| View previous topic :: View next topic |
| Author |
Message |
elestedt
Guru
Joined: 13 Mar 2005
Posts: 383
|
 Posted: Fri Sep 02, 2005 6:28 am Post subject: |
 |
|
| xtaski wrote: | Can I ask a dumb question? (ok, I will)
What does it mean that in 2.6.13 we can load another kernel from a currently executing kernel using Kexec? What's the benefit/ what can you do with this? It sounds cool, I'm just wondering what the implication is... |
You don't have to reboot when you've installed a new kernel. |
|
| Back to top |
|
 |
kottlettstanze
Tux's lil' helper
Joined: 20 Apr 2004
Posts: 113
|
| Posted: Fri Sep 02, 2005 8:30 am Post subject: |
|
|
| elestedt wrote: | | xtaski wrote: | Can I ask a dumb question? (ok, I will)
What does it mean that in 2.6.13 we can load another kernel from a currently executing kernel using Kexec? What's the benefit/ what can you do with this? It sounds cool, I'm just wondering what the implication is... |
You don't have to reboot when you've installed a new kernel. |
Well, it is sort of a reboot, so you lose all running apps. The kernel is loaded and init starts up your system. The only advantage is, that your mainboard doesn't reboot. So no bios, no bootloader but a direct boot into the new kernel. |
|
| Back to top |
|
|
dsd
Developer
Joined: 30 Mar 2003
Posts: 2162
Location: nr London
|
| Posted: Fri Sep 02, 2005 11:03 am Post subject: |
|
|
| Matteo Azzali wrote: | I'm still having this issue,
from dmesg: |
you should file a bug (and include the entire message, you missed 1 or 2 lines from the top when you pasted it here)
| elestedt wrote: | | I can't even boot it - goes into some endless loop at boot when ACPI is initialized (ACPI is compiled into the kernel)... Something about some event not being handled. |
you should file a bug. https://bugs.gentoo.org
_________________
http://dev.gentoo.org/~dsd |
|
| Back to top |
|
|
Raistlin
l33t
Joined: 17 May 2004
Posts: 691
Location: Boston, MA
|
| Posted: Fri Sep 02, 2005 11:11 am Post subject: |
|
|
| kottlettstanze wrote: | | elestedt wrote: | | xtaski wrote: | Can I ask a dumb question? (ok, I will)
What does it mean that in 2.6.13 we can load another kernel from a currently executing kernel using Kexec? What's the benefit/ what can you do with this? It sounds cool, I'm just wondering what the implication is... |
You don't have to reboot when you've installed a new kernel. |
Well, it is sort of a reboot, so you lose all running apps. The kernel is loaded and init starts up your system. The only advantage is, that your mainboard doesn't reboot. So no bios, no bootloader but a direct boot into the new kernel. |
All the hardware initialization is "skipped". No POST, no BIOS' device-discovering etc. This is usually the most time-consuming part of boot. Especially for larger computers: mainframes etc. You will see, that even the s390 (IBM zSeries) has kexec-code in the kernel 
Cheers, R.
_________________
Zwei Was Eins Initially
"Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." |
|
| Back to top |
|
|
frenkel
Veteran
Joined: 13 May 2003
Posts: 1034
Location: .nl
|
| Posted: Fri Sep 02, 2005 3:11 pm Post subject: |
|
|
| Guglie wrote: | in this kernel the suspend does not work with in the "traditional" way:
echo 3 > /proc/acpi/sleep for RAM suspend does not work (on my laptop the file /proc/acpi/sleep does not exist)
in 2.6.13 kernel , as written in the documentation of the kernel (Documentation/power), the file is now /sys/power/state and has 3 states:
- standby
- mem (ram suspend)
- disk (disk suspend)
but on my T42 only suspend-to-ram works fine |
/sys/power/state already exists since < 2.6.10 and the documentation already advised everybody to use that. They have now just removed the /proc/acpi/sleep thing. |
|
| Back to top |
|
|
xtaski
Apprentice
Joined: 20 Dec 2004
Posts: 168
Location: New York, NY
|
| Posted: Fri Sep 02, 2005 9:08 pm Post subject: |
|
|
| Thanks everyone - makes more sense now. |
|
| Back to top |
|
|
yaq
n00b
Joined: 02 Sep 2005
Posts: 50
Location: Poland/Krakow
|
| Posted: Fri Sep 02, 2005 9:59 pm Post subject: |
|
|
| elestedt wrote: | I can't even boot it - goes into some endless loop at boot when ACPI is initialized (ACPI is compiled into the kernel)... Something about some event not being handled.
