Re: [Xpert]compiling libXxf86dga.so.1
If you have a static lib, libXxf86dga.a, you don't need sources to make a shared lib. Just do this: ld -shared -soname libXxf86dga.so.1 -o libXxf86dga.so.1 /usr/X11R6/lib/libXxf86dga.a Joe Krahn k-essej wrote: i'm in need of libXxf86dga.so.1, i have the source of my current xfree86 build (4.2.1). now how do i compile the library? blah blah bluh send mail to [EMAIL PROTECTED] pleaze, no mayonaisse. yes, http://fastmail.fm i am at. ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
[Xpert]Strange: Netscape briefly blank video output
Have other people noticed this? In composing a message in netscape, add an attachment, then right-click on the icon in the attachment pane. When I do this, the video output blanks briefly before the menu comes up. It even makes my monitor click as it has to re-sync. I find it strange that Netscape could do this. Maybe it is in part a server bug? I've seen this on RedHat Linux with different Netscape 4.x and with different video cards. Anyone have any ideas? Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]dual head on a Radeon VE
Geoffrey wrote: > My assumption is that this patch deals with the fact the oem card has > two digital outputs? At least that's what I understand, mine has one > vga and one digital. I don't think this should matter though. I've got > a converter on my digital and before starting X, I've got mirrored > output to both monitors. I think it is actually due to a different BIOS on the card. The Dell card has a single digital socket, and a choice of two Y cables to plug in: dual digital or dual VGA. I'm using the dual VGA. I don't get mirrored output before starting X. SO, your card is quite different from Dell OEM, but maybe it's still worth trying a patched XFree86. I can send you a patched executable to test, if you want to try it. Joe ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]dual head on a Radeon VE
Geoffrey wrote: > > I've got a 'Screen 0' in my first Device section and 'Screen 1' in the > second. > > The examples I used were for single card with dual outputs. One for a > Matrox card that had two vga outputs (linux journal article) and another > was an example config file from someone who has a Radeon VE oem card. > He's using some hacked version of Xfree though. > Do you also have different identifiers for the two device sections? Here's the Device/Screen sections from a Radeon VE that works for me. BUT, I have a Dell OEM version and had to apply the patch/hack to XFree for the "Crt2Dac2" option. Does this work for you? Joe Krahn -- Section "Device" Identifier "Radeon0" Driver "radeon" VendorName "ATI Radeon AGP" BoardName "ATI Radeon AGP" BusID"PCI:1:0:0" Screen 0 EndSection Section "Device" Identifier "Radeon1" Driver "radeon" VendorName "ATI Radeon AGP" BoardName "ATI Radeon AGP" Option "Crt2Dac2" BusID"PCI:1:0:0" Screen 1 EndSection Section "Screen" Identifier "Screen0" Device "Radeon0" Monitor "DELL" DefaultDepth24 Subsection "Display" Depth 24 Modes "1280x1024" EndSubsection EndSection Section "Screen" Identifier "Screen1" Device "Radeon1" Monitor "DTI3D" DefaultDepth16 Subsection "Display" Depth 16 Modes "1024x768" EndSubsection EndSection ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Genlocking X
Christian Berger wrote: > > Devrim Erdem schrieb: > > > > Hi, > > > > I want to genlock two Nvidia Xfree running PCs to an external source. I > > am looking for a hardware or software solution. I appreciate any links > > or pointers on the subject ? > > Well there is a hardly documented feature on the feature connector which > might be used for that, it has some "external-sync-enable" pins. I've > tried to mess around with them, but had no luck. > > > If there is none, I am looking into writing a software genlock but I am > > clueless about how Xfree manages genlock ( or VSync and HSync ). > > I'm not even sure if that's possible by software, I mean you would have > to tell the graphics card when to start the image quite acurately. As > well as you have to have fine control over the clock. There is a software genlock at sourceforge: http://sourceforge.net/projects/softgenlock It is made for synchronizing for multi-screen stereo, where horizontal accuracy doesn't matter. If you're trying to get a precise alignment for overlaying video, software genlock is definitely not going to work. Getting genlock to work well also requires a really good SYNC signal. Getting it to work well is hard to do. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Redhat 7.3 | Radeon VE | Dual Head | Help Needed
