[fglrx] binary graphics drivers don't load with Xen (2.6.22 and 2.6.24)

[Jeff Rasmussen]

The fglrx module won't load into the linux-image-2.6.22-14-xen kernel.

When you run 'sudo modprobe fglrx' you get "FATAL: Error running install command for fglrx" and the following log information.

/var/log/messages:
Oct 10 10:09:02 hfa9495l kernel: [ 76.843649] fglrx: Unknown symbol kmap_prot
Oct 10 10:09:02 hfa9495l kernel: [ 76.843753] fglrx: Unknown symbol xen_l1_entry_update
Oct 10 10:09:02 hfa9495l kernel: [ 76.843846] fglrx: Unknown symbol kmap_atomic_prot

I'm running an ATI X1300 card. Here is a partial lspci output.
01:00.0 VGA compatible controller: ATI Technologies Inc RV516 [Radeon X1300/X1550 Series]
01:00.1 Display controller: ATI Technologies Inc RV516 [Radeon X1300 Pro] (Secondary)

[Mehdi Abaakouk (sileht)]

I have the same problem with the nvidia_new module

sudo modprobe nvidia

dmesg :
Oct 11 13:04:56 prout kernel: [ 1945.433977] nvidia: Unknown symbol kmap_prot
Oct 11 13:04:56 prout kernel: [ 1945.434263] nvidia: Unknown symbol kmap_atomic_prot

lspci | grep VGA:
03:00.0 VGA compatible controller: nVidia Corporation NV43 [GeForce 6600 GT] (rev a2)

[Pulu Anau (panau)]

Can report the same problem on a core2duo system using the 64bit Gutsy.

Same exact messages as the first comment.

My uname:
Linux panau-linux 2.6.22-14-xen #1 SMP Sun Oct 14 23:20:20 GMT 2007 x86_64 GNU/Linux

With Nvidia drivers I always get kernel panics on x startup. If it doesn't work with ATI, I guess I'm stuck with kvm.

Please let me know if you need more information, I'm more than happy to provide.

[mdaguete (manuel-aguete)]

The same problem with nvidia:
Oct 26 10:40:10 xxxkernel: [ 25.017352] nvidia: Unknown symbol kmap_prot
Oct 26 10:40:10 xxx kernel: [ 25.017842] nvidia: Unknown symbol kmap_atomic_prot

[thearthur (arthur-ulfeldt)]

is this bug about the nvidia driver or the ati?

[Jeff Rasmussen (jeffrasmussen)]

fglrx is the ati driver

On 10/30/07, thearthur <email address hidden> wrote:
>
> is this bug about the nvidia driver or the ati?
>
> --
> [Gutsy] fglrx driver doesn't load with linux-image-2.6.22-14-xen
> https://bugs.launchpad.net/bugs/151327
> You received this bug notification because you are a direct subscriber
> of the bug.
>

--
Jeff Rasmussen
GPG public key 0x9686C12F

[mdaguete (manuel-aguete)]

both.

[karlo (carlo-reggiani)]

On my AMD X2 64 (4200+) with a ATI Saphire X1300 Gnome is ok, but all the graphical operation (opening windows, scrolling text, etc) are very slow.

I'm getting:
sudo modprobe fglrx --> FATAL: Error running install command for fglrx

karlo

[karlo (carlo-reggiani)]

More info from dmesg:

[ 28.303858] fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel.
[ 28.304231] fglrx: Unknown symbol xen_l1_entry_update
[ 28.304511] fglrx: Unknown symbol flush_tlb_page

[ 477.364207] fglrx: Unknown symbol xen_l1_entry_update
[ 477.364492] fglrx: Unknown symbol flush_tlb_page
[ 815.122407] fglrx: Unknown symbol xen_l1_entry_update
[ 815.122701] fglrx: Unknown symbol flush_tlb_page

I did a distribution upgrade from 7.04.

karlo

[Johnathon (kirrus)]

Bug #157298 is duplicate of this report.

[Marc Quinton (mquinton)]

same problem with GeForce 7300 SE card.

[Marc Quinton (mquinton)]

to avoid this problem, don't use restricted card driver : go to System / Admin / Restricted drivers and suppress non open-source driver. In other word, avoid gglrx and nvidia in /etc/X11/xorg.conf ; prefer nv and ati.

[Matt Wagner (mwaggy)]

Unfortunately, some cards (such as the Mobility X1300) don't work under the older ati or radeon drivers, but instead fall back to vesa, so it's basically fglrx or nothing.

[Johnathon (kirrus)]

Hello,
Can someone who has this issue please run the following commands, and attach them to the bug report?
    * uname -a > uname-a.log
    * cat /proc/version_signature > version.log
    * dmesg > dmesg.log
    * sudo lspci -vvnn > lspci-vvnn.log
Thanks

[Martin (martinjedamzik)]

Hi Jonathon,

here you go.

