Op Sat Dec 9 2006, schreef Con Kolivas: (cut) > Changes (first significant changes since : > Added: > +sched-fix_iso_starvation.patch > A bug first introduced into 2.6.18-ck1/cks1 meant that SCHED_ISO tasks were > not being throttled when above their cpu limit. This presents a security > risk to any machine with user logins and upgrading for this issue should be > considered a high priority. > > +sched-make_softirqd_batch.patch > Instead of 'nice'ing ksoftirqd we can use the policy hint of SCHED_BATCH > which specifies it as not requiring low latency. This increases the cpu use > possible under very heavy softirq traffic (such as network loads) and > decreases the latency that might otherwise be seen (such as keyboard input > under heavy cpu load on slow machines). nice, con, nice! good to see you're back! how are things progressing in the japanese area? grtz Jos -- Disclaimer: Alles wat ik doe denk en zeg is gebaseerd op het wereldbeeld wat ik nu heb. Ik ben niet verantwoordelijk voor wijzigingen van de wereld, of het beeld wat ik daarvan heb, noch voor de daaruit voortvloeiende gedragingen van mezelf. Alles wat ik zeg is aardig bedoeld, tenzij expliciet vermeld.
Attachment:
pgpWJpmGV4B0k.pgp
Description: PGP signature
- References:
- 2.6.19-ck2
- From: Con Kolivas <[email protected]>
- 2.6.19-ck2
- Prev by Date: Re: What was in the x86 merge for .20
- Next by Date: Re: VCD not readable under 2.6.18
- Previous by thread: 2.6.19-ck2
- Next by thread: Recursive spinlocks for Network Recursion Bugs in 2.6.18
- Index(es):
 |