Bug 144597 (dual-head)
Description
Jim Cornette
2005-01-09 08:13:45 UTC
Created attachment 109526 [details]
xorg.conf post dual-head configuration
This is what configuring dual-boot did to the xorg.conf file. The file was
edited manually to get the 815 card to work before configuration. I see
multiple monitor entries in the config file.
Created attachment 109527 [details]
This is my current Xorg.0.log file
Referring to many monitor entries, I had to manually add resolutions for the
815 card to be set to a decent resolution.
I still was not able to get the secondary monitor to work.
Created attachment 109528 [details]
Attempting to start a fresh config file for xorg-x11
I got the attached error when attempting to run system-config-display
--reconfig
I guess something is amiss to be able to create a file from scratch via
s-c-display.
Jim
I found an earlier bug report where I could almost get dual display working, but had a server lockup due to two "primary" devices. https://bugzilla.redhat.com/bugzilla/attachment.cgi?id=105631&action=view I'd like to figure out where this prolem is located and how to resolve this problem. I'm unsure if this is a configuration problem or a server problem. Excerpt from old bug report (II) PCI-to-ISA bridge: (II) Bus -1: bridge is at (0:31:0), (0,-1,-1), BCTRL: 0x0008 (VGA_EN is set) (!!) More than one possible primary device found (--) PCI: (0:2:0) Intel Corp. 82815 CGC [Chipset Graphics Controller] rev 2, Mem @ 0x40000000/26, 0x44000000/19 (--) PCI: (1:0:0) ATI Technologies Inc 264VT [Mach64 VT] rev 64, Mem @ 0xfd000000/24, I/O @ 0xd800/8, BIOS @ 0xfe9f0000/16 I see that the device is not recognized by the program during this trial. (II) ATI: Candidate "Device" section "Videocard1". (WW) ATI: PCI/AGP Mach64 in slot 1:0:0 could not be detected! With the earlier bug report, I might have had the primary as ATI and the secondary as Intel 815. I do not recall with certainty. The earlier report was with the ATI card as the primary and Intel 815 as secondary. (II) Primary Device is: (II) ATI: Candidate "Device" section "Videocard0". (II) ATI: Shared PCI/AGP Mach64 in slot 1:0:0 detected. (II) ATI: Shared PCI/AGP Mach64 in slot 1:0:0 assigned to active "Device" section "Videocard0". Rounding up the radeon 7200 bug that at least recognized the two cards is below for comparison. https://bugzilla.redhat.com/bugzilla/attachment.cgi?id=104238&action=view Adding to this report is the fact that I installed Severn and found dual-display functionality was in the future and has not been created yet. After upgrading to FC2 from RHL 9.0.93, I found that FC2 does have dual-display configuration tool. I have several files to attach to the report and sort of a story to go beforehand related to trials and tribulations encountered. After upgrading, server started as setup in Severn. After configuring system for dual boot Operation, I ended up with XFree86Config. The server refused to start after saving the file, then exiting. The reason seems to be obvious in the previously mentioned file. After having the problem with XFree86 not starting, I edited the xorg.conf file that was a copy of the XFree86Config file. The editing was severe and I have the same problem that FC3 and the version in Rawhide both exhibit.I cannot get to the second display for the ATI card. 815 is primary and the ATI is secondary. attachments for xorg.config and logs for both to follow. Created attachment 109748 [details]
Log for Severn system used yesterday, no dual-head ability
This is for reference to the system with an Intel 815 card
Created attachment 109749 [details]
This is the log for the newly created dual display via s-c-display
xorg-x11-6.7.0-2, system-config-display-1.0.14-1 intalled
Created attachment 109750 [details]
This is the backup before configuration of the server
The date is later than XFree86Config
Created attachment 109751 [details]
This is what s-c-display did to the config file
I believe ATI was "detected" as both cards.
Created attachment 109752 [details]
I edited this file to get X up
This is an amatuar attempt to get by the problem. I am up and running with the
original problem.
