You state "x86_64-195.36.15 beta nvidia driver", … I assume you mean the nVidia proprietary graphic driver. Last I looked the 195.36.15 was not a beta, but rather it was the official release. For example according to the nvidia forum: Current NVIDIA Linux graphics driver releases - nV News Forums I know, a trivial inaccuracy, but where there is smoke there can be fire … ?
What happens when you reboot , and from there go direct to run level 3 (do NOT do this via “init 3”) and login as a regular, switch to root, and then run:
You are right. I checked the nvidia website and it turns out that 195.36.15 is listed as official. The interesting thing is that the openSUSE nvidia repos contain the 195.30 driver which nvidia lists as being beta. Go figure.
It’s obvious after some more investigating that there is no (or little) hardware acceleration happening with my install of Milestone 4. I think I’ll live with it for a while instead of using an xorg.conf. I’ll see what happens after future updates.
This also happened while installing from DVD build 0515 64bit. It also didn’t save the users home directory, consequently after reboot was not able to log in as user. Logged in as root and did ls on /home as below:
After reboot: Login could not enter /home/tom
laptop:~ # ls /home
lost+found <<<<<<<<<< no folder for user tom
laptop:~ #
This is interesting as the installer did create the user whose credentials are shown below (user is tom):
Sole user is tom. When that user failed, added bob.
From /etc/passwd:
bob:x:1001:100:Robert Webster:/home/bob:/bin/bash
tom:x:1000:100:Thomas Taylor:/home/tom:/bin/bash
From /etc/shadow:
bob:$2a$05$XliFOz78lP(25removed)LvPwO0y7pdEw3VB18O:14695:0:99999:7:::
tom:$2a$05$STpzYL43Ca(25removed)jbZ2cCvVa1qEZTNL7.:14695:0:99999:7:::
From /etc/group:
video:x:33:bob,tom
I had added the user “bob” to test and his home directory was created properly and could login without problem.
After creating a directory under /home for “tom”, am able to login and use that account with a few quirks. The icons under kickoff > leave are those normally assigned to root while in root they are the user style (full color).
Although my usb mouse/keyboard froze near the end of my installation, I did a cold boot. After the reboot, the mouse and keyboard were functionibg correctly.
Ran accross another problem carried over from 11.3-M3. Akonadi server self-test is still failing under M4. When I reinstalled with lxde and installed kde4 4.4.1 SC from kde svn, it works fine so apparently the problem is in the openSuSE implementation. Bug 591693 filed about this. It causes a problem with kmail not storing status changes and freezing on composition.
I encountered some very minor anomalous behaviour in LXDE desktop in a 32-bit M4 install (from DVD) of LXDE with file manager PCManFM when using sftp and when in the root directory of the remote PC. I wrote a bug report on it here: Access Denied
I booted my Dell Studio 1537 laptop (64-bit P8400 with ATI Radeon HD3450 graphics and Intel 5300 AGN wireless) to the M4 KDE4 64-bit liveCD.
Graphics worked with radeonhd opensource driver at 1920x1200 resolution (I checked driver in /var/log/Xorg.0.log and resolution with xrandr). Sound worked. Webcam worked.
The Device Notifier for hot plugging USB sticks was inexplicably missing ! So when I plugged in a USB stick, no indication that it was plugged in. But USB stick did show up in “fdisk -l” and could be manually mounted. Has anyone else observed this ?
Wireless was a bit cr*ppy, with the connection constantly dropping and restarting. I’ve seen this behaviour before in 11.1 and 11.2, but not this bad. Possibly not being able to disable ipv6 made it worse, and possibly the Intel wireless driver just got worse. I’m not certain, but wireless functionality appears worse (albeit the Network Manager ( ? ) GUI for connecting has improved). I need to check to see if anyone has raised a bug report on the Intel wireless driver in the kernel.
