|
Bugzilla – Full Text Bug Listing |
|
Description
andreas bittner
2009-01-25 23:30:41 UTC
so i have tested a little bit more, and i have to say, opensuse 11.0 x86 and opensuse 11.1 x86 are very sluggish on this board.
graphics doesnt get detected properly at all, graphics only seem to work in safe mode (boot selection from grub)
so i have now tested knoppix 5.3 which may be a similar system to 11.0 (kernel wise).
knoppix 5.3 (dvd edition) boots up fine, mouse (usb) and keyboard (usb) are doing fine. no sluggishness. also there is a standard kde3.5.9 starting up, although at a little bit of different (or default?) resolution of 1024x768 (unconfigure).
some outputs from knoppix:
> root@Knoppix:~# uname -a
> Linux Knoppix 2.6.24.4 #2 SMP PREEMPT Tue Mar 25 21:35:28 CET 2008 i686 GNU/Linux
knoppix 5.3 also has hpet activated in kernel, like opensuse 11.1 and 11.0
will attach some more knoppix 5.3 logs and then will report back with logs from 11.1 and 11.0
and i will also try to find out which even older opensuse release works better without sluggishness at usb devices.
Created attachment 268384 [details]
logfiles archived from knoppix 5.3 dvd boot
logfiles archived from knoppix 5.3 dvd boot
when using knoppix 5.3 dvd boot, this system also has minimal load, or next to nothing.
top and w shows 0.00 in longterm
contrary to the opensuse 11.0 and 11.1 systems which showed load 1.xx+ but top not displaying any processes at all that caused the load. weird....
additional info:
11.0:
the system comes up "normaly" only when booting failsafemode on grub and with additional kernel parameter hpet=off
usb devices (mouse/keyboard) work only with hpet=off in failsafe mode. normal failsafe mode (apparently doesnt turn off hpe) isnt sufficient.
also normal mode is completely insane. keyboard hits are completely randomly swallowed/forgotton or duplicated n-times, making it almost impossible to enter passwords at logon prompt where you dont see the characters (at least number of characters, stars, x-ses, dots....) if they arrived at the system during key-press or not...
also mouse is almost unusable.
what i have found out additionally: when there is some kind of load on the cpu/system the usb-problem seems to be a little bit less severe.
i just created a little crap script which loops "ls -lartR /" and is running the whole time on some "konsole" kde-window. this little load seems to be enough to let me type this bugreport in here right now actually from a normal boot with only one special parameter, hpet turned off.
> Kernel command line: root=/dev/disk/by-id/scsi-SATA_SAMSUNG_HD753LJS13UJ1NQB00974-part2 resume=/dev/sda1 splash=silent vga=0x317 hpet=off
from time to time it still swallows certain keystrokes and mouse events.
the most normal system behaviour is during failsafe with additional hpet=off parameter.
there is definitely something wronn either in the kernel, or one of the modules/drivers or something the like.
i also turned off rcpowersaved and some various standard services, but havent found out if anything during runtime lessens this problem.....
--------------
lsmod output:
Module Size Used by
ip6t_LOG 23428 7
xt_tcpudp 19584 2
xt_pkttype 18304 3
ipt_LOG 22788 8
xt_limit 19076 15
af_packet 38656 0
snd_pcm_oss 64256 0
snd_mixer_oss 33408 1 snd_pcm_oss
snd_seq 73664 0
snd_seq_device 25100 1 snd_seq
ip6t_REJECT 21376 3
nf_conntrack_ipv6 33528 4
ipt_REJECT 20352 3
xt_state 18944 8
iptable_mangle 19712 0
iptable_nat 23688 0
nf_nat 35736 1 iptable_nat