Anyone know anything about this? |
hp laptop?
http://bugzilla.kernel.org/show_bug.cgi?id=5161 |
|
| Back to top |
|
|
spiralvoice
Apprentice
Joined: 09 Mar 2004
Posts: 238
|
| Posted: Fri Sep 02, 2005 10:37 pm Post subject: |
|
|
| Matteo Azzali wrote: | I'm still having this issue,
from dmesg:
| Code: | | Process kgameportd (pid: 851, threadinfo=dfa8a000 task=dfca95d0) |
It is probabily a mine issue. Anyone knows what this mean? |
I have this issue, too. |
|
| Back to top |
|
|
elestedt
Guru
Joined: 13 Mar 2005
Posts: 383
|
| Posted: Sat Sep 03, 2005 5:48 am Post subject: |
|
|
| yaq wrote: | | elestedt wrote: | I can't even boot it - goes into some endless loop at boot when ACPI is initialized (ACPI is compiled into the kernel)... Something about some event not being handled.
Anyone know anything about this? |
hp laptop?
http://bugzilla.kernel.org/show_bug.cgi?id=5161 |
As a mater of fact: yes.
Thanks for the link. |
|
| Back to top |
|
|
kottlettstanze
Tux's lil' helper
Joined: 20 Apr 2004
Posts: 113
|
| Posted: Sat Sep 03, 2005 8:39 am Post subject: |
|
|
I got it running now 
The ipw2200 thing seems to be a router-issue as well, which is why not everybody is affected. Compnerd and brix fixed that issue, it is in portage now (ipw2200-1.0.6-r2).
The issue with powering off, is a software suspend bug. It is fixed in suspend2-sources-2.6.13-r1 (patch is in /usr/portage/sys-kernel/suspend2-sources/files/linux-2.6.13-poweroff.patch).
Well, the last issue still exists: If I log out from KDE (plain logout, w/o choosing reboot or power off), the notebook starts to hibernate during the logout process. Hitting escape or just resuming, doesn't help. As soon the sys is up again, it starts to hibernate. Only the noresume2 kernel line helps then 
This might be a KDE thing, perhaps klaptopdaemon is going weird. |
|
| Back to top |
|
|
Jakub
Guru
Joined: 04 Oct 2003
Posts: 377
Location: Warsaw, Poland
|
| Posted: Sat Sep 03, 2005 10:21 am Post subject: |
|
|
| dsd wrote: |
| Jakub wrote: | I don't know if these can be considered bugs but:
1) the nvidia device nodes are not created (maybe some warning when installing the kernel should be issued?) |
it "just works" here (v1.0.7676) - i'm guessing that the nvidia-glx component creates the device nodes. please specify which version is giving you problems..
| Quote: | | 2) there must have been some changes in the 2.6.13-rcX kernels' iptables code - loopback connections made by KDE were treated as invalid, so if you blocked invalid connections first and then allowed loopback connections, kde would hang at "Initializing peripherals" (it would load eventually, after several minutes). I'm not sure if it's fixed by now. |
this bug was fixed in the 2.6.12 series shortly after it hit bugzilla (which was weeks after it was discussed on the forums.. ) - are you saying that this bug is still present, or just speculating that it might be?
|
1) As for nvidia, I have the nvida-kernel-1.0.6629-r4 and nvidia-glx-1.0.6629-r6 installed which doesn't work for the 2.6.13 kernel (because of the missing device nodes; the problem already occurred on one of the 2.6.13-rcX kernels). I use those 'old' versions because for me (and almost everybody else, I guess) the renderAccel option in xorg doesn't work with the nvidia 7XXX drivers.
