Skip to contentSkip to conversation

2.4.9-ac15 sluggish

DISCUSSION Xavier Bestel 4 participants 5 messages Sep 27, 2001

lore
Share & tools
lore
#0 lore raw

I gave 2.4.9-ac15 a try on my dual-pIII, 700MB

I tried to run /usb/bin/automake on the gstreamer project (current automake has a bug which sucks all ram, gstreamer provides its own)

with -ac10 no real bad behavior, just automake is working like crazy.

with -ac15 the system starts disk-trashing immediately, xterms, ssh or telnet sessions are unresponsive for 20mn (after that I gave up and rebooted)

#1 lore raw
Quoted 10 lines · Xavier Bestel

I gave 2.4.9-ac15 a try on my dual-pIII, 700MB

I tried to run /usb/bin/automake on the gstreamer project (current automake has a bug which sucks all ram, gstreamer provides its own)

with -ac10 no real bad behavior, just automake is working like crazy.

with -ac15 the system starts disk-trashing immediately, xterms, ssh or telnet sessions are unresponsive for 20mn (after that I gave up and rebooted)

Did you try Rik's patch on http://www.surriel.com/patches?

I had the same problem and it is fixed with the patch. patch name is 2.4.9-ac15-age+launder

Roberto

#2 lore raw
In reply to Xavier Bestel
Quoted 5 lines · Xavier Bestel

with -ac10 no real bad behavior, just automake is working like crazy.

with -ac15 the system starts disk-trashing immediately, xterms, ssh or telnet sessions are unresponsive for 20mn (after that I gave up and rebooted)

We discovered a merge bug, -ac15 has a few lines in try_to_swap_out() 10 lines higher than they were in the patch I sent to Alan ;)

This is fixed in the age+launder patch from my home page ard in the vmscan.c I sent to Alan for -ac16.

If you want the patch:
http://www.surriel.com/patches/

Rik

Signature · 5 lines
IA64: a worthy successor to i860.

http://www.surriel.com/		http://distro.conectiva.com/

Send all your spam to [email protected] (spam digging piggy)
#3 lore raw
Quoted 6 lines · le jeu 27-09-2001 at 17:13 Rik van Riel a écrit :

We discovered a merge bug, -ac15 has a few lines in try_to_swap_out() 10 lines higher than they were in the patch I sent to Alan ;)

This is fixed in the age+launder patch from my home page ard in the vmscan.c I sent to Alan for -ac16.

Mmmh ... i'm now compiling (make -j5) -ac16 under -ac15-age+launder and it's sluggish: xmms drops audio sometimes, my mouse pointer is jerky.

Xav

#4 lore raw
In reply to Xavier Bestel
Quoted 4 lines

I gave 2.4.9-ac15 a try on my dual-pIII, 700MB

I tried to run /usb/bin/automake on the gstreamer project (current automake has a bug which sucks all ram, gstreamer provides its own)

ac15 has a merge error that causes bad VM behaviour. Ac16 is brewing