iptable_filter 19840 1
ip6table_mangle 19584 0
nf_conntrack_netbios_ns 19200 0
nf_conntrack_ipv4 27652 7 iptable_nat,nf_nat
nf_conntrack 79188 6 nf_conntrack_ipv6,xt_state,iptable_nat,nf_nat,nf_conntrack_netbios_ns,nf_conntrack_ipv4
ip_tables 30224 3 iptable_mangle,iptable_nat,iptable_filter
cpufreq_conservative 24456 0
cpufreq_userspace 22660 0
cpufreq_powersave 18176 0
powernow_k8 31748 0
ip6table_filter 19712 1
ip6_tables 31376 3 ip6t_LOG,ip6table_mangle,ip6table_filter
x_tables 33668 11 ip6t_LOG,xt_tcpudp,xt_pkttype,ipt_LOG,xt_limit,ip6t_REJECT,ipt_REJECT,xt_state,iptable_nat,ip_tables,ip6_tables
Module Size Used by
ip6t_LOG 23428 7
xt_tcpudp 19584 2
xt_pkttype 18304 3
ipt_LOG 22788 8
xt_limit 19076 15
af_packet 38656 0
snd_pcm_oss 64256 0
snd_mixer_oss 33408 1 snd_pcm_oss
snd_seq 73664 0
snd_seq_device 25100 1 snd_seq
ip6t_REJECT 21376 3
nf_conntrack_ipv6 33528 4
ipt_REJECT 20352 3
xt_state 18944 8
iptable_mangle 19712 0
iptable_nat 23688 0
nf_nat 35736 1 iptable_nat
iptable_filter 19840 1
ip6table_mangle 19584 0
nf_conntrack_netbios_ns 19200 0
nf_conntrack_ipv4 27652 7 iptable_nat,nf_nat
nf_conntrack 79188 6 nf_conntrack_ipv6,xt_state,iptable_nat,nf_nat,nf_conntrack_netbios_ns,nf_conntrack_ipv4
ip_tables 30224 3 iptable_mangle,iptable_nat,iptable_filter
cpufreq_conservative 24456 0
cpufreq_userspace 22660 0
cpufreq_powersave 18176 0
powernow_k8 31748 0
ip6table_filter 19712 1
ip6_tables 31376 3 ip6t_LOG,ip6table_mangle,ip6table_filter
x_tables 33668 11 ip6t_LOG,xt_tcpudp,xt_pkttype,ipt_LOG,xt_limit,ip6t_REJECT,ipt_REJECT,xt_state,iptable_nat,ip_tables,ip6_tables
ipv6 281064 13 ip6t_REJECT,nf_conntrack_ipv6,ip6table_mangle
binfmt_misc 28040 1
fuse 66332 1
loop 35332 0
dm_mod 78676 0
rtc_cmos 27168 0
rtc_core 37148 1 rtc_cmos
rtc_lib 19328 1 rtc_core
serio_raw 23428 0
snd_hda_intel 386716 4
ohci1394 48432 0
i2c_piix4 25484 0
ieee1394 107016 1 ohci1394
snd_pcm 100100 2 snd_pcm_oss,snd_hda_intel
i2c_core 41108 1 i2c_piix4
snd_timer 40712 2 snd_seq,snd_pcm
snd_page_alloc 27400 2 snd_hda_intel,snd_pcm
snd_hwdep 26372 1 snd_hda_intel
snd 79544 16 snd_pcm_oss,snd_mixer_oss,snd_seq,snd_seq_device,snd_hda_intel,snd_pcm,snd_timer,snd_hwdep
soundcore 24264 1 snd
shpchp 50196 0
sr_mod 33448 0
pci_hotplug 46116 1 shpchp
joydev 28224 0
cdrom 50588 1 sr_mod
wmi 24488 0
button 25360 0
sky2 63236 0
sg 52020 0
usbhid 60260 0
hid 53708 1 usbhid
ff_memless 21896 1 usbhid
ohci_hcd 39940 0
ehci_hcd 53004 0
usbcore 164812 4 usbhid,ohci_hcd,ehci_hcd
sd_mod 45336 7
edd 26440 0
ext3 156040 5
mbcache 25348 1 ext3
jbd 73504 1 ext3
fan 22660 0
ahci 46344 6
pata_atiixp 24320 0
libata 176604 2 ahci,pata_atiixp
scsi_mod 168436 4 sr_mod,sg,sd_mod,libata
dock 27536 1 libata
thermal 40348 0
processor 67756 2 powernow_k8,thermal
----------------
Created attachment 268395 [details]
opensuse 11.0 normal boot - sluggish usb everything - logfiles
logfiles from
opensuse 11.0 normal boot - sluggish usb everything
additional test: 11.0: i am running a no-output commandline process which only works on the cpu/cores and doesnt create any output in konsole. so cpu-load only, and almost no activity inside x/kde. the sluggishness is also almost gone during cpu-load-only. so the problem might not be at a graphic driver level or graphic/screen activity but rather at cpu/sheduling? level..... Created attachment 268398 [details]
11.0 logs from failsafe boot with additional parameter hpet=off
11.0 logs from failsafe boot with additional parameter hpet=off
mouse / keyboard (both usb) seem to work normally.
