4:46 p.m.
Hello list,
I have a dual core Dell XPS and have been experiencing mysterious lockups
every now and then with FC4 and the newlly installed FC5.
The kernel version is 2.6.15-1.2054_FC5smp and everything appears to be
running smoothly until the keyboard and mouse lockup (USB
both) and the whole system appears to freeze. The only activity I see is on
the gnome-netstatus-applet on the panel. Anybody
else experiencing something similar? Any clues? Any ideas? Anything would be
greatly appreciated!
4:55 p.m.
On 4/17/06, Filippos Klironomos <presariod(a)gmail.com> wrote:
Hello list,
I have a dual core Dell XPS and have been experiencing mysterious lockups
every now and then with FC4 and the newlly installed FC5.
The kernel version is 2.6.15-1.2054_FC5smp and everything appears to be
running smoothly until the keyboard and mouse lockup (USB
both) and the whole system appears to freeze. The only activity I see is on
the gnome-netstatus-applet on the panel. Anybody
else experiencing something similar? Any clues? Any ideas? Anything would be
greatly appreciated!
have you checked /var/log/messages for any hints?
--
As a boy I jumped through Windows, as a man I play with Penguins.
5:04 p.m.
Yep, aboslutely nothing is logged in after a lockup. I wish I could have a
hint or a lead. The only thing I suspect is the "reversed-enineered"
ethernet driver 'forcedeth' for the nForce NVIDIA card, but that is only a
suspicion. I also removed the gnome-netstatus-applet from the panel and will
wait to see if I have another lockup.
Thanks for answering!
On 4/17/06, Arthur Pemberton <pemboa(a)gmail.com> wrote:
On 4/17/06, Filippos Klironomos <presariod(a)gmail.com> wrote:
> Hello list,
>
> I have a dual core Dell XPS and have been experiencing mysterious
lockups
> every now and then with FC4 and the newlly installed FC5.
> The kernel version is 2.6.15-1.2054_FC5smp and everything appears to be
> running smoothly until the keyboard and mouse lockup (USB
> both) and the whole system appears to freeze. The only activity I see is
on
> the gnome-netstatus-applet on the panel. Anybody
> else experiencing something similar? Any clues? Any ideas? Anything
would be
> greatly appreciated!
>
have you checked /var/log/messages for any hints?
--
As a boy I jumped through Windows, as a man I play with Penguins.
--
fedora-list mailing list
fedora-list(a)redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
5:08 p.m.
Filippos Klironomos wrote:
The kernel version is 2.6.15-1.2054_FC5smp and everything appears to be
running smoothly until the keyboard and mouse lockup (USB
both) and the whole system appears to freeze. The only activity I see is
on the gnome-netstatus-applet on the panel. Anybody
else experiencing something similar? Any clues? Any ideas? Anything
would be greatly appreciated!
If you switch to a VT (Ctrl+Alt+F1) and then back (Alt+F7), do the
keyboard and mouse start working again?
5:28 p.m.
If you switch to a VT (Ctrl+Alt+F1) and then back (Alt+F7), do the
keyboard and mouse start working again?
No, I tried that too. Both are completely unresponsive! There have been
cases that I could remotely
login via ssh to the machine and cleanly reboot it but in some cases I
couldn't even do that and only a hard "Power off" could restore the
system...
11:38 p.m.
On 4/17/06, Filippos Klironomos <presariod(a)gmail.com> wrote:
Hello list,
I have a dual core Dell XPS and have been experiencing mysterious lockups
every now and then with FC4 and the newlly installed FC5.
The kernel version is 2.6.15-1.2054_FC5smp and everything appears to be
running smoothly until the keyboard and mouse lockup (USB
both) and the whole system appears to freeze. The only activity I see is on
the gnome-netstatus-applet on the panel. Anybody
else experiencing something similar? Any clues? Any ideas? Anything would be
greatly appreciated!
Have you tried upgrading to the kernel-2.6.16-1.2080?
12:14 p.m.
Yep, it's the same. Today I got another lockup but this time I found the
following log messages:
kernel: ata1: command 0x35 timeout, stat 0x50 host_stat 0x24
looking around I found that disabling APIC might prevent this from happening
so I'll give it a try.
Have you tried upgrading to the kernel-2.6.16-1.2080?
--
fedora-list mailing list
fedora-list(a)redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
12:29 p.m.
