I have tried to grab details about the kernel panic stack trace to make it easier for Joerg to identify and fix the issue. I have downloaded the latest Kubuntu Live ISO image which is based on kernel 4.2.0. I have changed the boot parameters to boot with intel_iommu=on. I was able to capture the kernel panic with slow motion video recording. The hard lockup appears to occur when calling "flush_context".

Testing with:

Intel I7 980

MSI Big Bang X-Power (X58)

Here are the images of the kernel panic.

http://s31.postimg.org/5s0zzdu63/Screenshot_from_2016_04_23_00_04_32.png

http://s31.postimg.org/juris3oxn/Screenshot_from_2016_04_23_00_47_00.png

Exactly same kernel panic happens on Intel DX58SO board with a Xeon processor when using intel_iommu=on commandline parameter.

On Sat, Apr 23, 2016 at 12:35 AM, Dheeraj CVR <cvr.dheeraj@gmail.com> wrote:

I am currently running the following 2 setups.

Intel Xeon X5670
Intel DX58SO (x58)

Intel I7 980
MSI Big Bang X-Power (x58)

I am using IOMMU along with VFIO to passthrough a VGA card on both of my setups. Everything worked fine till kernel 4.1, but after the introduction of "iommu/vt-d: Make root entry visible for hardware right after allocation", both the boards no longer boot with "intel_iommu=on" commandline parameter.

The logs show the following:
DMAR: RMRR base: 0x000000bf7e8000 end: 0x000000bf7fffff
DMAR: ATSR flags: 0x0
DMAR: dmar0: Using Queued invalidation
------------[ cut here ]------------
WARNING: CPU: 4: PID: 1 at /home/dhiru1602/linux-git/kernel/watchdog.c:311 watchdog_overflow_callback+0x79/0xa0()
Watchdog detected hard LOCKUP on cpu 4
Modules Linked in:
CPU: 4 PID: 1 Comm: swapper/0 Tainted: G I 4.2.0-iommu-test

The kernel panic occurs very early in the boot process and I was not able to grab the full kernel panic log.

I have reverted the patch and recompiled the kernel. I was able to boot fine on both the boards with intel_iommu=on and I was able to see the IOMMU groups.

On a sidenote, Both the boards have broken interrupt remapping and the vendors have not provided BIOS updates to fix this problem. There was a report for the Intel board made on the Intel Community Forums here: https://communities.intel.com/thread/44752 but there hasn't been an updated BIOS released by Intel to fix this issue.

Please let me know if more information is needed to diagnose this issue.
--
Regards,
Dheeraj CVR.

Regards,
Dheeraj CVR.