Arch Linux

AlmostSurelyRob

Member

From: UK

Registered: 2012-03-17

Posts: 77

[SOLVED] After kernel update X freezes the system

Sorry, I didn't know how to make the subject more descriptive.

I have been running Arch with Xmonad as WM for a good couple of years now. I updated my installation yesterday and run

xmonad -compile

which completed without errors. After restart X became extremely unstable freezing the whole system after a couple of minutes. I logged in remotely and run

dmesg -wH

which allowed me to see the following message:

[  +0.000002] DMAR: DRHD: handling fault status reg 2
[  +0.000002] DMAR: [DMA Read] Request device [00:02.0] fault addr 1239000 [fault reason 05] PTE Write access is not set
[  +0.000002] DMAR: DRHD: handling fault status reg 3
[  +0.000001] DMAR: [DMA Read] Request device [00:02.0] fault addr 1237000 [fault reason 05] PTE Write access is not set
[  +0.000003] DMAR: DRHD: handling fault status reg 2
[  +0.000002] DMAR: [DMA Read] Request device [00:02.0] fault addr 1225000 [fault reason 05] PTE Write access is not set
[  +0.000003] DMAR: DRHD: handling fault status reg 3
[  +0.000001] DMAR: [DMA Read] Request device [00:02.0] fault addr 1223000 [fault reason 05] PTE Write access is not set
[  +0.000003] DMAR: DRHD: handling fault status reg 2
[  +0.000002] DMAR: [DMA Read] Request device [00:02.0] fault addr 1220000 [fault reason 05] PTE Write access is not set
[  +0.000001] DMAR: DRHD: handling fault status reg 2
[  +0.000002] DMAR: [DMA Read] Request device [00:02.0] fault addr 123f000 [fault reason 05] PTE Write access is not set
[  +0.000002] DMAR: DRHD: handling fault status reg 2
[  +0.000002] DMAR: [DMA Read] Request device [00:02.0] fault addr 1241000 [fault reason 05] PTE Write access is not set
[  +0.052261] [drm:fw_domains_get [i915]] *ERROR* render: timed out waiting for forcewake ack to clear.

Looking through the forum, I found the following issue https://bbs.archlinux.org/viewtopic.php?id=176398. So I edited my grub kernel parameters line and added "intel_iommu=igfx_off" and the problem appears to have gone away.

I thought I'll share this in case anyone has similar issues.