Debugging

Revision 16 as of 2007-10-19 09:40:26

Clear message

X and Ubuntu

The X Windows System is a critical component in the Ubuntu operating system. X is not without its bugs, but fortunately debugging X issues is not rocket science.

The vast majority of Ubuntu X issues fall into one of several distinct categories, and based on the way they manifest, there are several different tactics that can be employed in a nearly paint-by-numbers fashion to isolate them.

Even non-developers can help! The goal of this handbook is to give folks a toolset for rendering these bugs easily solveable. By making Ubuntu's X strong, we can help drive Open Source to world domination!

Bug Reporting

The lifecycle of a bug report begins, unsurprisingly, with the preliminary report. How a bug is initially reported can have a huge effect on how it's handled and how quickly it gets fixed.

Choosing a Good Title

Your title should communicate two things: The symptom you're seeing, and whatever is unique or unusual about your system. Otherwise, your bug may not get proper attention.

Examples:

BAD:

Crazy screen issues on boot

BAD:

Multiple problems with CD today

BAD:

Randomly doesn't work

GOOD:

[Feisty] Screen briefly corrupts during boot with -nv (NVidia 6100)

GOOD:

[Hardy Alpha-3] Alt-CD (only) selected wrong driver (Matrox / BenQ FP91+)

GOOD:

[Gutsy] Periodic crashes w/ high CPU on Dell Latitude D505 (-intel 855GM)

GOOD:

[Dapper,Edgy] Wrong default refresh rates on 16:10 LCD panels

Do's and Don't's

DON'T:

Assume "they must already know about this"

DO:

Look for existing bug reports that match your problem

DON'T:

Assume a "similar" bug is exactly what you're seeing

DO:

File a new bug, but mention the ID's of all bugs that sound similar. Someone can dupe them together later.

DON'T:

Add "me too" responses. Wastes everyone's time.

DO:

Add missing data (photos, logs) to add to an existing bug's "knowledge base". Or if you just wish to be notified, then Subscribe yourself to the bug.

DON'T:

Post bugs with only a brief description of the problem

DO:

Post relevant logs, config files, and data (see table below) ALWAYS ATTACH YOUR /var/log/Xorg.0.log

DON'T:

Assume "everyone" is seeing this same bug

DO:

Consider what is unique about your system

DON'T:

Assume others will "just know" how the bug occurs

DO:

Itemize the exact steps that result in the issue. Can you reproduce it at will?

DON'T:

Fire and forget. Abandoned bugs rarely get fixed.

DO:

Follow up on your bug from time to time, even if it seems ignored. Report if the issue goes away or remains when new Ubuntu's come out.

What to Include in Bug Reports

Problem class:

Things to Include:

General X bug

Description of problem

Paste in output of lspci -nn | grep VGA

Attach /etc/X11/xorg.conf

Attach /var/log/Xorg.0.log

Attach output of lspci -vvnn

Wrong resolutions, refresh rates, or monitor specs

Resolution, rate, or other parameter expected

Resolutions, rates, or other parameters actually obtained

/etc/X11/xorg.conf

/var/log/Xorg.0.log

output of lspci -vvnn

output of sudo ddcprobe

output of xrandr

Wrong font dpi or size

Are you running GNOME, KDE, XFCE, or ...?

Affected (and unaffected) applications

/var/log/Xorg.0.log

output of sudo ddcprobe

Screenshot showing font differences

X crash, lockup, freeze, exit, or doesn't start/shutdown

Detailed description of problem

List any versions you tried that did not have this issue

Detailed list of steps to reproduce

How complete is the X failure? BR + Does ctrl+alt+f1 take you to a console? BR + Does ctrl+alt+backspace restart X? BR + Does mouse pointer still move? BR + Does the keyboard LED come on when hitting the CAPSLOCK key?

/etc/X11/xorg.conf

/var/log/Xorg.0.log

/var/log/Xorg.0.log.old

~/.xsession-errors

output of lspci -vvnn

Keyboard, touchpad, and mouse issues

Description of the problem

/var/log/Xorg.0.log

output of xprop -root

output of gconftool-2 -R /desktop/gnome/peripherals

Screen display corruption

Photo of the screen

Description of the problem

Does it also occur if DRI is disabled?

/var/log/Xorg.0.log

Bad video playback

/etc/X11/xorg.conf

/var/log/Xorg.0.log

output of lspci -vvnn

Bug Triage

Ubuntu receives a huge amount of bug reports, many of which are important and valid issues needing attention. Even so, nearly all X bugs are initially reported without information necessary for classification and analysis. This is where the bug triaging role comes in.

Bug triaging for Ubuntu's Xorg components does not require any particular expertise with X, just regular Linux know-how should be sufficient. As a bug triager, your role is twofold: First as a coach to help bug reporters in maximizing their chances of getting the bug addressed by providing complete information, and second as a filter to help developers focus their time on important and/or easy-to-fix bugs.

