Re: i810 VBE problems - was Re: DDC Atoms Xinerama
On 17 Mar 2003, Ben Guthro wrote: In xc/programs/Xserver/hw/xfree86/CHANGELOG 754. Fixed VBE EDID read: due to a missing register setting read ended in endless loop on certain systems (Egbert Eich). On a good day an endless loop would be aborted. It would be worth trying v4.2 Typo; I meant it would be worth trying 4.3. I am currently running version 4.2.0 (18 Jan 2002) I have up to change 690 in my CHANGELOG. So, this may be my problem. However, before trying the most recent version, I set my ATI Radeon card up to run in dual head mode, one off the DVI connector, one off the VGA. This, interestingly produced the same results xprop -display :0.1 -root 8x XFree86_DDC_EDID1_RAWDATA XFree86_DDC_EDID1_RAWDATAAborted Do you still think that its a driver problem? Sorry, driver was the wrong word. However the radeon DDC code has been seriously overhauled since 4.2. I really think you should upgrade to 4.3 before going further. -- Dr. Andrew C. Aitchison Computer Officer, DPMMS, Cambridge [EMAIL PROTECTED] http://www.dpmms.cam.ac.uk/~werdna ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: i810 VBE problems - was Re: DDC Atoms Xinerama
Andrew C Aitchison writes: In xc/programs/Xserver/hw/xfree86/CHANGELOG 754. Fixed VBE EDID read: due to a missing register setting read ended in endless loop on certain systems (Egbert Eich). On a good day an endless loop would be aborted. You should tell this to the NVIDIA BIOS programmers ;-) Egbert. ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: DDC Atoms Xinerama
Other than that I've never seen an XFree86_DDC_EDID2_RAWDATA property, so would be interested to see your log file. Unfortunately, I have been unable to reproduce this result. I swapped out My Mistubishi Diamond Pro 2060u for a Lacie electron 19b IV for some other testing. When I put the 2060u back on to attempt to reproduce this DDC_EDID2 anomoly, it did not reproduce it. The log file is attached below, in case you might be able to extract any pertinent information on this, even if I could not reproduce the results. However, I'm still left with the dilemma of being unable to inventory all the monitors on the system. I realize that the network transarency schema says that in some instances this does not make sense to do. However, in Windows, Mac OS9, and OS X there is a mechanism to retrieve information on the monitors on the desktop (like each one's EDID), be it virtual or not. Is there no equivalent in X? Ben Guthro I have more machines than monitors, so haven't looked very hard at the dual head case, and wouldn't be suprised if there are cases where the second head EDID data is wrong or not made available. log file XFree86 Version 4.2.0 (Red Hat Linux release: 4.2.0-8) / X Window System (protocol Version 11, revision 0, vendor release 6600) Release Date: 23 January 2002 nlIf the server is older than 6-12 months, or if your card is nlnewer than the above date, look for a newer version before nlreporting problems. (See http://www.XFree86.Org/) Build Operating System: Linux 2.4.17-0.13smp i686 [ELF]=20 Build Host: daffy.perf.redhat.com nl=20 Module Loader present Markers: (--) probed, (**) from config file, (=3D=3D) default setting, nl (++) from command line, (!!) notice, (II) informational, nl (WW) warning, (EE) error, (NI) not implemented, (??) unknown. (=3D=3D) Log file: /var/log/XFree86.0.log, Time: Mon Mar 17 07:55:37 2003 (=3D=3D) Using config file: /etc/X11/XF86Config-4 (=3D=3D) ServerLayout XFree86 Configured (**) |--Screen Screen0 (0) (**) | |--Monitor NEC FP1375X (**) | |--Device ATI|Radeon 7500 QW (**) |--Screen Screen1 (1) (**) | |--Monitor MitsDPro2060 (**) | |--Device IBM (**) |--Input Device Mouse0 (**) |--Input Device Keyboard0 (**) Option XkbLayout us (**) XKB: layout: us (=3D=3D) Keyboard: CustomKeycode disabled (**) FontPath set to unix/:7100 (=3D=3D) RgbPath set to /usr/X11R6/lib/X11/rgb (=3D=3D) ModulePath set to /usr/X11R6/lib/modules (--) using VT number 7 nlnl (II) Open APM successful (II) Module ABI versions: nlXFree86 ANSI C Emulation: 0.1 nlXFree86 Video Driver: 0.5 nlXFree86 XInput driver : 0.3 nlXFree86 Server Extension : 0.1 nlXFree86 Font Renderer : 0.3 (II) Loader running on linux (II) LoadModule: bitmap (II) Loading /usr/X11R6/lib/modules/fonts/libbitmap.a (II) Module bitmap: vendor=3DThe XFree86 Project nlcompiled for 4.2.0, module version =3D 1.0.0 nlModule class: XFree86 