Created attachment 336322 [details]
screen-noise-investigation_file.tar.bz2

This is *Regression* issue related to Xorg mga driver.
I belive combination of G200e and mga driver on xorg is very problematic.

**** General Escalation Information ****

State the problem

 1. Provide time and date of the problem
N/A

 2. Indicate the platform(s) (architectures) the problem is being
    reported against.
RHEL4.8 *Beta*, x86_64, xorg-x11-6.8.2-1.EL.62

 3. Provide clear and concise problem description as it is
    understood at the time of escalation
        * Observed behavior
Bottom part of the screen is garbled

        * Desired behavior
The screen is displayed without the noise.

 4. State specific action requested of SEG
----
I belive there is several combination of issues/bugs on X with G200e @ 24bpp.
IMO, this is problem how to detect/treat G200e series (Chipset:
"mgag200 SE A PCI") on Xorg.

For RHEL5.4/4.8 and RFEs (Backgrounds, I believe)

Basically, Xorg mga driver had no capability for 24bpp on X with MGA
G200e, if video memory would be recognized properly, because of hard
coding 16bpp on X (please see BZ#479919 for RHEL5.3).

So there's already RFE request for RHEL4.8 (by Fujitsu-Siemens) and
RHEL5.4 (by NEC).  Besides, it seemed there's much complicated that
MGA200e has 2 versions (Kronos-1 has 1.7MB Video Ram, Kronos-2 has
8MB) and MGA200 has 2 version (Pilot 1 and Pilot 2). About this,
please see BZ#469651.

For this issue

NEC have Great concern about that many server already shipped use
MGA200e (Pilot1, Kronos-1, pci-id 102b:0522) affected by this
problem and they think this as very important (prioritized) one.

In this time, they reported display corruption problem on RHEL4.8
Beta on this ticket, and memory detection problem on other ticket
(IT#278328).
I believe that both of issues should be fixed on
RHEL4.8, (and/or on z-stream?) through BZ#469651, on which
Fujitsu-Siemens also repoted similar problem.
----

 5. State whether or not a defect in the product is suspected
        * Provide Bugzilla if one already exists
RHEL4 for BZ#469651, RHEL5 for BZ#479919 (maybe related).

 6. If there is a proposed patch, make sure it is in unified diff
    format (diff -pruN)

 7. Refrain from using the word "hang", as it can mean different
    things to different people in different contexts. Use a better
    and more specific description of your problem.
----
Bottom part of display is garbled in X-windows on servers using MGA G200e.

This problem occurred on Pilot small VideoRAM if 24bpp is specified in RHEL4.8 Beta x86_64.
This problem did not occur on Pilot large VideoRAM (Pilot 2nd gen) even if 24bpp is specified in RHEL4.8 Beta x86_64.

  VideoRAM size
    Pilot small VideoRAM : 1696 kByte
    Pilot 2nd gen        : 8128 kByte

- This problem does not occur on RHEL4.8 Beta X86_64(select 16bpp).
- This problem does not occur on RHEL4.8 Beta i386.
  + RHEL4.8 Beta i386 cannot select 800x600 24bpp with the GUI tool.

[detail]
+----------------------+--------------+---------+
| config               | RHEL4.7      | RHEL4.8 |
| arch(Modes Depth)    | GA           | Beta    |
+----------------------+--------------+---------+
|x86_64(800x600 24bpp) | OK           | NG      |
|x86_64(800x600 16bpp) | OK           | OK      |
|i386(800x600 24bpp)   | Can't select | NG(*1)  |
|i386(800x600 16bpp)   | OK           | OK      |
+----------------------+--------------+---------+
*1 Adding 800x600 to the line of Modes of xorg.conf
----

 8. This is especially important for severity one and two
    issues. What is the impact to the customer when they experience
    this problem?
        * For example: "This system houses our accounts payable
          database. When the system crashes we are unable to process
          payroll, and other payable functions. This is especially
          critical as we approach end of our quarter."
----
This is a regression issue which affects servers already shipped.
----

**** Provide supporting info ****

 1. State other actions already taken in working the problem:
        * tech-list, google searches, fulltext, consulting with
          another engineer
        * Provide any relevant data found

 2. Attach sosreport
Attached screen-noise-investigation_file.tar.bz2 including sosreport-Suruga3B-248113-e3e7e9.tar.bz2.

 3. Attach other supporting data

Also rhel4.8beta_x86_64_screen.png (screen shot which show garbled screen) is included in screen-noise-investigation_file.tar.bz2.

