ntfs problem with Slackware 14.0
 

I have a problem in Slackware 14.0 with acessing my ntfs partitionon on a dual boot system from the windows manager (either KdE or Xfce)
This partition is accessible from windows without any problems. chkdsk marks the partition as without errors
In Slackware 14.0 partition is mounted in my fstab
Accessing the partition and browsing directories and files is no problem from the console with no windows manager running.
However when starting the windows manager and browsing through the partition through a file manager the file manager locks and
then backing out of the windows manager to the console prevents me accessing the partition again. (System is giving boot sector error etc, but we all know that system messages are most of the time reporting an other problem then the real one)
I tested Slacko (the Puppy Linux variant based on Slackware 13) and in Slacko I have no problem at all when using the windows manager to access the ntfs partition. Since the partition is properly mounted an acessible in the console, I don't think it is udev related.
I am expecting that it is either a problem with dbus or with policykit. Since I have no experience with either I am hoping for an answer from anyone who has and can give me directions to find the problem.

Hi SlackInstalled,
Which ntfs driver are you using?
The stock kernel driver or the ntfs-3g driver?

ntfs driver
 

I started with using the ntfs-3g driver, but also tested Slackware 14.0 with the kernel ntfs driver.
Both drivers are giving me the same problem

I'm sorry posted on the wrong thread

Try using the ntfs-3g driver and do a reboot, KDE and all these wonderful DE's use dbus, (if dbus is unaware of the change then it will still deny you access)

ntfs problem
 

Hi dev/random

As stated before, i tried both drivers, the ntfs-3g driver and the kernel ntfs driver and rebooted the system.
No result whatsoever.
By the way, I tested Slackware 13.1 and in 13.1 there is no problem at all.
Will today also test Slackware 13.37 in order to see if the regression occured there or that it only 14.0 related

SlackInstalled

ntfs problem
 

Ok,

For anyone who whants to know :

I tested Slackware 13.37 and Slackware 13.37 has the same problem as Slackware 14.0
So it seems that my ntfs regression occured beginning with the 13.37 release
Have now to investigate what major differences there are between the 13.1 and 13.37 release

SlackInstalled

During Slackware installation, you are giving a selection of choices in the way you are going to make your NTFS partition accessible.

You get the following choices, note that the driver which is going to be used is mentioned at the end. The mount options which are going to be added to "/etc/fstab" are shown at the beginning of the below lines.
You do not see that in the setup dialog while you are installing, but I left them in for informational purposes:
Can you post the line in /etc/fstab which you had for that NTFS partition while your were experiencing these issues? I run Slackware 14 on this computer in the living room which has a Windows7 partition for the rest of the family and I can browse, read and write to the NTFS partition, even from a graphical filemanager like Dolphin (I run KDE) without problems.

Please give more details about your installation. What Desktop Environment are you using, do you have a full Slackware installation, are you booting in runlevel 3 or 4?

Eric

During installation I made the following choice

"fmask=111,dmask=000" "All users can read/write to any file (ntfs-3g)

This resulted in the following line in my fstab file

/dev/sdb1 /Data ntfs-3g fmask=111,dmask=000 1 0

I prefer the KDE desktop and that is what I currently are testing with
However I also tested with Xfce and had the same problem.
The problem occurs both with Dolphin and Konquerer
The installation was the full installation with no additional packages selected (eg cups etc)
For the time being I am only using runlevel 3, but to make sure had already tested also with runlevel 4.
With runlevel 4 I have the same problem

What is strange is that when I am using Dolphin in Slackware 13.1 (Where everything works normal), my ntfs partiton shows only up as /Data in the Dolphin window, while in 14.0 it shows up as both the /Data mount point and also as a disk drive in the left pane of Dolphin. (just as if Dolphin did not recognize that it was already mounted)


SlackInstalled

Does using ntfs-3g manually work successfully?

e.g.
Perhaps if it doesn't, it may give us some clues.

CUPS is part of the 'a' package series, as is the ntfs-3g package.
What is the output of 'ls /var/log/packages | grep ntfs'?
You should see 'ntfs-3g-2012.1.15-x86_64-2' on a 64bit system.

you have problem with file managers, what about from a terminal like xterm or konsole ? can you access /Data that way at all ? also when you do ls -l /Data, who is the owner of the directory ? It might not make a difference but if the owner is root:root , you can mount it with your uid and gid as owner to see if that make any difference.

Does using ntfs-3g manually work successfully?

No, it gives me the same problems

SlackInstalled

What is the output of 'ls /var/log/packages | grep ntfs'?

bash-4.2# ls -l /var/log/packages |grep ntfs
-rw-r--r-- 1 root root 3257 Oct 20 01:16 ntfs-3g-2012.1.15-i486-2

SlackInstalled

What about from a terminal like xterm or konsole ?

Trying to access the partition within Konsole gives the same problems

SlackInstalled

This is the output on my system
There is a discrepancy in the file sizes. Try reinstalling the ntfs-3g package.