2) The iptables code has apparently been fixed, thanks. |
|
| Back to top |
|
|
474
l33t
Joined: 19 Apr 2002
Posts: 714
|
| Posted: Sat Sep 03, 2005 2:02 pm Post subject: |
|
|
| HRH_H_Crab wrote: | very depressing. speedtouch seems broken. rolled back to 2.6.12 and all well again.
first time for ages ive not been able to run the latest stable kernel. |
I have an explanation for this. The modem_run tool which is used to upload the firmware to the modem simply will not work with 2.6.13. Unfortunately, the net-dialup/speedtouch package makes use of this tool.
As of 2.6.10, the speedtch module has acquired the ability to make use of the in-kernel firmware loading mechanism and this is how it should really be done now. For now please take a look at a related bug and this wiki entry which explains how to accomplish this. Note that you do not actually need to install the speedtouch package at all, but you do still need to use ppp built with atm support. The only aspect that may initially seem unusual is that you have to "split" the firmware file into two using a firmware extraction tool. But it's not difficult and works very well
I am co-operating with Paul Bredbury (the author of that wiki entry) in order to devise a new package which will fully support the use of a speedtouch modem in gentoo in conjunction with >=2.6.13 kernels. Hopefully this will develop into something that could make for a feasible replacement for the current speedtouch package. |
|
| Back to top |
|
|
Vim_Fuego
n00b
Joined: 25 Jun 2004
Posts: 23
Location: Rotterdam, The Netherlands
|
| Posted: Sun Sep 04, 2005 7:49 am Post subject: |
|
|
Recently installed 2.6.13. Worked fine yesterday, but today dmesg and Xorg are complaining that it can't find / read fixed fonts. I tracked that problem back to a problem with reading (XFS?) partitions with the 2.6.13 kernel.
For example, with the 2.6.13 kernel I was not not able to read the /usr/share dir, doing an a simple ls command results in an unknown error. With the 2.6.11 kernel the directory reads just fine. Any ideas???? |
|
| Back to top |
|
|
dsd
Developer
Joined: 30 Mar 2003
Posts: 2162
Location: nr London
|
| Posted: Tue Sep 06, 2005 5:08 pm Post subject: |
|
|
| yaq wrote: | | elestedt wrote: | I can't even boot it - goes into some endless loop at boot when ACPI is initialized (ACPI is compiled into the kernel)... Something about some event not being handled.
Anyone know anything about this? |
hp laptop?
http://bugzilla.kernel.org/show_bug.cgi?id=5161 |
looks like a solution is available on that bug, could one of you please file a bug report at https://bugs.gentoo.org describing the problem and referencing the kernel bugzilla URL? thanks
_________________
http://dev.gentoo.org/~dsd |
|
| Back to top |
|
|
elestedt
Guru
Joined: 13 Mar 2005
Posts: 383
|
| Posted: Tue Sep 06, 2005 5:29 pm Post subject: |
|
|
| dsd wrote: | | yaq wrote: | | elestedt wrote: | I can't even boot it - goes into some endless loop at boot when ACPI is initialized (ACPI is compiled into the kernel)... Something about some event not being handled.
Anyone know anything about this? |
hp laptop?
http://bugzilla.kernel.org/show_bug.cgi?id=5161 |
looks like a solution is available on that bug, could one of you please file a bug report at https://bugs.gentoo.org describing the problem and referencing the kernel bugzilla URL? thanks |
Done - https://bugs.gentoo.org/show_bug.cgi?id=105046
It might not be the best bug-report - but it gets the message across. |
|
| Back to top |
|
|
dsd
Developer
Joined: 30 Mar 2003
Posts: 2162
Location: nr London
|
| Posted: Sun Sep 11, 2005 3:01 pm Post subject: |
|
|
| spiralvoice wrote: | | Matteo Azzali wrote: | I'm still having this issue,
from dmesg:
| Code: | | Process kgameportd (pid: 851, threadinfo=dfa8a000 task=dfca95d0) |
It is probabily a mine issue. Anyone knows what this mean? |
I have this issue, too. |
on the bug report it appears that this problem is only present on gentoo-sources-2.6.13 and not on the vanilla-sources-2.6.13 release. this is quite puzzling in itself. would you mind also testing vanilla-sources-2.6.13 to get a 2nd confirmation that this is a gentoo-specific problem? thanks.