Cheers, Martin

[Johnathon (kirrus)]

Confirming (multiple reporters), assigning to Kernel Team (lots of kernel messages in the comments). This is the end of what I can do for you, just need to wait for the kernel team.

[Johnathon (kirrus)]

Thanks for the report, and thanks for the files Martin.

[Omar Benhamid (omar-benhamid)]

Hi all, I found a problem with xenified kernel source in arch/i386/mm/hypervisor.c and arch/x86_64/kernel/smp-xen.c : xen_l1_entry_update and flush_tlb_page are not exported to kernel symbol table. I exported the symbols and the module loads (well causing some trouble on my machine ...).
Sorry i'm not very familiar with debian bulid system to give the right things but here is the diff I applied to debian/build/custom-source-xen in linux kernel source :

diff -urN custom-source-xen/arch/i386/mm/hypervisor.c custom-source-xen.modified/arch/i386/mm/hypervisor.c
--- custom-source-xen/arch/i386/mm/hypervisor.c 2008-02-08 17:27:50.000000000 +0000
+++ custom-source-xen.modified/arch/i386/mm/hypervisor.c 2008-02-08 13:18:35.000000000 +0000
@@ -51,6 +51,7 @@
        u.val = __pte_val(val);
        BUG_ON(HYPERVISOR_mmu_update(&u, 1, NULL, DOMID_SELF) < 0);
 }
+EXPORT_SYMBOL(xen_l1_entry_update);

 void xen_l2_entry_update(pmd_t *ptr, pmd_t val)
 {
 diff -urN custom-source-xen/arch/x86_64/kernel/smp-xen.c custom-source-xen.modified/arch/x86_64/kernel/smp-xen.c
--- custom-source-xen/arch/x86_64/kernel/smp-xen.c 2008-02-08 17:27:50.000000000 +0000
+++ custom-source-xen.modified/arch/x86_64/kernel/smp-xen.c 2008-02-08 16:43:43.000000000 +0000
@@ -293,10 +293,16 @@
 { return; }
 void flush_tlb_current_task(void)
 { xen_tlb_flush_mask(&current->mm->cpu_vm_mask); }
+EXPORT_SYMBOL(flush_tlb_current_task);
+
 void flush_tlb_mm (struct mm_struct * mm)
 { xen_tlb_flush_mask(&mm->cpu_vm_mask); }
+EXPORT_SYMBOL(flush_tlb_mm);
+
 void flush_tlb_page(struct vm_area_struct * vma, unsigned long va)
 { xen_invlpg_mask(&vma->vm_mm->cpu_vm_mask, va); }
+EXPORT_SYMBOL(flush_tlb_page);
+
 void flush_tlb_all(void)
 { xen_tlb_flush_all(); }
 #endif /* Xen */

Regards

[Timo Aaltonen (tjaalton)]

Hardy now has fglrx enabled for xen, please test.

[netslayer (netslayer007)]

Hardy Alpha 5 (fresh install cd + online update) + xen + fglrx fails with the same errors above. sources 2.6.24 xen
I re installed with Gutsy (fresh install cd) + xen + fglrx and I get the same..

** shrugs..

[Leann Ogasawara (leannogasawara)]

Hi netslayer,

Can you attach the entire dmesg output that also contains the errors when running the latest Hardy Beta kernel? Thanks.

[drivinghome (drivinghome)]

I'm running Hardy up to date, i've installed ''linux-restricted-modules-2.6.24-16-xen'', it seems that there is no fglrx module:

$sudo modprobe fglrx
FATAL: Could not open '/lib/modules/2.6.24-16-xen/volatile/fglrx.ko': No such file or directory

[John Dong (jdong)]

[ 101.170910] fglrx: Unknown symbol xen_invlpg_mask
[ 101.171091] fglrx: Unknown symbol kmap_prot
[ 101.171866] fglrx: Unknown symbol kmap_atomic_prot

Same problem on up to date Hardy

[Saúl Romero (spacetree)]

I have the same problem on Hardy AMD64 up to date. I suppose I have 'linux-restricted-modules-2.6.24-16-xen' installed since I can use my wireless drivers. Are you still accepting the logfile?

[zavarian (zavarian)]

linux-restricted-modules-2.6.24-19-xen still giving problems

I'm getting fglrx: Unknown symbol xen_invlpg_mask When loading the module. Just wondering if there is a progress update for this bug.

[Alberto Milone (albertomilone)]

Currently the FGLRX driver doesn't support Xen and there is nothing we can do without ATI's support.

[selric (rickard-sellstedt)]

Any news regarding this issue?

[Launchpad Janitor (janitor)]

Per a decision made by the Ubuntu Kernel Team, bugs will longer be assigned to the ubuntu-kernel-team in Launchpad as part of the bug triage process. The ubuntu-kernel-team is being unassigned from this bug report. Refer to https://wiki.ubuntu.com/KernelTeamBugPolicies for more information. Thanks.