additional bughunting to narrow down the source of the sluggishness. started/rebooted a variations of the failsafe parameters and without the parameters. i have figured out, that as soon as i have acpi=off on the kernel commandline parameters, everything works "fine" as far as i can tell. with acpi=off, usb mouse/keyboard sluggishness are gone. i dont even need to turn off hpet. acpi=off is enough, the system works normally with usb mouse/keyboard. during the whole time an additional ps/2 keyboard has been connected, and even during usb sluggishness, the ps/2 keyboard was working normally in all the apps and during login, x/kde, console, wherever. Created attachment 268405 [details]
11.0 standard boot with acpi=off as addtl. parameter
11.0 standard boot with acpi=off as addtl. parameter
usb keyboard/mouse problems are gone when using acpi=off as a single additional essential parameter during boot.
whenever acpi=off isnt supplied (acpi=on?) the whole system is almost unusable, not only the usb-devices, but the boot takes much much longer to bring up the boot messages, the harddisk/partition checking, the printout of all the services and processes, and the time til the desktop appears is much much longer (minutes) than with apci=off.
with acpi=off as a parameter, the system boots up very quickly (less than a minute), and shows kde logon screen very rapidly.
did some tests with clean 11.1 x86. during install (first booting up of the 11.1 setup from dvd media) the 11.1 setup-system is also very sluggish once it finished loading the yast2 and setup parts. its the same symptom as with 11.0 in kde. when the system has a little load on the cpu (created a little load again (while true; do ls -lartR /; done;)) the whole system is more responsive and i can use the usb keyboard and usb mouse to go through the steps of 11.1 clean install, partinioning and package installing. after another one or two reboots of the 11.1 system the installed 11.1 came up (X is still not working on this board/gfx, it did work with 11.0, but i will file another bugreport for that problem). it was once again sluggish even in text/console/framebuffer mode on the commandline. a ps/2 keyboard worked fine. so then i booted 11.1 once again with additional parameter acpi=off, and all of a sudden the whole system booted up like madman (quickly) and usb keyboard/mouse was reacting normally again. so basically, 11.1 behaves the same way like 11.0. there is a huge problem with acpi on this board. and once again as a reminder: its not just that usb keyboard/mice dont work properly with acpi turned on, but its also the whole system is taking ages compared to acpi=off to boot, to start, to start various services, to start x/kde (in 11.0) and so on.... will attach the 11.1 logs here after this message. can i be of any more help? greetings. Created attachment 268426 [details]
cleaninstall 11.1 x86, logs after the final step. first login
cleaninstall 11.1 x86, logs after the final step. first login
normal boot of 11.1 x86 (so acpi=on i guess). usb mouse/keyboard almost unusable if the system is completely idle (=no cpu load)
once there is a little cpu load, the usb components are more reactive and dont create fake events and/or usb events get recognized properly.
Created attachment 268427 [details]
logs from 11.1 x86 bootup normally with acpi=off as addtl. parameter
logs from 11.1 x86 bootup normally with acpi=off as addtl. parameter
once being booted with acpi=off as additional parameter (grub), the system boots up really quickly, and usb mice/keyboards work normally.
acpi is apparently the only parameter that essentially messes up the system (lags, usb problems, etc....)
