[Intel-gfx] Approximately every other day video will freeze for no reason at all and no specific time between freezes.

[Martin Steigerwald]

Hi Chris, hi Intel graphics developers,

Am Freitag, 2. Oktober 2015, 16:06:30 CEST schrieb Chris:
> To elaborate on this. The video is frozen to include the clock however I
> can move the mouse cursor around the screen. I can not however switch
> between window desktops. This happened on 28 Sept 2015 after an uptime

I have seen exactly this issue on ThinkPad T520 with Sandybridge graphics.

With one exception: At least sometimes I have been able to switch to tty1 and 
do killall -u <affecteduser> to get back to sddm login screen.

I didn´t do anything about it so far. Driver is running in uxa mode, cause 
with sna I experienced hard freezes with mouse freezing and even data loss. I 
lost my GPG public key ring after a hard freeze after a gpg --recv-keys 
operation (restored from backup meanwhile). This was at DebConf and a 
developer I worked with recommended me to switch from SNA to UXA. And some 
time ago I had it that kwin sometimes told me it restarted compositing due to 
a hang in the graphics driver.

I find freezing issues extremely hard to debug. First off often they happen 
when I am not willing to spend even a single minute on them. Second: What to 
do anyway? On a hard freeze, I could try Sys-Rq key combo to sync disk 
contents, and I think I may better enable it anyway to avoid data loss in 
those situations, and hope that *some* information on *why* the hard freeze 
happened still lands in kern.log or Xorg log. On soft freezes I can do 
something if I can still switch to tty1. But what? I have no idea what 
information I can gather *efficiently* to provide at least a somewhat 
meaningful initial bug report within less than 15 to 30 minutes. I don´t just 
want to do *something* in that case, but I want something that has a good 
chance on giving a real insight on why the bug is happened and a good hint at 
what can be done to fix it.

git-bisect is out of question for me for something that happens so 
irregularily. It would take an insane amount of time with running kernels that 
are probably way to unstable for production use – as it may include pre-rc1 
kernels - anyway. Even tough I go to the limits on this machine regarding 
production use, this is still a production machine and I usually wait till rc2 
kernels to try out a new kernel.

So I appreciate any pointers what to *efficiently* do about issues like this. 
In case I miss any documentated bug triaging howto please point me to it. 
Ideally some kind of bug-reporting tool would be nice. Just run a script which 
collects all data that can be helpful into a file that I can attach to a bug 
report.


That said in general I am really quite happy with the Intel drivers.


martin at merkaba:~> phoronix-test-suite system-info

Phoronix Test Suite v5.2.1
System Information

Hardware:
Processor: Intel Core i5-2520M @ 3.20GHz (4 Cores), Motherboard: LENOVO 
42433WG, Chipset: Intel 2nd Generation Core Family DRAM, Memory: 16384MB, 
Disk: 300GB INTEL SSDSA2CW30 + 480GB Crucial_CT480M50, Graphics: Intel HD 3000 
(1300MHz), Audio: Conexant CX20590, Monitor: P24T-7 LED, Network: Intel 
82579LM Gigabit Connection + Intel Centrino Advanced-N 6205

Software:
OS: Debian unstable, Kernel: 4.3.0-rc3-tp520-btrfstrim+ (x86_64), Desktop: KDE 
Frameworks 5, Display Server: X Server 1.17.2, Display Driver: intel 2.99.917, 
OpenGL: 3.3 Mesa 11.0.2, Compiler: GCC 5.2.1 20150911, File-System: btrfs, 
Screen Resolution: 3840x1080

lspci -nnv:

00:02.0 VGA compatible controller [0300]: Intel Corporation 2nd Generation 
Core Processor Family Integrated Graphics Controller [8086:0126] (rev 09) 
(prog-if 00 [VGA controller])
        Subsystem: Lenovo Device [17aa:21cf]
        Flags: bus master, fast devsel, latency 0, IRQ 26
        Memory at f0000000 (64-bit, non-prefetchable) [size=4M]
        Memory at e0000000 (64-bit, prefetchable) [size=256M]
        I/O ports at 5000 [size=64]
        Expansion ROM at <unassigned> [disabled]
        Capabilities: [90] MSI: Enable+ Count=1/1 Maskable- 64bit-
        Capabilities: [d0] Power Management version 2
        Capabilities: [a4] PCI Advanced Features
        Kernel driver in use: i915

Ciao,
-- 
Martin