Ok , Here are som things I found out.
Looking at the syslog of Slackware 14.0 revealed the following :

Oct 21 16:48:45 darkstar org.kde.powerdevil.backlighthelper: QDBusConnection: system D-Bus connection created before QCoreApplication. Application may misbehave.
Oct 21 16:48:53 darkstar kernel: [ 68.296738] ata2.00: failed to read SCR 1 (Emask=0x40)
Oct 21 16:48:53 darkstar kernel: [ 68.296742] ata2.01: failed to read SCR 1 (Emask=0x40)
Oct 21 16:48:53 darkstar kernel: [ 68.296748] ata2.15: exception Emask 0x50 SAct 0x0 SErr 0x800 action 0x6 frozen
Oct 21 16:48:53 darkstar kernel: [ 68.296750] ata2.15: irq_stat 0x08000000, interface fatal error
Oct 21 16:48:53 darkstar kernel: [ 68.296752] ata2.15: SError: { HostInt }
Oct 21 16:48:53 darkstar kernel: [ 68.296755] ata2.00: exception Emask 0x100 SAct 0x1 SErr 0x0 action 0x6 frozen
Oct 21 16:48:53 darkstar kernel: [ 68.296758] ata2.00: failed command: READ FPDMA QUEUED
Oct 21 16:48:53 darkstar kernel: [ 68.296762] ata2.00: cmd 60/20:00:0f:34:54/00:00:57:00:00/40 tag 0 ncq 16384 in
Oct 21 16:48:53 darkstar kernel: [ 68.296763] res 50/00:00:00:00:00/00:00:00:00:00/00 Emask 0x100 (unknown error)
Oct 21 16:48:53 darkstar kernel: [ 68.296765] ata2.00: status: { DRDY }
etc

The line about "QDBusConnection: system D-Bus connection created before QCoreApplication. Application may misbehave" is followed by errors on my ata2 connection which is possibly the cause of the misbehavior of my ntfs partition.
If you Google on this then it turns out to be also an assigned bug in KDE.

Further Googling on a Gentoo forum revealed that this problem probably is related to dev-libs/libdbusmenu-qt as well as qt-dbus. This indeed looks to be a d-bus related issue. Someone seems to have solved it by rebuilding qt first and then rebuilding KDE.
Since Xfce uses also qt (at least for applications like Dolphin) it explains why it has the same bug.

I don't know if all the above statements are true and if so if they are the real cause of my problem, but would like to keep everyone updated

SlackInstalled

This is the output on my system
Quote:
bash-4.2$ ls -l /var/log/packages |grep ntfs
-rw-r--r-- 1 root root 3230 Jul 18 21:59 ntfs-3g-2012.1.15-i486-2
There is a discrepancy in the file sizes. Try reinstalling the ntfs-3g package.

I don't think the difference in reported seizes is the issue since as I reported earlier I have the same problem with the stock ntfs kernel driver

SlackInstalled

This may be a stupid question ...

Why dolphin? Why not just switch over to thunar (at least until the bug gets fixed) and call it a day?

Why dolphin? Why not just switch over to thunar (at least until the bug gets fixed) and call it a day?

Unluckily Thunar in Xfce gives the same problem
I don't know why, since it is (if I am right) GTK based and not QT
Maybe the underlying problem is a dbus problem and not qdbus (which uses dbus I suppose)

To make sure I tested again with Slackware 13.1 and that version is giving me no problems at all
Also I upgraded My Slackware 14.0 QT to version 4.8.3, but that did not solve the problem either.

ISSUE SOLVED

IT WAS NOT AN NTFS PROBLEM

The "drive" with the ntfs partiton I have is in fact a hardware raid 1 array.
(Dawicontrol DC-5200)
As such the operating system sees just the "drive" and can access it. No driver is needed.
(It is not a software raid or fake array)
However the raid controller is using the SATA port multiplier functionality.
The port multiplier functionality is used for an application program under Windows to be able to monitor the raid status. As such there are three sub ports on this
controller, Sub-port 0 is the normal raid access port. In addition to that is uses also sub ports 1 and 15.
When the kernel starts, it sees the multiplier functionality, but does not do anything with it. So when accessing the drive from the console it only uses port 0.
However when trying to access the "disk" from the windows manager it tries to open all three ports as disks (so also sub ports 1 and 15) and it fails on those.
I don't know why the windows manager tries to do this different from the console, but it is causing my problems.
I also don't know why this functionality of the windows manager has changed between releases 13.1 and 13.37.

To solve the problem I recompiled the kernel with the kernel configuration parameter CONFIG_SATA_PMP=n (instead of y) in order to disable the port multiplier functionality.
After that change I can now access my ntfs partition from the windows manager without any problems

So the problem for me is solved, however there seems to be a bug in the windows manager chain.

I did all that and I am in runlevel 4 in 3 even and I cannot see them in my file manager this is my fstab --

/dev/sda3 swap swap defaults 0 0
/dev/sda2 / ext4 defaults 1 1
/dev/sda1 /ntfs ntfs-3g fmask=133,dmask=022 1 0
#/dev/cdrom /mnt/cdrom auto noauto,owner,ro,comment=x-gvfs-show 0 0
/dev/fd0 /mnt/floppy auto noauto,owner 0 0
devpts /dev/pts devpts gid=5,mode=620 0 0
proc /proc proc defaults 0 0
tmpfs /dev/shm tmpfs defaults 0 0


I can cd to in in a terminal though and see the files and such but how does one cd into Documents and Settings ?