On 4/18/06, Filippos Klironomos <presariod(a)gmail.com> wrote:
Yep, it's the same. Today I got another lockup but this time I found the
following log messages:
kernel: ata1: command 0x35 timeout, stat 0x50 host_stat 0x24
looking around I found that disabling APIC might prevent this from happening
so I'll give it a try.
So you have any other OSs on your machine? I would try a live cd, this
could be a hardware issue . . . quite possible the machine just needs
a thorough cleaning.
--
As a boy I jumped through Windows, as a man I play with Penguins.
12:37 p.m.
This is greatly unlikely, it's a brand new Dell XPS not even a year old. It
seems that it's a hardware issue (SATA disks) but the fact that it appeared
only when FC5 was installed on it points somewhere else. On the FC4 side the
lockups appeared only when newer kernels were installed. Back in December
the machine was running for 3 months uninterrupted doing heavy calculations
with absolutely no problem.
I'll turn off APIC and see, but this cannot be a permanent solution since
the machine is dual core...
So you have any other OSs on your machine? I would try a live cd,
this
could be a hardware issue . . . quite possible the machine just needs
a thorough cleaning.
--
As a boy I jumped through Windows, as a man I play with Penguins.
--
fedora-list mailing list
fedora-list(a)redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
12:47 p.m.
On 4/18/06, Filippos Klironomos <presariod(a)gmail.com> wrote:
This is greatly unlikely, it's a brand new Dell XPS not even a year old. It
seems that it's a hardware issue (SATA disks) but the fact that it appeared
only when FC5 was installed on it points somewhere else. On the FC4 side the
lockups appeared only when newer kernels were installed. Back in December
the machine was running for 3 months uninterrupted doing heavy calculations
with absolutely no problem.
I'll turn off APIC and see, but this cannot be a permanent solution since
the machine is dual core...
I am not trying to take away blame from Fedora, just suggesting
possible alternate problems which could be easier resolved - have you
taken a look at your BIOS? Maybe there is an option there which FC4
dealth with better or ignored.
A computer that new should have a pretty advanced BIOS setup.
--
As a boy I jumped through Windows, as a man I play with Penguins.
1:22 p.m.
Unfortunatelly that's where everything points at. The DSDT table (surprise,
surprise) might be
buggy (surprise, surprise). From what I read here:
http://www.al-najjar.co.uk/
it's most likely that, and of course there is no precompiled DSDT for me to
use, so I have to dig myself. Ahhh the good old days are back! Unfortunately
I don't have the time to do that so if the crude patch of disabling APIC
altogether works for me, I'll stick to it. The only downside is that there
is no interrupt balancing between the two CPUs but I can still use the
second, which is what matter to me for now.
If indeed this works, I'll come back with another posting for the next poor
fellow with a Dell XPS...
I am not trying to take away blame from Fedora, just suggesting
possible alternate problems which could be easier resolved - have
you
taken a look at your BIOS? Maybe there is an option there which FC4
dealth with better or ignored.
A computer that new should have a pretty advanced BIOS setup.
--
As a boy I jumped through Windows, as a man I play with Penguins.
--
fedora-list mailing list
fedora-list(a)redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
7:20 a.m.
Filippos Klironomos wrote:
Unfortunatelly that's where everything points at. The DSDT table
(surprise,
surprise) might be
buggy (surprise, surprise). From what I read here:
http://www.al-najjar.co.uk/
it's most likely that, and of course there is no precompiled DSDT for me to
use, so I have to dig myself. Ahhh the good old days are back! Unfortunately
I don't have the time to do that so if the crude patch of disabling APIC
altogether works for me, I'll stick to it. The only downside is that there
is no interrupt balancing between the two CPUs but I can still use the
second, which is what matter to me for now.
I'm not convinced by that site.
DSDTs are part of the BIOS interface to ACPI (Advanced Configuration and
Power Interface). BIOS ACPI implementations are notoriously buggy, and
can easily crash kernels. So you may need to turn that off (noacpi --
you'll lose hardware power-off).
An APIC (Advanced Programmable Interrupt Controller, which is
responsible for interrupt balancing, as you say) is pretty unrelated.
You may well be able to turn off ACPI without turning off APIC.
Hope this helps,
James.
--
E-mail address: james | For every complex problem, there is a solution
@westexe.demon.co.uk | that is simple, neat, and wrong.
1:18 p.m.