After initially reported, a bug is reviewed and several basic things are checked by the bug triager:

1. Is it definitely an X bug? Sometimes things get misfiled, and sometimes reports are just invalid, or are really just support requests and should be directed to Launchpad Answers instead. If unsure, leave it as is.

2. Is it clearly a dupe of an already known bug? Ideally reporters should do a cursory scan of existing bug reports to see if it's obviously already in the system, but not all reporters do. If unsure, don't dupe it - someone can handle this later.

3. Is there at least the basic minimum amount of data present? If not, mark it Incomplete and see below for a table of what kinds of files and command output is needed. Once you think the basic required info is present, move it to the Confirmed state.

4. Review log files for error messages or other obvious anomalies. Highlight these in the bug report, and search launchpad for other reports of that same error message. Mention these as potential dupes, or dupe them where obvious.

5. Tidy up the bug report. This may involve improving the bug's title or wordsmithing the description to clarify it.

Since step 3 requires waiting on replies from bug reporters, you can't go through all five of the above steps in one bug triaging session. Instead, when doing bulk triage work, you can consider dividing the workflow into two types of sessions:

INITIAL TRIAGE: (steps 1->2->3)

  • Do a query for NEW bugs.
  • For each bug, review according to steps 1, 2, and 3 above
  • Post a request for more information and set to INCOMPLETE

FINAL TRIAGE: (step 3->4->5)

  • Do a query for INCOMPLETE-WITH-RESPONSE bugs.
  • For each bug, read the reporter's reply and information posted
  • If still insufficient info, ask for more info and leave INCOMPLETE
  • Otherwise, do steps 4 and 5, and mark CONFIRMED

Once this basic triage work is in place, a reviewer (generally a developer or official bug master) reviews CONFIRMED bugs and just doublechecks that all the necessary stuff has been done. They then set the bug to TRIAGED state.

Bug Research

For many bugs, a little googling and searching in upstream bug trackers can reveal important additional info.

1. Review all attached log files for error messages.

2. Look for other similar/duplicate bug reports to gain additional perspectives and look for obvious commonalities, like same error messages, driver, hardware, etc. Places to search:

  • google.com
  • bugs.launchpad.net/ubuntu
  • bugs.debian.org/
  • bugs.freedesktop.org/
  • ubuntuforums.org
  • If you find the same issue reported in Launchpad, mark the less complete and/or newer bug as a dupe of the other. If you find the same issue reported in debian or xorg, mark it as "Also affects project..." With Google, it often helps to include "ubuntu" in the search string. Also, you can use "site:freedesktop.org" or "site:debian.org" to narrow the search to a specific domain.

3. Try reproducing the issue, especially if you have similar hardware.

4. Look for a newer version of the package, and review its changes to see if there's a fix for this issue.

  • If so, check apt-get update; apt-cache madison $pkgname to see if the new version is already packaged. If not, ask a packager to produce a test package of the new release to test for this bug.

5. Upload any patches you run across directly to Launchpad, and be sure to tick the "patch" checkbox, so patches can be queried for later.

6. Have them try an older Ubuntu Live CD, or have them downgrade a specific package. For example, to downgrade the xserver:

         apt-get install xserver-xorg-core=2:1.3.0.0.dfsg-4ubuntu
  • If an older version fixes the issue, then possibly you can bisect things down to find a specific patch causing the issue. See the Analysis section for how to do this.

7. Unless you've been lucky and found the fix already, finish up the research phase by doing the following:

  • Summarize your findings. Restate the problem, describe progress made, outline remaining suspicions or questions.
  • If appropriate, report the bug upstream to Debian and/or Xorg, attaching all relevant files and a link to the Ubuntu bug report. Summarize the research you did, patches that were tested, and any other details that may be relevant.

Analyzing X Problems

For hard bugs the analysis phase is the most important, and most challenging part of bug work. It often requires both strength of insight and skill with code.

Depending on how the bug is behaving, there are multiple directions to investigate the issue. Here's some different approaches:

Problem manifested only recently

If the issue has been narrowed to occur only after (or before) a given point in time or software version, then it is possible to narrow in on the specific cause of the issue through a "Bi-Section" strategy.

Essentially, if you know it occurred in Version 1, but not Version 8, have a person able to replicate the issue try Version 4. If it's there, then have them try Version 6, otherwise Version 2.

If the problem is in the current Ubuntu, but not in the prior Ubuntu, it can be useful to have them test the intermediate Alpha versions of the new release.

Once you have bracketed it down to a specific version of something, you can then go through the individual patches included in that version compared with the prior one. Sometimes the patch descriptions can give a strong clue to this. If there are a number of changes, then rather than trying each patch one-by-one you may want to simply disable the latter half of patches, and bisect that way.