_________________
http://dev.gentoo.org/~dsd |
|
| Back to top |
|
|
elestedt
Guru
Joined: 13 Mar 2005
Posts: 383
|
| Posted: Sun Sep 11, 2005 3:30 pm Post subject: |
|
|
| dsd wrote: | | spiralvoice wrote: | | Matteo Azzali wrote: | I'm still having this issue,
from dmesg:
| Code: | | Process kgameportd (pid: 851, threadinfo=dfa8a000 task=dfca95d0) |
It is probabily a mine issue. Anyone knows what this mean? |
I have this issue, too. |
on the bug report it appears that this problem is only present on gentoo-sources-2.6.13 and not on the vanilla-sources-2.6.13 release. this is quite puzzling in itself. would you mind also testing vanilla-sources-2.6.13 to get a 2nd confirmation that this is a gentoo-specific problem? thanks. |
already done - it is there on vanilla-sources. |
|
| Back to top |
|
|
Riekr
n00b
Joined: 21 Jun 2002
Posts: 47
Location: Italy
|
| Posted: Tue Sep 13, 2005 11:29 am Post subject: |
|
|
i tried to install 2.6.13 and 2.6.13-r1 (whole system with ~x86) but when booting the system will take half an our to come up, it seems that when reaching init it takes a while to execute anything.
tried many configurations but really now i don't know what to change
any clues?
_________________
Our continuing mission to boldly code where no one has man page 4. |
|
| Back to top |
|
|
MerlinTheWizard
Apprentice
Joined: 06 Apr 2003
Posts: 270
|
| Posted: Tue Sep 13, 2005 11:33 am Post subject: |
|
|
| Riekr wrote: | i tried to install 2.6.13 and 2.6.13-r1 (whole system with ~x86) but when booting the system will take half an our to come up, it seems that when reaching init it takes a while to execute anything.
tried many configurations but really now i don't know what to change
any clues? |
Uh huh, see if this is not actually a baselayout problem. |
|
| Back to top |
|
|
Riekr
n00b
Joined: 21 Jun 2002
Posts: 47
Location: Italy
|
| Posted: Tue Sep 13, 2005 4:10 pm Post subject: |
|
|
i'm using baselayout-1.12.0_pre8-r2 anyone using my same version?
_________________
Our continuing mission to boldly code where no one has man page 4. |
|
| Back to top |
|
|
MerlinTheWizard
Apprentice
Joined: 06 Apr 2003
Posts: 270
|
| Posted: Tue Sep 13, 2005 6:27 pm Post subject: |
|
|
| Riekr wrote: | | i'm using baselayout-1.12.0_pre8-r2 anyone using my same version? |
Currently I'm using baselayout-1.12.0_pre8-r2 as well...
Not sure what your problem is. At what point during the boot process does it seem to hang exactly? |
|
| Back to top |
|
|
UTgamer
Veteran
Joined: 10 Aug 2003
Posts: 1326
Location: Troisdorf (Köln) Deutschland
|
| Posted: Wed Sep 14, 2005 12:15 am Post subject: |
|
|
Hey, kernel 2.6.13-gentoo-r1 was a must be for me, due too I have an AMD64 x2 on a MSI Mainbord with nForce 4 chipset with an PCIe GF 7800 grafik card and I liked to use the onboard ca0106 sound chip capture functionality.
Nothing, niente, nichts, no failure at all.
I'm so glad with this kernel, you can't believe.
It is so speedy, I didn't see that with kernel-2.6.12-gentoo-r10, which I did use until today.
Thanks the kernel team for this.
_________________
AMD Phenom II x4 >> CFLAGS="-march=amdfam10 -O2 -mmmx -msse3 -mfpmath=sse,387 -pipe -ffast-math" is stable and here in use.
Did Intel produce at any time bugfree HW?
http://www.urbanmyth.org/microcode/
http://www.heise.de/newsticker/meldung/91748 |
|
| Back to top |
|
|
Riekr
n00b
Joined: 21 Jun 2002
Posts: 47
Location: Italy
|
| Posted: Wed Sep 14, 2005 8:18 am Post subject: |
|
|
| MerlinTheWizard wrote: | | Not sure what your problem is. At what point during the boot process does it seem to hang exactly? |
It does not really hang, it only take a huge amount of time to do anything, during this time no hdd activity is shown. For example, populating udev takes a couple of seconds with 2.6.12 with hdd led always on, with 2.6.13 it takes about 30-60 seconds and with very low hdd activity.