DSDTs are part of the BIOS interface to ACPI (Advanced Configuration and
Power Interface). BIOS ACPI implementations are notoriously buggy,
and
can easily crash kernels. So you may need to turn that off (noacpi --
you'll lose hardware power-off).
Passing the 'noapic' option to just turn off APIC fixed the freeze problem
along with
the 'ata1:' related messages. So that was the solution. I can't turn off
ACPI completely
because not only I lose poweroff but the second CPU (it's a dual core
machine).
An APIC (Advanced Programmable Interrupt Controller, which is
responsible for interrupt balancing, as you say) is pretty
unrelated.
You may well be able to turn off ACPI without turning off APIC.
Yep, that's what I did and believe it or not the system has been up and
running
for 5 days now with no freezes, when it took less than 24hours to freeze
previously. So I still have ACPI managing the second CPU (the system is
admitedly slower compared to when APIC is on) and for now I'm happy
with the solution. Older FC4 kernel versions did not freeze the machine so
I'm hoping eventually a newer kernel will fix this.
Filippos
6:46 p.m.
New subject: [UPDATE] Re: mysterious keyboard, mouse lockups with FC5
Digging around on this problem I have an update to announce. Instead of
turning off APIC completely
by passing the 'noapic' option to the kernel, I instead forced the local
APIC to take over with the 'lapic'
commnad. This enabled interrupt sharing (and increase of speed) between the
2 CPUs so it is
an improvement!
The system has been up for 5 days now with no freezes.
Filippos
On 4/24/06, Filippos Klironomos <presariod(a)gmail.com> wrote:
DSDTs are part of the BIOS interface to ACPI (Advanced Configuration and
> Power Interface). BIOS ACPI implementations are notoriously buggy, and
> can easily crash kernels. So you may need to turn that off (noacpi --
> you'll lose hardware power-off).
Passing the 'noapic' option to just turn off APIC fixed the freeze problem
along with
the 'ata1:' related messages. So that was the solution. I can't turn off
ACPI completely
because not only I lose poweroff but the second CPU (it's a dual core
machine).
An APIC (Advanced Programmable Interrupt Controller, which is
> responsible for interrupt balancing, as you say) is pretty unrelated.
> You may well be able to turn off ACPI without turning off APIC.
Yep, that's what I did and believe it or not the system has been up and
running
for 5 days now with no freezes, when it took less than 24hours to freeze
previously. So I still have ACPI managing the second CPU (the system is
admitedly slower compared to when APIC is on) and for now I'm happy
with the solution. Older FC4 kernel versions did not freeze the machine so
I'm hoping eventually a newer kernel will fix this.
Filippos
10:31 p.m.
Hi James,
It tooks me more than a week to fine tune my new FC5_64 box. Now I can
use onboard LAN, onboard sound and adding a PCI-E card without
suffering from PC hanging.
DSDTs are part of the BIOS interface to ACPI (Advanced Configuration
and
Power Interface). BIOS ACPI implementations are notoriously buggy,
and
can easily crash kernels. So you may need to turn that off (noacpi --
you'll lose hardware power-off).
Poweroff without PC power-off is one of the remaining problem on this
FC5_64 box. The cause is adding "noacpi" on kernel option at boot.
Without adding it PC will hang frequently. I don't know whether
upgrade BIOS will be its solution. OR there is other remedy.
Any suggestion? TIA
B.R.
SL
11:37 a.m.
Stephen Liu wrote:
Poweroff without PC power-off is one of the remaining problem on
this
FC5_64 box. The cause is adding "noacpi" on kernel option at boot.
Without adding it PC will hang frequently. I don't know whether
upgrade BIOS will be its solution. OR there is other remedy.
Any suggestion? TIA
Well, there are two ways to do soft power-off. APM doesn't work in SMP
mode, so you'd have to turn off one of your cores (you've got a dual
core box, IIRC).
The other way is APCI.
Yes, getting a new BIOS sounds like the best approach. Otherwise,
there's a small chance that new kernels will be able to work around
whatever bug you're experiencing.
I assume that you've moaned to your vendor...
James.
--
E-mail address: james | For every complex problem, there is a solution
@westexe.demon.co.uk | that is simple, neat, and wrong.
6556
days inactive
6577
days old
15 comments
6 participants
participants (6)
-
Arthur Pemberton -
Filippos Klironomos -
Gordon Messmer -
James Wilkinson -
Kam Leo -
Stephen Liu