On 03/29/2010 01:16 PM, oldcpu wrote:
> Wireless was a bit cr*ppy, with the connection constantly dropping and
> restarting. I’ve seen this behaviour before in 11.1 and 11.2, but not
> this bad. Possibly not being able to disable ipv6 made it worse, and
> possibly the Intel wireless driver just got worse. I’m not certain, but
> wireless functionality appears worse (albeit the Network Manager ( ? )
> GUI for connecting has improved). I need to check to see if anyone has
> raised a bug report on the Intel wireless driver in the kernel.
You CAN disable ipv6. Either use ipv6.disable=1 in the GRUB boot option line
(temporary fix), or add it to the startup line in /boot/grub/menu.lst
(permanent). I do not have Intel wireless hardware, thus I cannot comment on
that version of their driver. With most of my drivers, NM works. There is a
problem with kernel 2.6.33-6.3 and the p54usb driver - it does not transmit, but
when I build that same kernel from source, it works fine. NM is the only
solution I use for systems with wireless.
I booted to a liveCD (64-bit KDE4 M4), as I am not keen on installing milestone software on this laptop (I use it often for business and even I am very conservative as to what I put on it as I take no risks). I suppose I could see about adding “ipv6.disable=1” as an option when the liveCD boots - never tried that before
Any recommendations as to what log files I should include if I write a bug report on the wireless behaviour (so as to minimize the effort of the developers / packagers ) … this could of course be upstream problem …
… I don’t think it our wireless router as this PC connects ok with wireless with WinXP, our Fujitsu-Siemens laptop connects ok (with openSUSE-11.1 and winXP) to the router, my wife’s ipod connects ok to the router, and a friend’s mac laptop (model ?? ) connected ok to the router.
On 03/29/2010 02:26 PM, oldcpu wrote:
>
> lwfinger;2144812 Wrote:
>> You CAN disable ipv6. Either use ipv6.disable=1 in the GRUB boot option
>> line (temporary fix), or add it to the startup line in
>> /boot/grub/menu.lst (permanent). I booted to a liveCD (64-bit KDE4 M4), as I am not keen on installing
> milestone software on this laptop (I use it often for business and even
> I am very conservative as to what I put on it as I take no risks). I
> suppose I could see about adding “ipv6.disable=1” as an option when the
> liveCD boots - never tried that before
>
> Any recommendations as to what log files I should include if I write a
> bug report on the wireless behaviour (so as to minimize the effort of
> the developers / packagers ) … this could of course be upstream
> problem …
>
> … I don’t think it our wireless router as this PC connects ok with
> wireless with WinXP, our Fujitsu-Siemens laptop connects ok (with
> openSUSE-11.1 and winXP) to the router, my wife’s ipod connects ok to
> the router, and a friend’s mac laptop (model ?? ) connected ok to the
> router.
You can certainly add the ipv6.disable=1 info to the NET install CD and the DVD.
I’m not sure about the live CD as it uses a different booting setup.
When you file a bug, please include any dmesg output about deauthentication.
I booted again to the live CD and this time I could not get the Network Manager to connect. I’m convinced I used same sequence of steps as first time. So I went to YaST and selected traditional method to connect via wireless, which worked. Worked for a while that is. Then it disconnected and refused to connect again.
I decided even though I know NOTHING about wireless, this was too much. No one else has written a bug report on 11.3 M4 on this Intel Wireless 5300 AGN that I could find, so I wrote one: Access Denied
I note this does work reasonably well on 11.1 and 11.2 although even on those two versions there are dropouts.
The above bug report was fixed with an update from factory of libfm0-0.1.9beta+git20100330-23.1, libfm-0.1.9beta+git20100330-23.1, and pcmanfm-lang-0.9.3beta+git20100324-45.1 .
However I ran into a new bug report on LXDE, where follow on logon sessions to an external computer using sftp and PCManFM as the same user fails. It does work as a different user. https://bugzilla.novell.com/show_bug.cgi?id=592432
As you will know by the time you see this, I took the info from the Bug report
and sent E-mail to the Intel group and the wireless mailing list. This problem
is in the kernel, not in anything that openSUSE has done.