 4. Provide issue repro information:
        * List steps or...
        * ...reference the specific ticket update which contains the
          steps.
        * Provide location and access information for repro machine,
          if available
----
How reproducible:
Always

Steps to Reproduce:
1. Install RHEL4.8 Beta on x86_64.
2. After the installation is completed, reboot the system.
3. Display the X graphical login screen.
----

 5. List any known hot-fix packages on the system
N/A

 6. List any customer applied changes from the last 30 days
N/A

**** Regressions ****

 1. What is the last known working version (specific
    name-version-release) of the package?
xorg-x11-6.8.2-1.EL.52

 2. What is the specific name-version-release of the package where
    the customer discovered the regression?
xorg-x11-6.8.2-1.EL.62

 3. Please paste the contents of the change log between these two
    versions. Not the entire changelog contents. (rpm -qp [package]
    --changelog
----
2009-03-06 22:00:00
Adam Jackson <ajax> 6.8.2-1.EL.62:
- xorg-x11-6.8.2-mga-g200se-dont-force-16bpp.patch: Fix option parsing to
work (and not be quite so crashy). (#485156)

2009-02-12 22:00:00
Adam Jackson <ajax> 6.8.2-1.EL.60:
- xorg-x11-6.8.2-mga-g200se-dont-force-16bpp.patch: Do vram probe
earlier, and fix inverted test sense. (#485156)

2009-02-12 22:00:00
Adam Jackson <ajax> 6.8.2-1.EL.61:
- xorg-x11-6.8.2-dev-input-mice.patch: Try /dev/input/mice when
searching for automatic mouse device. (#485305)

2009-01-28 22:00:00
Dave Airlie <airlied> 6.8.2-1.EL.59:
- xorg-x11-6.8.2-fix-powerpc-fonts.patch (#300911)

2009-01-27 22:00:00
Kristian Høgsberg <krh> 6.8.2-1.EL.58:
- Fix segfault on xfs drone server exit path.
- Update Sun6 keyboard patch to not drop the \| key.

2009-01-23 22:00:00
Dave Airlie <airlied> 6.8.2-1.EL.57:
- xorg-x11-6.8.2-xkbcomp-add-ibm_spacesaver.patch (#454831)
- xorg-x11-6.8.2-mga-g200se-dont-force-16bpp.patch (#469651)

2009-01-12 22:00:00
Adam Jackson <ajax> 6.8.2-1.EL.56:
- xorg-x11-6.8.2-g200-ev-wb.patch: Add support for Maxim and Winbond
variants. (#445498, #454780)

2009-01-08 22:00:00
Adam Jackson <ajax> 6.8.2-1.EL.55:
- Fix libGLU.a build on x86_64. (#432087)

2008-12-08 22:00:00
Adam Jackson <ajax> 6.8.2-1.EL.53:
- xorg-x11-6.8.2-drm-device-permissions.patch: Fix initial DRI device
permissions. (#462106)

2008-12-08 22:00:00
Adam Jackson <ajax> 6.8.2-1.EL.54:
- xorg-x11-6.8.2-fb-traps-crash.patch: Fix crash in trapezoid rendering
triggered by Firefox, among other things. (#463137)
----

 4. Have you been able to narrow the regression down to a specific version?
        * If so, which one?

 5. Have you been able to narrow the regression down to a specific
    patch committed? If so, which one?

 6. Provide the exact steps to reproduce the regression so
    SEG/engineering can narrow down where the regression was
    introduced if not already known and/or analyse how the
    regression is occurring.
Please see above **** Provide supporting info ****, section 4.

Thanks in advance.

Regards,
Masaki Furuta