Created attachment 109753 [details]
current log after editing file
I cannot get the ATI card and the screen won't pan. I believe the ATI fails to
initialize. I am using kernel-2.6.5-1.358 for the kernel.
Created attachment 109851 [details]
rawhide freshly generated xorg.conf file
I ran system-config-display after eliminating the conf file for xorg-x11. I did
not get an error and was able to havethis file generated with the tool.
I cannot get the secondary display to work because of the below error.
(--) PCI: (1:0:0) ATI Technologies Inc 264VT [Mach64 VT] rev 64
(WW) ATI: PCI/AGP Mach64 in slot 1:0:0 could not be detected!
The kernel being run is 2.6.10-1.1089_FC4 and the version of X is
xorg-x11-6.8.1.902-1. The version of S-C-display is
system-config-display-1.0.25-1.
I assume that s-c-display is working somewhat, except my top resolution with
the tool is only 800x600 for the primary card "Intel 815" and likewise for the
ATI card mentioned with the error above.
(no editing performed to file)
Created attachment 109852 [details]
This is the current log for full details with tool generated config file.
My wife can live with 800x600. I however see the screen as too restrictive for
any useability. The maximum resolution for the LCD on the secondary is 800x600,
would this limit the max resolution for the 815 card?
Next test. ATI as primary, 815 as secondary. The last attempt resulted in an x
lockup but was recoverable from. X exited with error.
changing video card from 815 primary to ATI primary - internal secondary resulted in the i810 not even being an option within the dual configuration setup. There was not even an option to select another video card for the secondary. (just in case the configuration tool was not able to detect,configure the secondary video) As a reference, Windows ME displays a text message with directions to enable the display on screen. The message basically is straightforward. lspci detects the 815 video card with the flipped arrangement. I also seem to recall that previous versions of system-config-display allowed some configuration of video cards for the dual head setup. Without being able to manually select your secondary video card type or the tool being able to detect and configure the secondary properly, the tool is very ineffective. Created attachment 109897 [details]
switching ati 815 on fc2
This file was generated by system-config-display-1.0.14-1 with the ati as
primary and 815 as secondary.
Highlights of this configuration are that the two cards overlap memory
addresses. The login screen in runlevel 5 is on the ati and the mouse moves
within a messed up 815 display. Moving the mouse results in a block that moves
with the cursor. Eventually the display locks up. On the ati, the
initialization message keeps on cycling.
I have to hard reset to get the computer to halt. ctl-alt-del, alt-sysreq-b or
ctl-alt-backspace do nothing.
I will attach the log file for the x lockup also. No story though!
Created attachment 109898 [details]
the log that goes w/ the story
memory dual allocation. ATI has internal memory, 815 is well known what it
tries to do.
Looking at the settings that are used with a system that can successfuly run dual displays, I got the below values for each card. ATI ati_m64: Interrupt Request 11 Memory Range FD000000 - FDFFFFFF Input Output Range D000 - D0FF Memory Range FC900000 - FC90FFFF Intel 815: Memory Range 000A0000 - 000AFFFF Memory Range 000B0000 - 000BFFFF Input Output Range 03B0 - 03BB Input Output Range 03C0 - 03DF Interrupt Request 11 Memory Range F8000000 - FBFFFFFF Memory Range FEB80000 - FEBFFFFF Memory Range 000C0000 - 000CBFFF With the above values, I can set the 815 (primary display) to 1280x1024 and set the ATI (secondary display) to 800x600. Since these results are on a primitive OS, how can the same (or better) functionality be obtained using system-config-display? Or manual entries into xorg.conf to prevent memory range overlap