Please attach your xorg log and dmesg output.

xorg.log -> http://pastebin.com/EMpPunjc
dmesg -> http://pastebin.com/XtdNe65A

Is this error related to my problem: 
[Firmware Bug]: Duplicate ACPI video bus devices for the same VGA controller, please try module parameter "video.allow_duplicates=1"if the current driver doesn't work.

What version of cairo are you using?  This may be a duplicate of bug 47266.

(In reply to comment #3)
> What version of cairo are you using?  This may be a duplicate of bug 47266.

I don't know how to check version of cairo. I use LMDE update pack 4. In the synaptic I can see I have these pachages
libcairo2 -> 1.10.2-7
gir1.2-freedesktop -> 1.31.10-1
libcairo-gobject2 -> 1.10.2-7
libcairo-perl -> 1.090-1
libcairomm-1.0-1 -> 1.10.0-1
libmono-cairo2.0-cil -> 2.10.8.1-1
libmono-cairo4.0-cil -> 2.10.8.1-1
python-cairo -> 1.8.8-1+b2
python-gi-cairo -> 3.1.0-2
gtk2-engines-murrine -> 0.98.1.1-3
libpango-perl -> 1.222-1+b1
libpixman-1-0 -> 0.24.4-1

Do you have a test program for bug 47266? I mean a program that can simulate that bug. Then we would be able to check if this is the duplicate of the bug 47266.

updating to libcairo2 1.12.0-2 did not solve the problem, i.e. gnome freezes  when I scroll in the firefox. 

I'm not sure whether it helps you identify the bug or not, but let me tell you something: 

In the 32 bit version of the kernel, when I build the radeon driver in the kernel binary (I mean when radeon driver is not a module), kernel crashes and do not boot.

Alex Deucher I should thank you for helping us fixing this problem.

I noticed that you have disabled MSI for RV515 (my card) in 3.2.16 version of the kernel. I'm going to test that now. Please inform me if you have any more ideas or if you need more info.

I don't think it's bug 47266, it can't really cause freezes directly.

What are the symptoms of the freeze? Can you still log into the machine remotely? If yes, can you attach gdb to gnome-shell or the X server and get a backtrace? ...

(In reply to comment #2)
> xorg.log -> http://pastebin.com/EMpPunjc
> dmesg -> http://pastebin.com/XtdNe65A

Please attach files here directly.

(In reply to comment #5)
> In the 32 bit version of the kernel, when I build the radeon driver in the
> kernel binary (I mean when radeon driver is not a module), kernel crashes and
> do not boot.

Are you building all the required firmware files into the kernel as well? Otherwise the attempt to load them will time out.

> What are the symptoms of the freeze? 

When I scroll in the firefox or some other programs screen freezes. I mean I can't click on anything, but mouse cursor is visible and moving. There is no tearing or artifacts on the screen but it remains frozen. One day I was compiling the kernel when this freeze happened but my compilation job finished to the end.

I really encourage you to see this link for more info about the symptoms other people experienced. I've read all of it and I have experienced most them as well.
https://bbs.archlinux.org/viewtopic.php?id=129481&p=1

> Can you still log into the machine remotely? If yes, can you attach gdb to 
> gnome-shell or the X server and get a backtrace? ...

I can switch to linux shell (I mean alt + ctrl + F1) works. However, I'm not familiar with gdb. If you give me some basic instructions I would provide you with the required debugging info.

> xorg.log -> http://pastebin.com/EMpPunjc
> dmesg -> http://pastebin.com/XtdNe65A
> 
> Please attach files here directly.

Ok. I would do that.
 
> (In reply to comment #5)
> > In the 32 bit version of the kernel, when I build the radeon driver in the
> > kernel binary (I mean when radeon driver is not a module), kernel crashes and
> > do not boot.
> 
> Are you building all the required firmware files into the kernel as well?
> Otherwise the attempt to load them will time out.

yes I am. My config was like this.
CONFIG_FW_LOADER=y
CONFIG_FIRMWARE_IN_KERNEL=y
CONFIG_EXTRA_FIRMWARE="/radeon/R520_cp.bin"
CONFIG_EXTRA_FIRMWARE_DIR="/lib/firmware/"

I should add that similar config works fine on the 64 bit version of the kernel.

Created attachment 60476 [details]
xorg.log

Created attachment 60477 [details]
dmesg

Created attachment 60478 [details]
kernel config

This the kernel config which fails to boot in the 32 bit version of the LMDE.

(In reply to comment #5)
> I noticed that you have disabled MSI for RV515 (my card) in 3.2.16 version of
> the kernel. I'm going to test that now.

How did that go? I think an IRQ problem could explain your symptoms.

(In reply to comment #11)
> (In reply to comment #5)
> > I noticed that you have disabled MSI for RV515 (my card) in 3.2.16 version of
> > the kernel. I'm going to test that now.
> 
> How did that go? I think an IRQ problem could explain your symptoms.

It seems everything works fine.

Thank you Michel Dänzer and Alex Deucher.

My last question: Do you think I should file a new bug for the kernel crash that happens after building the radeon driver in the 32 bit version of the kernel?

today I switched to my 64 bit version of LMDE to test check 3.2.16 there. There is no freeze in 64 bit system either. However, I should say that my previous report that I building the radeon driver in the kernel in the 64 bit system works was wron. Kernel crashes at boot when I build radeon as module in both 32 and 64 bit systems.

(In reply to comment #13)
> today I switched to my 64 bit version of LMDE to test check 3.2.16 there. There
> is no freeze in 64 bit system either.

Great, resolving this report as fixed.

(In reply to comment #12)
> My last question: Do you think I should file a new bug for the kernel crash
> that happens after building the radeon driver in the 32 bit version of the
> kernel?

If it's not a firmware loading problem (i.e. it doesn't continue booting even after a couple of minutes), please do.