Once x&gnome are up the system is slower but i just used it for few minutes then rolled back to 2.6.12-r10 since i was in need to restart to work
I'm using gentoo-sources on an asus l5g notebook (i865, p4m ht, 1gb mem, ati-drivers, regparams, preempt). The problem doesn't disappear if i boot without splash.. i'm starting to think it's something related to the hd controller even if it shouldn't be so exotic..
The only other time that i've seen something similar was when i switched from iso-8865-15 (i think) to utf8 and made the first boot.. i got a big delay when init was cleaning /tmp but it happened only the first time.
locale output:
| $ locale wrote: | LANG=it_IT.utf8
LC_CTYPE="it_IT.utf8"
LC_NUMERIC="it_IT.utf8"
LC_TIME="it_IT.utf8"
LC_COLLATE="it_IT.utf8"
LC_MONETARY="it_IT.utf8"
LC_MESSAGES="it_IT.utf8"
LC_PAPER="it_IT.utf8"
LC_NAME="it_IT.utf8"
LC_ADDRESS="it_IT.utf8"
LC_TELEPHONE="it_IT.utf8"
LC_MEASUREMENT="it_IT.utf8"
LC_IDENTIFICATION="it_IT.utf8"
LC_ALL=it_IT.utf8
|
after a su-
| # locale wrote: | LANG=
LC_CTYPE="POSIX"
LC_NUMERIC="POSIX"
LC_TIME="POSIX"
LC_COLLATE="POSIX"
LC_MONETARY="POSIX"
LC_MESSAGES="POSIX"
LC_PAPER="POSIX"
LC_NAME="POSIX"
LC_ADDRESS="POSIX"
LC_TELEPHONE="POSIX"
LC_MEASUREMENT="POSIX"
LC_IDENTIFICATION="POSIX"
LC_ALL=
|
_________________
Our continuing mission to boldly code where no one has man page 4. |
|
| Back to top |
|
|
dsd
Developer
Joined: 30 Mar 2003
Posts: 2162
Location: nr London
|
| Posted: Wed Sep 14, 2005 9:43 am Post subject: |
|
|
Riekr, boot with init=/bin/bash and see if some basic programs are still slow even at that point
_________________
http://dev.gentoo.org/~dsd |
|
| Back to top |
|
|
Riekr
n00b
Joined: 21 Jun 2002
Posts: 47
Location: Italy
|
| Posted: Wed Sep 14, 2005 12:39 pm Post subject: |
|
|
I used init=/bin/bash and launched some commands using "time". Thinking is a disk access problem i did some tests:
time find /usr/local > /dev/null
is giving me real times from 20ms to 1.7secs excluding the first run which should fill the cache. I tried the same command in 2.6.12-r10 and times excluding first run are from 5 to 7 ms. I'm not very sure it's related to disk acces.. maybe some scheduler issue? or maybe reiserfs?
I saved dmesg output from the 2 versions..:
2.6.13
Calibrating delay using timer specific routine.. 5991.81 BogoMIPS (lpj=11983624)
Calibrating delay using timer specific routine.. 5985.48 BogoMIPS (lpj=11970977)
Total of 2 processors activated (11977.30 BogoMIPS).
2.6.12
Calibrating delay loop... 5931.00 BogoMIPS (lpj=2965504)
Calibrating delay loop... 5980.16 BogoMIPS (lpj=2990080)
Total of 2 processors activated (11911.16 BogoMIPS).
lpj is about 5 times greater with 2.6.13
My mount at init:
| # cat mount.txt wrote: | rootfs on / type rootfs (rw)
/dev/root on / type reiserfs (rw,noatime)
proc on /proc type proc (rw,nodiratime)
sysfs on /sys type sysfs (rw)
udev on /dev type tmpfs (rw,nosuid)
devpts on /dev/pts type devpts (rw)
usbfs on /proc/bus/usb type usbfs (rw) |
I'm using rtclock support and hpet, may something make me run into weird?
_________________
Our continuing mission to boldly code where no one has man page 4. |
|
| Back to top |
|
|
|
Kernel & Hardware