between the two cards? Sifting through the Xorg.log that I submitted earlier, I found these values for the ati and the 815 cards relating to memory. It appears to me thatsome of the values are the same and others are quite different. I hope you can decipher the relations to the values and the comparative ranges that the working OS relays. ati > [0] -1 0 0xffe00000 - 0xffffffff (0x200000) MX[B](B) 815 > [0] -1 0 0xffe00000 - 0xffffffff (0x200000) MX[B](B) ati > [1] -1 0 0x00100000 - 0x3fffffff (0x3ff00000) MX[B]E(B) 815 > [1] -1 0 0x00100000 - 0x3fffffff (0x3ff00000) MX[B]E(B) ati > [2] -1 0 0x000f0000 - 0x000fffff (0x10000) MX[B] 815 > [2] -1 0 0x000f0000 - 0x000fffff (0x10000) MX[B] ati > [3] -1 0 0x000c0000 - 0x000effff (0x30000) MX[B] 815 > [3] -1 0 0x000c0000 - 0x000effff (0x30000) MX[B] ati > [4] -1 0 0x00000000 - 0x0009ffff (0xa0000) MX[B] 815 > [4] -1 0 0x00000000 - 0x0009ffff (0xa0000) MX[B] ati > [5] -1 0 0xfeb80000 - 0xfebfffff (0x80000) MX[B](B) 815 > [5] -1 0 0xfeb80000 - 0xfebfffff (0x80000) MX[B](B) ati > [6] -1 0 0xf8000000 - 0xfbffffff (0x4000000) MX[B](B) 815 > [6] -1 0 0xf8000000 - 0xfbffffff (0x4000000) MX[B](B) ati > [7] 0 0 0x000a0000 - 0x000affff (0x10000) MS[B] 815 > [7] -1 0 0x0000ffff - 0x0000ffff (0x1) IX[B] ati > [8] 0 0 0x000b0000 - 0x000b7fff (0x8000) MS[B] 815 > [8] -1 0 0x00000000 - 0x000000ff (0x100) IX[B] ati > [9] 0 0 0x000b8000 - 0x000bffff (0x8000) MS[B] 815 > [9] -1 0 0x0000df80 - 0x0000df9f (0x20) IX[B] ati > [10] -1 0 0x0000ffff - 0x0000ffff (0x1) IX[B] 815 > [10] -1 0 0x0000ef00 - 0x0000ef3f (0x40) IX[B] ati > [11] -1 0 0x00000000 - 0x000000ff (0x100) IX[B] 815 > [11] -1 0 0x0000e800 - 0x0000e8ff (0x100) IX[B] ati > [12] -1 0 0x0000df80 - 0x0000df9f (0x20) IX[B] 815 > [12] -1 0 0x0000efa0 - 0x0000efaf (0x10) IX[B] ati > [13] -1 0 0x0000ef00 - 0x0000ef3f (0x40) IX[B] 815 > [13] -1 0 0x0000ef80 - 0x0000ef9f (0x20) IX[B] ati > [14] -1 0 0x0000e800 - 0x0000e8ff (0x100) IX[B] 815 > [14] -1 0 0x0000ffa0 - 0x0000ffaf (0x10) IX[B] 815 > [15] -1 0 0x0000efa0 - 0x0000efaf (0x10) IX[B] 815 > [16] -1 0 0x0000ef80 - 0x0000ef9f (0x20) IX[B] 815 > [17] -1 0 0x0000ffa0 - 0x0000ffaf (0x10) IX[B] 815 > [18] 0 0 0x000003b0 - 0x000003bb (0xc) IS[B] 815 > [19] 0 0 0x000003c0 - 0x000003df (0x20) IS[B] adding the descrambling (or scrambling!) information from the working OS 815 >> Input Output Range 03B0 - 03BB 815 >> Input Output Range 03C0 - 03DF ati >> Input Output Range D000 - D0FF 815 >> Interrupt Request 11 ati >> Interrupt Request 11 815 >> Memory Range 000A0000 - 000AFFFF 815 >> Memory Range 000B0000 - 000BFFFF 815 >> Memory Range 000C0000 - 000CBFFF skipped MemoryRange 000CC000 - F7FFFFFF (what value, I don't know!) 815 >> Memory Range F8000000 - FBFFFFFF ati >> Memory Range FC900000 - FC90FFFF skipped MemoryRange FC910000 - FCFFFFFF (speculating, math rusty) ati >> Memory Range FD000000 - FDFFFFFF skipped MemoryRange FE000000 - FEB7FFFF (rusty?) 815 >> Memory Range FEB80000 - FEBFFFFF I replaced the ati card with a nvidia agp card. I found that the AGP card and the nvidia card do not seem to work together. I either see the nvidia card or see the intel 815 using system-config-display. Doing an /sbin/lspci showed the intel 815 in the listing. Anyway, since the radeon 7200, ati (both pci) overlapped in memory regions, I thought that I'd see what this 815/nvidia combination would do. I did not check the processes in windows ME yet to see if AGP and an internal card would work in tandem. I assume that an AGP card and an internal should work together. additional note. The resolution for the Internal 815 card went back up to how it was before adding either pci card. Also the nvidia card is using the supplied driver (nv) and not the binary, closed-source driver. Created attachment 110422 [details]
This is with nvidia as primary, no abilityto configure 815
Created attachment 110423 [details]
This is with the intel 815 as primary
The pci value for the nvidia driver was 2:0:0.