Font Renderer nlABI class: XFree86 Font Renderer, version 0.3 (II) Loading font Bitmap (II) LoadModule: pcidata (II) Loading /usr/X11R6/lib/modules/libpcidata.a (II) Module pcidata: vendor=3DThe XFree86 Project nlcompiled for 4.2.0, module version =3D 0.1.0 nlABI class: XFree86 Video Driver, version 0.5 (II) PCI: Probing config type using method 1 (II) PCI: Config type is 1 (II) PCI: stages =3D 0x03, oldVal1 =3D 0x8001480c, mode1Res1 =3D 0x8000 (II) PCI: PCI scan (all values are in hex) (II) PCI: 00:00:0: chip 8086,7120 card , rev 02 class 06,00,00 hdr = 00 (II) PCI: 00:01:0: chip 8086,7121 card 8086,7123 rev 02 class 03,00,00 hdr = 00 (II) PCI: 00:1e:0: chip 8086,2418 card , rev 01 class 06,04,00 hdr = 01 (II) PCI: 00:1f:0: chip 8086,2410 card , rev 01 class 06,01,00 hdr = 80 (II) PCI: 00:1f:1: chip 8086,2411 card , rev 01 class 01,01,80 hdr = 00 (II) PCI: 00:1f:2: chip 8086,2412 card , rev 01 class 0c,03,00 hdr = 00 (II) PCI: 00:1f:3: chip 8086,2413 card , rev 01 class 0c,05,00 hdr = 00 (II) PCI: 01:08:0: chip 127a,4320 card 1235,4320 rev 00 class 04,01,00 hdr = 80 (II) PCI: 01:08:1: chip 127a,4321 card 1235,4321 rev 00 class 07,80,00 hdr = 80 (II) PCI: 01:08:2: chip 127a,4322 card 1235,4322 rev 00 class 09,80,00 hdr = 80 (II) PCI: 01:09:0: chip 1113,1211 card 103c,1207 rev 10 class 02,00,00 hdr = 00 (II) PCI: 01:0a:0: chip 1002,5157 card 1002,013b rev 00 class 03,00,00 hdr = 00 (II) PCI: End of PCI scan (II) LoadModule: scanpci (II) Loading /usr/X11R6/lib/modules/libscanpci.a (II) Module scanpci: vendor=3DThe XFree86 Project nlcompiled for 4.2.0, module version =3D 0.1.0 nlABI class: XFree86 Video Driver, version 0.5 (II) UnloadModule: scanpci (II) Unloading /usr/X11R6/lib/modules/libscanpci.a (II) Host-to-PCI bridge: (II) PCI-to-ISA bridge: (II) PCI-to-PCI bridge: (II) Bus 0: bridge is at (0:0:0), (-1,0,0), BCTRL: 0x08 (VGA_EN is set) (II) Bus 0 I/O range: nl[0] -1 0x - 0x (0x1) IX[B] (II) Bus 0 non-prefetchable memory
Re: DDC Atoms Xinerama
On 17 Mar 2003, Ben Guthro wrote: However, I'm still left with the dilemma of being unable to inventory all the monitors on the system. I realize that the network transarency schema says that in some instances this does not make sense to do. However, in Windows, Mac OS9, and OS X there is a mechanism to retrieve information on the monitors on the desktop (like each one's EDID), be it virtual or not. Is there no equivalent in X? If you want to inventory all monitors connected to a machine, ddcprobe (ships with Red Hat 8.0) should do that independently of X. I suspect that it doesn't handle multiple monitors very well. If you want to inventory the monitors on a desktop, xprop -root 0x XFree86_DDC_EDID1_RAWDATA should work even across the network. However I haven't tried it on a dual head with two graphics cards. I know that drivers for several dual head cards don't correctly return the EDID info for both displays; do the other OSes you mention get this right ? What does EDID info even mean on a laptop display - my laptop BIOS doesn't return EDID info for the builting screen, but does return the EDID from an external monitor if connected. (II) Loading sub module vbe (II) LoadModule: vbe (II) Loading /usr/X11R6/lib/modules/libvbe.a (II) Module vbe: vendor=The XFree86 Project compiled for 4.2.0, module version = 1.0.0 ABI class: XFree86 Video Driver, version 0.5 (II) Loading sub module int10 (II) LoadModule: int10 (II) Reloading /usr/X11R6/lib/modules/linux/libint10.a (II) I810(1): initializing int10 (EE) I810(1): Cannot read V_BIOS (II) I810(1): this driver cannot do DDC without VBE This looks like a driver problem - does the DDC work if you driver this card as a single head ? -- Dr. Andrew C. Aitchison Computer Officer, DPMMS, Cambridge [EMAIL PROTECTED] http://www.dpmms.cam.ac.uk/~werdna ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: DDC Atoms Xinerama
I know that drivers for several dual head cards don't correctly return the EDID info for both displays; do the other OSes you mention get this right ? Yes, they do. What does EDID info even mean on a laptop display - my laptop BIOS doesn't return EDID info for the builting screen, but does return the EDID from an external monitor if connected. In both Windows, and Mac, something IS returned , though it may not be defined...something to the effect of Default Monitor is returned. My IBM Thinkpad doesn't seem to return anything either, unless a monitor is plugged in. This looks like a driver problem - does the DDC work if you driver this card as a single head ? yes. When I remove the ATI Radeon 7500, the i810 correctly gets an EDID. Ben Guthro ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: DDC Atoms Xinerama
