 76268 267936 0 8180 2206 991 15 9 0 77
8 2 8352 76268 268064 0 6504 3534 677 34 16 0 50
5 4 8432 76272 268064 0 5783 1555 1025 9 4 0 87
4 3 8424 76276 268064 0 2305 1497 372 37 16 0 47
1 0 12264 76284 264624 0 9 1233 1597 49 12 0 39
0 0 9832 76312 268192 0 0 1216 3298 43 50 7 0
0 0 9832 76312 268192 0 0 1123 1581 1 0 99 0

--
Christophe Saout <christophe@saout.de>
Please avoid sending me Word or PowerPoint attachments.
See http://www.fsf.org/philosophy/no-word-attachments.html

-
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/

\
 
 \ /
  Last update: 2005-03-22 13:47    [W:0.046 / U:0.648 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site