Red Hat Bugzilla – Bug 490659
[RHEL5.3G] mga driver can not provide 24bpp on X
Last modified: 2010-09-24 10:51:16 EDT
Created attachment 335531 [details]
MGAoutput.zip - sosreport appears to be in the zip file
State the problem
1. Provide time and date of the problem
2. Indicate the platform(s) (architectures) the problem is being
RHEL5.3GA, i686, x86_64, xorg-x11-drv-mga-1.4.2-10.el5
3. Provide clear and concise problem description as it is
understood at the time of escalation
* Observed behavior
The issue is re-create from IT#268374 on NEC-Certification and this NEC's blocker for RHEL5.4.
This is problem about xorg-x11-drv-mga-1.4.2-10.el5 on RHEL5.3 and
this prevent 24bit color on many NEC servers using "MGA G200e [Pilot] ServerEngines (SEP1) (rev 02)".
Since this driver is hard coded for using 16bpp with X, though video card have efficien memory (8MB) for 24bpp.
This cause limitation issue for NEC customers.
* Desired behavior
Mga driver can privide 24bpp on X.
4. State specific action requested of SEG
NEC tell me that this is must be fixed on RHEL5.4.
In the case I need exception for RHEL5.4.
5. State whether or not a defect in the product is suspected
* Provide Bugzilla if one already exists
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.
Hardware Certification Test failed with "Warning ! Current color depth: 16 does not meet the minimum requirements !" for G200SE with mga driver. Video.py failed on normal kernel, Xen kernel, GuestOS
There is excerpt of error from video.py:
(EE) MGA(0): Forcing 16bpp on G200SE due to low memory.
(EE) MGA(0): Using settings from depth 24 section
(WW) MGA(0): xf86UnMapVidMem: cannot find region for [0x2b59d563e000,0x7f0000]
FreeFontPath: FPE "unix/:7100" refcount is 2, should be 1; fixing.
x11perf returned 0
Screen resolution: 1280x768
Color depth: 16
Warning ! Current color depth: 16 does not meet the minimum requirements !
Video test on device video-card FAILED
...finished running ./video.py, exit code=1
8. This is especially important for severity one and two
issues. What is the impact to the customer when they experience
* 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."
Many NEC server used MGA G200, "MGA G200e [Pilot] ServerEngines (SEP1) (rev 02)".
where RHEL Certification requires 24bpp on X, but xorg-x11-drv-mga-1.4.2-10.el5 can not provide 24bpp.
They need mga driver for supporting customers, so this is big impact for NEC's server business.
**** Provide supporting info ****
1. State other actions already taken in working the problem:
* tech-list, google searches, fulltext, consulting with
* Provide any relevant data found
I talked this with Rob Landry-san about BZ#479919, and very much, great helpful for current RHEL5.3 Hardware Certification.
But NEC strongly need to be fixed on *RHEL5.4* for this problem.
2. Attach sosreport
Attached MGAoutput.zip, this include sosreport and result of video.py in hts.
# unzip -l MGAoutput.zip
Length Date Time Name
-------- ---- ---- ----
95554 02-18-09 13:37 MGAoutput.log
104019 02-18-09 13:37 VESAoutput.log
566241 02-18-09 13:37 T120aMel5x64.2009021792353.tar.bz2
765814 3 files
3. Attach other supporting data
4. Provide issue repro information:
* List steps or...
* ...reference the specific ticket update which contains the
* Provide location and access information for repro machine,
Always on X.
5. List any known hot-fix packages on the system
6. List any customer applied changes from the last 30 days
The error pops out for mga driver claiming the card is G200SE chip,
which hardcoats and caps the bpp at the max to 16. I'm guessing there
are variants of cards with the same chip but different ram, as lspci
indicates the card ready with 8MB, which would allow more bpp than
it is limited at the moment (if it's designed that way).
The ultimate of the vendor's wish is certifying the box, so we have
two things on our mind: 1. lift off 24bpp hurdle on 5.3 hts certification
and allow 16bpp as it was in 5.2 cert, or 2. Lift off 16bpp limit
or do anything to let the specific card get its way around it.
We are not sure if 2. is feasible for any card that returns reply
claiming G200SE, so if this is the path we are to take, then we
need to ask someone that's good with mga driver/products. Note
we are still not sure if this is actually functional with lifted-off
*** This bug has been marked as a duplicate of bug 479919 ***