On 17 Mar 2003, Ben Guthro wrote: This looks like a driver problem - does the DDC work if you driver this card as a single head ? yes. When I remove the ATI Radeon 7500, the i810 correctly gets an EDID. Do you get both EDIDs if you make the i810 the first head in your XF86Config ? --- I've hooked up a dual head system with two ATI cards (r128 and mach64) and confirmed that xprop -display :0.0 -root 8x XFree86_DDC_EDID1_RAWDATA and xprop -display :0.1 -root 8x XFree86_DDC_EDID1_RAWDATA do report the correct EDID info for each monitor. Of course this wont work for Xinerama, since that deliberately hides the display distinction. -- Dr. Andrew C. Aitchison Computer Officer, DPMMS, Cambridge [EMAIL PROTECTED] http://www.dpmms.cam.ac.uk/~werdna ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
i810 VBE problems - was Re: DDC Atoms Xinerama
On 17 Mar 2003, Ben Guthro wrote: I've hooked up a dual head system with two ATI cards (r128 and mach64) and confirmed that xprop -display :0.0 -root 8x XFree86_DDC_EDID1_RAWDATA and xprop -display :0.1 -root 8x XFree86_DDC_EDID1_RAWDATA do report the correct EDID info for each monitor. I tried this, and got a response of XFree86_DDC_EDID1_RAWDATAAborted when trying on the i810. This occured with both setting the i810 as the first head, as well as the second. Any thoughts? In xc/programs/Xserver/hw/xfree86/CHANGELOG 754. Fixed VBE EDID read: due to a missing register setting read ended in endless loop on certain systems (Egbert Eich). On a good day an endless loop would be aborted. It would be worth trying v4.2 Other than that, I think you will have to go to the i810 driver maintainers. -- Dr. Andrew C. Aitchison Computer Officer, DPMMS, Cambridge [EMAIL PROTECTED] http://www.dpmms.cam.ac.uk/~werdna ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: i810 VBE problems - was Re: DDC Atoms Xinerama
In xc/programs/Xserver/hw/xfree86/CHANGELOG 754. Fixed VBE EDID read: due to a missing register setting read ended in endless loop on certain systems (Egbert Eich). On a good day an endless loop would be aborted. It would be worth trying v4.2 I am currently running version 4.2.0 (18 Jan 2002) I have up to change 690 in my CHANGELOG. So, this may be my problem. However, before trying the most recent version, I set my ATI Radeon card up to run in dual head mode, one off the DVI connector, one off the VGA. This, interestingly produced the same results xprop -display :0.1 -root 8x XFree86_DDC_EDID1_RAWDATA XFree86_DDC_EDID1_RAWDATAAborted Do you still think that its a driver problem? Ben Guthro ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: DDC Atoms Xinerama
On 13 Mar 2003, Ben Guthro wrote: Xinerama aside, even in a dual head system running without Xinerama, I am having a difficult time retrieving the XFree86_DDC_EDID1_RAWDATA for the second (non-primary) screen xlsatoms |grep EDID yields 69XFree86_DDC_EDID1_RAWDATA 209 XFree86_DDC_EDID1_RAWDATA 489 XFree86_DDC_EDID2_RAWDATA however when running xprop -root |grep EDID on the second screen, it yields no results, and running it on the first screen prints one atom. so...what window are those other two attached to so I can retrieve that information with XGetWindowProperty? Wow, I'd forgotten all about that. In your case xprop -root on the first screen prints the data in atom 209. Atom 69 is a hack; because the EDID data is found before the screen is created, there isn't yet a root window to attach it to. Instead we create a screenless atom (69), then copy it when we create the screen (in xf86Init.c xf86CreateRootWindow() ). XFree86_DDC_EDID2_RAWDATA isn't the data for the second screen. It is used when the monitor returns data in a EDID v2 data structure (256 bytes as opposed to 128). I've never seen a monitor which returns that. There are however monitors which return 128 bytes, but report that they are version 2, since they return it in the structure described in version 2 of the EDID standard (this is really EDID v1.1). We detect this case by the EDID checksum, and create both properties. grepping for EDID in the log file should alert you to that case. Other than that I've never seen an XFree86_DDC_EDID2_RAWDATA property, so would be interested to see your log file. I have more machines than monitors, so haven't looked very hard at the dual head case, and wouldn't be suprised if there are cases where the second head EDID data is wrong or not made available. -- Dr. Andrew C. Aitchison Computer Officer, DPMMS, Cambridge [EMAIL PROTECTED] http://www.dpmms.cam.ac.uk/~werdna ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