If you've narrowed it to an upstream version change, then you may wish to use git's bisecting functionality to assist with this.

Problem manifests only with specific configuration options

TODO

Problem manifests only with a particular driver

If the research found that most people with this problem were all using the same driver, then obviously it makes sense to explore it from that aspect.

Note that for most graphics hardware, there are at least two different drivers. It can be worthwhile to test the alternate driver to verify it's a driver issue.

  • NVidia: -nv (open) and -nvidia (proprietary)
  • ATI: -ati (open) and -fglrx (proprietary)
  • Intel: -intel (current) and -i810 (legacy).

Each driver has its own source code package, which can be retrieved via xserver-xorg-video-<driver>. The open source drivers also have git repositories at http://gitweb.freedesktop.org.

Resolving these issues will generally require patching the driver code, although some driver-specific issues end up requiring changes to other pieces of code, like the xserver.

The following can be added to your /etc/X11/xorg.conf to provide additional debug information:

Section "Device"
        ...
        Option "ModeDebug" "true"
        ...
EndSection

Problem manifests only with particular kind of hardware

Many issues are highly specific to a particular kind of hardware, such as only Intel 855, or only a particular monitor model. Sometimes these end up being general bugs, but often they require adding hardware-specific quirks to the driver or to xserver.

TODO: Process for adding these changes

Problem manifests under seemingly random conditions

Few bugs are truly random; usually this just means more data is needed. Things to consider:

  • Resource utilization over time
  • Specific to one piece of hardware? If so, is that HW faulty?
  • How is the system being used when it fault occurs? If it's idle,
    • could it be a screensaver, power savings, or something?
  • Fluctuating Network/power conditions?

Problem manifests itself during video playback

TODO

Problem manifests itself when using 3D software (compiz, games, GL...)

TODO

Problem manifests as a performance degradation issue

TODO

Problem results in screen display corruption

Nearly all screen corruption issues will be due to a bug in a driver. Identify the driver and the specific steps to produce the corruption. Then run the xserver through gdb to identify the line or lines immediately prior to the corruption.

From here, things to try could include checking for invalid/undefined values, adding usleep() calls to add delay, or even disabling the lines in question.

Once a preliminary patch exists, post it to the upstream xorg list for feedback. Often they can suggest a better patch.

Problem results in X crash, lockup, freeze, or exit

In some cases, an error message will be printed before the fault; these can be used to identify where in the codebase the fault occurred, and often give an explanation as to why.

Otherwise, use gdb to get a backtrace. Once the issue is found, step through the code leading up to the line where the fault occurred. Look for invalid/undefined values, or questionable logic. Try disabling the line or lines where the fault occurred, adding usleep() before it, or etc.

Once a preliminary patch exists, post it to the upstream xorg list for feedback. Often they can suggest a better patch.

Problem involves wrong resolutions, refresh rates, or monitor specs

These issues can be narrowed down by checking a few things:

  • Does xorg.conf have the correct values? If so, then something is
    • wrong in how xserver is interpreting them. Review Xorg.0.log.
  • Is the hardware new? If so, is it's pciid registered properly?
  • xresprobe - Is ddcprobe outputting the right parameters? Is
    • xresprobe selecting the correct one from this set?

Problem involves wrong font dpi or size

TODO

Problem involves buggy EDID from monitor

If the monitor is clearly advertising an incorrect mode (such as not advertising a preferred mode), a quirk can be added to the xserver to prefer a specific mode.

TODO: What's the code change to do this?

Problem involves startup/shutdown, hibernate, suspend, or tty switch

TODO

Problem involves missing support for some keyboard keys

TODO

Problem involves missing support for mouse or touchpad functions

TODO

Problem involves GUI application that crashes with an X error message

TODO

Problem may be due to prior installs of a binary driver

Try:

 dpkg -l '*fglrx*'

and

 locate fglrx

to see if there is still some proprietary bits around causing problems.

Debugging Memory Issues

top xrestop

Limiting the ram X uses to 80% to prevent memory leaks via ulimit -m in the X startup script

Troubleshooting Common Error Messages

Common Intel Driver Error Messages

   (II) intel(0): [drm] removed 1 reserved context for kernel
   (II) intel(0): [drm] unmapping 8192 bytes of SAREA 0xf89c1000 at 0xb7b65000

These appear only on system shutdown, and generally don't indicate an issue.

   (EE) intel(0): I830 Vblank Pipe Setup Failed 0

This is because the X driver calls the DRM_I915_SET_VBLANK_PIPE ioctl after de-initializing the DRM. It should be harmless.

   (II) AIGLX: Suspending AIGLX clients for VT switch
   Error in I830WaitLpRing(), timeout for 2 seconds

This is a generic error indicating that the GPU locked up. It could be caused by a variety of issues.