looking through the log file supplied in comment #20, I noticed that a particular section refers to the i810 driver, but references the ATI graphics card. The 000C000-000CBFFF is for the 815 card utilization. The overlap range is in comment #22 for the [3] overlapping range. As additional information, the 370SSM motherboard does not configure two cards when using an AGP card on either windows or Linux when choosing either card as the primary video. The PCI cards seem to be where it is possible to enable dual-display. (II) I810(1): Primary V_BIOS segment is: 0xc000 (II) I810(1): VESA BIOS detected (II) I810(1): VESA VBE Version 2.0 (II) I810(1): VESA VBE Total Mem: 16384 kB (II) I810(1): VESA VBE OEM: ATI MACH64 (II) I810(1): VESA VBE OEM Software Rev: 1.0 (II) I810(1): VESA VBE OEM Vendor: ATI Technologies Inc. (II) I810(1): VESA VBE OEM Product: MACH64VT (II) I810(1): VESA VBE OEM Product Rev: 01.00 excerpt from comment #22 ati > [3] -1 0 0x000c0000 - 0x000effff (0x30000) MX[B] 815 > [3] -1 0 0x000c0000 - 0x000effff (0x30000) MX[B] Also from log in comment #22 is this excerpt. The ATI has builtin memory and should not require shared memory allocation. The shared memory seems to be the problem that causes the overlapping resources. (II) ATI: Shared PCI/AGP Mach64 in slot 1:0:0 detected. (II) ATI: Shared PCI/AGP Mach64 in slot 1:0:0 assigned to active "Device" section "Videocard0". Created attachment 120037 [details]
Development, still problems, different symptom, xorg.conf
Strange changes happened with the cards with present rawhide. The Intel is set
to primary and starts the boot process up to a point. Then the Matrox (yes,
changed to anotherPCI card, same problem with released version, no display at
all) - The boot process continues on the Matrox. When the GUI starts, the GUI
is displayed on the Intel card.
As a note, a question was asked on the xorg-x11 list and this log is for
follow-up.
Created attachment 120038 [details]
Here is the xorg-x11 log file
part 2 for development with Intel internal Matrox Millenium G200 secondary
video
Created attachment 120091 [details]
Xorg log after configuring with system-config display
No editing and a fresh configuration file
Created attachment 120093 [details]
xorg log after using s-c-display to configure
This shows a normal Intel screen but a garbled mga screen.
Created attachment 120094 [details]
trying with mgs replaced with vesa
With vesa, the screen only displays anything when you switch to vt 1-6. Moving
the mouse does not exit the screen toward any corner.
Created attachment 120095 [details]
just changed to vesa for display 2 (MGA)
Created attachment 120096 [details]
change to rightof
I added pci values for the Intel primary card. Since vesa did not work any
different than single head configuration, I tried with mga again. The screen
moves right of the main screen.
Created attachment 120097 [details]
this is the last config file for the display
last file for now
Though this problem still exists, I am closing this bug since dual head operation is not a resource payed attention to and no developer actions in 10 months. |