DDC Atoms Xinerama
I am attempting to do some things that are somewhat unconventional, and have so far been unable to find documentation on how to accomplish this. I have poted a similar query to some newsgroups, but so far have not got an answer I need. I am in the process of writing hardware / software monitor color calibration software for linux under Qt. In doing so, determining the monitor hardware that is currently running is quite paramount. The DDC seems to probe the monitors, then store this EDID data in an atom called XFree86_DDC_EDID1_RAWDATA While this method seems to work great in a single monitor environment, a Xinerama environment seems to be different. Since there seems to be a shared Screen between the Xinerama displays - is this EDID data still stored somewhere? Though the full EDID data would be nice, ultimately, I merely need a way off accociating specific screen coordinates to a monitor Vendor / Model / serial number triplet Any insight into this matter would be greatly apprecated Ben Guthro Sequel Imaging, Inc. -- Ben Guthro [EMAIL PROTECTED] Sequel Imaging, Inc. 603.425.2170 ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: DDC Atoms Xinerama
On 13 Mar 2003, Ben Guthro wrote: I am in the process of writing hardware / software monitor color calibration software for linux under Qt. In doing so, determining the monitor hardware that is currently running is quite paramount. The DDC seems to probe the monitors, then store this EDID data in an atom called XFree86_DDC_EDID1_RAWDATA While this method seems to work great in a single monitor environment, a Xinerama environment seems to be different. Since there seems to be a shared Screen between the Xinerama displays - is this EDID data still stored somewhere? Though the full EDID data would be nice, ultimately, I merely need a way off accociating specific screen coordinates to a monitor Vendor / Model / serial number triplet Any insight into this matter would be greatly apprecated Xcms is an existing industry standard color management scheme for X servers. It also uses atoms (such as XDCCC_LINEAR_RGB_MATRICES) to maintain state and make it available to applications/libraries. See man xcmsdb for the way one implementation handles these atoms. I haven't looked at the Xcms code for creating these atoms, so I don't know whether it copes with Xinerama. (I do have a program which can be used with xcmsdb to read XFree86_DDC_EDID1_RAWDATA and create the XDCCC_ atoms, but conventional wisdom is that the color data in many monitor EDIDs is worse than using established defaults - and I've seen figures to prove that). Xinerama didn't exist when I wrote the code for the atoms XFree86_DDC_EDID1_RAWDATA and XFree86_DDC_EDID2_RAWDATA, and I've never got around to looking at what to do. As you say, Xinerama presents a single screen, and thus properties on a single root window, but there are multiple sets of EDID data to present. I've heard talk of ways of refering to the separate heads for use in commands which take a display as argument, but I haven't seen anything working. With Xinerama, since a window may be moved between screens at any time (or appear on more than one at the same time) it isn't really appropriate for a client to ask about the monitor it is displayed on, but I can see that it may be sensible for configuration type applications to have access to both sets of monitor information at the same time. The X log file contains most of the EDID data (and I have a program which can parse most of it) so that might be the easiest way to get what you need. I haven't tried very hard, but I can't remember ever seeing correct EDID data for two monitors into the XFree86_DDC_EDID1_RAWDATA atoms of a dual head setup, and I would definitely not like to rely on it for a dual-head card. Sorry I can't be very positive. -- Dr. Andrew C. Aitchison Computer Officer, DPMMS, Cambridge [EMAIL PROTECTED] http://www.dpmms.cam.ac.uk/~werdna ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