Thanks for your input. I suspected that this was in the kernel with M4, but I thought it best to start here on openSUSE-11.3 M4 and see where the packagers send me from here. … I’m committed to getting this to work, so I’ll likely end up supporting any kernel bug report, regardless as to who raises it .
wrt giving support for kernel investigations, I do note while I’m comfortable installing openSUSE packaged kernel rpms, I’m not comfortable compiling my own kernel (mainly because I refuse to spend the time to learn this - I have other priorities). … I guess we all draw the line in some areas as to how far we are willing to go for support.
i just installed OS 11.3 MS 4 (DVD 64 bit) on an Thinkpad X200t. Nearly everything is working, including basic tablet-screen-setup (i didn’t figure out yet how to calibrate).
Some minor issues:
Some Fn+Fx combinations are not working (Submitted bug report)
Tablet buttons are not working, too (They never did in openSuSE before, so it’s no bug, but an unimplemented feature)
KDE compositing is not enabled by default after installation, but works without problems after enabling.
Regards, pepun.
Addendum: My X200t has the intel-cpu-whine problem (high-frequency sound from cpu when changing c-states), in OS 11.1 i simply removed the uhci_hcd-module (which i mostly don’t need) and whining had gone. In OS 11.2 in both kernel (-desktop and -default), the uhci_hcd driver was compiled into the kernel. With 11.3 MS 4 at least the -default-kernel ships with uhci_hcd configured as modul. Fine. I hope this remains until release of OS 11.3-final.
The sax2 bug report I raised was closed. It was declared a duplicate of another bug report, and a ‘work around’ to use gcc-4.3 was put in place: Access Denied
… this was called a “work around” fix, presumably because sax2 is depreciated.
lwfinger, I want to thank you publicly for your support in the Access Denied bug report, and pointing out that
This problem is known to the Intel developers and is caused by the watchdog timer kicking in for some as yet unknown reason. A set of 3 patches posted in Kernel Bugzilla (15374 – iwlagn Microcode SW error detected) has a partial fix. The watchdog timer kicks in, but the event is a lot less
disruptive.
These patches should be applied to the 2.6.33 source. They are in 2.6.34, thus
the problem will disappear from 11.3 MX as soon as the kernel changes.
This Intel PRO/Wireless 5300 AGN is on my 1+ year old Dell Studio 1537 laptop, and I use that laptop extensively for business. On occasion I also need Internet access for business, and being able to remain in Linux when working in business meetings, is very useful for me. Often my Laptop display (with Linux running) is projected on a big screen, and everyone in the meeting will see my functioning with Linux. So my having a fully working laptop (including wireless) is much appreciated by me, and I think it may have actually turned the occasional person (who attended the meetings I go to) to considering an OS other than Microsoft. I’ll likely stick with openSUSE-11.2 (recently installed) on this laptop for the near future, until the fix for this glitch is out. While I observed this on 11.1, its interesting that it got a bit progressively worse on 11.2 (but still manageable) and much worse on 11.3. I sort of blaime myself for not following up on this on 11.1 and 11.2, but it was not so severe then.
Again, thanks for your efforts here. You are a very valued member of our forum.
On 03/31/2010 09:16 AM, oldcpu wrote:
> Again, thanks for your efforts here. You are a very valued member of
> our forum.
Thanks. On the opensuse-kernel mailing list, I was told that M5 will
have the 2.6.34-rc3 kernel, and that it will be moved to Factory as soon
as it finishes building. The fixes for the IW5300 are in it.
Note: I have been running 2.6.34 kernels since -rc1. My only issues have
been with lock dependencies, and most people would never see those.
Admittedly, my main interest is in production and I do not need any
higher graphics support for my nVidia adapter than is provided by the
framebuffer driver, thus I have not tested the nouveau code.