update: i am de-bugging another bug (xorg doesnt start up on this system with my ati/amd radeon card), so i was testing 11.1 x86 kde4 live cd version. what i am wondering now is, that this 11.1 x86 live cd system boots up quickly and except for the xorg problems "normally". this 11.1 x86 kde4 live system doesnt show the laggy usb keyboard (could test mouse due to lack of working xorg). also there is no system-load and the booted system sits perfectly idle it seems. i have created the logs in an attachment over at my other 11.1 xorg bug: https://bugzilla.novell.com/attachment.cgi?id=268461&action=edit please see the boot.log file and others there. you will notice that it doesnt say acpi=off anywhere in the kernel parameters. ------ from boot.msg <5>Kernel command line: BOOT_IMAGE=linux initrd=initrd ramdisk_size=512000 ramdisk_blocksize=4096 splash=silent vga=0x345 and i see lots of acpi messages when i do dmesg | more so acpi is basically running fine with this kde4 live version of 11.1 x86 so whats the difference between the kernels of 11.1 normal install/running system and the 11.1 kde4 live kernel? maybe this is the way to figure out whats going wrong and whats the difference or regression bug(s) between these kernels. thanks and regards. i am still suffering from this bug. opensuse 11.1 and 11.0 is unsubale on this current asus mainboard. so i was looking around for these "kernel of the day" thingies and found in ftp://ftp.suse.com/pub/projects/kernel/kotd/SL111_BRANCH/i586/ the pae files, and downloaded the three rpms ftp://ftp.suse.com/pub/projects/kernel/kotd/SL111_BRANCH/i586/kernel-pae-2.6.27.15-SL111_BRANCH_20090207032226_dcf3eb30.i586.rpm ftp://ftp.suse.com/pub/projects/kernel/kotd/SL111_BRANCH/i586/kernel-pae-base-2.6.27.15-SL111_BRANCH_20090207032226_dcf3eb30.i586.rpm ftp://ftp.suse.com/pub/projects/kernel/kotd/SL111_BRANCH/i586/kernel-pae-extra-2.6.27.15-SL111_BRANCH_20090207032226_dcf3eb30.i586.rpm and did an rpm -Uvh --force --nodeps on all three rpms after rebooting 11.1 with this kernel, i dont have to supply any additional parameters to turn off acpi or anything. dmesg displays loads of acpi things, and usb mouse/keyboard work normal and well it seems. so this kernel doesnt feature the acpi bug related to usb keyboard/mouse sluggishness as i have displayed here in this bug. will attach dmesg log could anybody incorporate whatever fixes and patches necessary to make the official 11.1 kernel work with acpi? maybe the attached logs and output messages will help. would be great. or other than that, is there any faq or best practices document around on how to work with the kernel of the day kernels, and how to apply to current 11.1 or 11.0 systems? thanks for helping and fixing. Created attachment 272605 [details]
kernel of the day bootup messages - dmesg
kernel of the day bootup messages - dmesg
kernel boots up fine and usb mice/keyboards are usable normally with normal acpi settings.
no "acpi=off" is needed with this kernel to boot up and use opensuse 11.1 (x86)
thanks and regards.
Can someone try the following the boot options?
a. idle=poll
b. nolapic_timer
Thanks.
Will you please attach the output of acpidump, kernel configuration?
Thanks.
sorry for the delay, but something is driving me crazy with this system. basically: the acpi problem with usb mouse/keyboard is somehow non-existant any more. odd enough: the system idled like a week or two down below my desk and was turned off, but never modified (hardware) nor anything else changed on it. at the moment (yesterday) i have tried to create the acpidump information, and i realized, that both opensuse 11.0 rtm (and also all online_updates) and opensuse 11.1 rtm (and all updates too) states dont give me the sluggish usb mouse/keyboard behaviour any more, and both systems boot up fine and normally (quickly). i am simply at a lack of explanation for this behaviour, at least i think i am not that retarded to complain and create extenisve bugreports about non-existant bugs. i didnt mess with any bios settings on this board either. so i tried to clean install opensuse 11.0 x86, and cleaninstall 11.1 x86. attached the usb keyboard/mouse in all kinds of different usb ports on the mainboard atx panel. the usb problem doesnt seem to exist any more. never the less, i have created dmesg, /var/log and acpidump files for 11.0 right after installer, after onlineupdates and also for 11.1 right after install and after online_updates and also with different usb-port mouse/keyboard combinations. maybe some expert can see any differences when comparing these latest logs with the early logs that i have already provided in this bugreport. thanks for the help and sorry for the turmoil if i created any. maybe there is still some bug in the kernels, but i currently have no idea where the heck those usb mouse/keyboard lags and loss of usb events came from. will upload an archive with all the recent 11.0 and 11.1 logs. the lower numbers in logfiles are the early (after setup) ones, the higher numbers are the later ones (after onlineupdates, with various usb keyboard/mouse positions) thanks. :( Created attachment 273206 [details]
11.0 and 11.1 x86 clean logs and after various onlineupdates
here is the archive for dmesg, acpidump and /var/log for 11.0 and 11.1 clean install, and also after online_update.
lower numbers are the earlier stages in each set.
greetings.
Strange. If you're unable to reproduce and powering the system off for a while seems to have fixed it, I'm going to blame hardware. If the problem reoccurs, please reopen this report. |