Alex Valys wrote: > > Having just installed RedHat 7.3, in the hopes that this would be easier (it's > not), I'm still trying to coax dual-head support out of my new Radeon VE > (DELL OEM). > > When I run XFree86 with the configuration file below, it works fine on the > primary display. Nothing appears on the secondary, and it doesn't make any > noises either. Any help would be greatly appreciated, as I've been at this > for a few hours. > Unfortunately, the DELL OEM Radeon has a different BIOS. For now, it requires a patch to the XFree86 source, which was posted a while back to Xpert. It hasn't made it in to the CVS. A patch relative to the current XFree86 CVS is below. You also need to set Option "Crt2Dac2" in the 2nd screen entry. Joe Krahn *** radeon_driver.c Tue Jun 18 21:10:56 2002 --- radeon_driver.c.patched Tue Jun 18 21:12:52 2002 *** *** 137,142 --- 137,143 OPTION_CLONE_HSYNC, OPTION_CLONE_VREFRESH, OPTION_FBDEV, + OPTION_CRT2_DAC2, OPTION_VIDEO_KEY } RADEONOpts; *** *** 163,168 --- 164,170 { OPTION_CLONE_HSYNC,"CloneHSync", OPTV_ANYSTR, {0}, FALSE }, { OPTION_CLONE_VREFRESH, "CloneVRefresh",OPTV_ANYSTR, {0}, FALSE }, { OPTION_FBDEV, "UseFBDev", OPTV_BOOLEAN, {0}, FALSE }, + { OPTION_CRT2_DAC2, "Crt2Dac2", OPTV_BOOLEAN, {0}, FALSE }, { OPTION_VIDEO_KEY, "VideoKey", OPTV_INTEGER, {0}, FALSE }, { -1,NULL, OPTV_NONE,{0}, FALSE } }; *** *** 2376,2382 pPriv = xf86GetEntityPrivate(pScrn->entityList[0], gRADEONEntityIndex); pRADEONEnt = pPriv->ptr; ! if (pRADEONEnt->BypassSecondary) return FALSE; pRADEONEnt->pSecondaryScrn = pScrn; } else { DevUnion *pPriv; --- 2378,2390 pPriv = xf86GetEntityPrivate(pScrn->entityList[0], gRADEONEntityIndex); pRADEONEnt = pPriv->ptr; ! /* XXX Note to Maintainer: The only reason we would be here ! * is if the user entered a second "Screen" definition for ! * this card, presumably because there are two screens. ! * Shouldn't this automatically override presumed BIOS settings? ! if(pRADEONEnt->BypassSecondary) return FALSE; ! */ ! pRADEONEnt->pSecondaryScrn = pScrn; } else { DevUnion *pPriv; *** *** 4056,4061 --- 4064,4072 ((info->SavedReg.disp_output_cntl & ~(CARD32)RADEON_DISP_DAC_SOURCE_MASK) | RADEON_DISP_DAC_SOURCE_CRTC2); + else if (xf86ReturnOptValBool(info->Options, OPTION_CRT2_DAC2, FALSE)) + save->dac2_cntl = info->SavedReg.dac2_cntl + | RADEON_DAC2_DAC2_CLK_SEL; else save->dac2_cntl = (info->SavedReg.dac2_cntl /* | RADEON_DAC2_DAC2_CLK_SEL */ ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]LocalClient extensions vs. ssh
Egbert Eich wrote: > > Joe Krahn writes: > > Aside from LocalClient returning false, DGA requests should fail, right? > > LocalClient() returns TRUE therefore DGA doesn't fail. Oh... I understand now. I didn;t know about the LocalClient fucntion, but it's whole purpose is to supply info so that, for example, DGA can do the right thing. Clearly, LocalClient needs a foolproof way to determine if a client really is local. Not understanding how DGA memory access works, here's a question: If the tunneled server gives permission for direct memory access, how does the local user get access to write into system memory without simply getting a segfault? Also, what happens when a client is accessing server memory, and you switch to a different virtual terminal? Thanks, Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]LocalClient extensions vs. ssh
Egbert Eich wrote: > > When running X request thru a ssh tunnel LocalClient() returns > true as the sshd connects to the Xserver from the same machine. > > This leads to unpredictable results and even system crashes > if an application that can only run locally is started on a remote > system. > > If for example an application wants to address the framebuffer > directly thru DGA it will try to map the framebuffer on the machine > it is started on not the one where the Xserver lives on. > > Even though this requires root privileges I do consider this to > be a problem that should be adressed. > > Any ideas? > Aside from LocalClient returning false, DGA requests should fail, right? Trying GL over the tunnel correctly gives an error message and reverts to indirect rendering. I don't know much about DGA, but my understanding is that it has a lot of issues because it is sort of a big hack (but useful), and the real solution is to replace it with real protocols like XVideo and a similar capture extension. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
[Xpert]DRI and glXCopyContext, or one context for 2 windows
I am trying to use one context state for two windows. Is glXCopyContext supposed to work for DRI? Alternatively, I can actually use one context and point it to the window I want to render in. Apparently, this is an unusual thing to do. It works on SGI (as uaual) but was able to crash XFree86 DRI, Nvidia, and Xig servers. Nvidia and Xig fixed the problem, but DRI is still broken. DRI doesn't crash from this any more, but it gives serious artifacts. Without glXCopyContext, a workaround is really hard, other than not using DRI. Any ideas? This is probably another case of trying the DRI CVS. Thanks, Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]how to stop window gravity
etienne wrote: > > Hello, > > I am running an application whose output is going to a > projector. I have configured my geforce2 so that the second > output takes the top left hand corner (640 x 480) of my screen (1400 x > 1050). > > Since my window has a border created by the window manager, I want to > position my X window at -2-10 (or something similar) so that the window > manager's border is not visible ... but this always defaults to the > bottom right hand corner of the screen. > > how do I turn window gravity off ? > > this doesn't work: > > hint.win_gravity = ForgetGravity; > hint.flags |= PWinGravity; > > Also, is there a way to stop the window manager form drawing a screen at > all ? > > etienne > You are supposed to be able to get the window manager to use static gravity, so that positioning to 0,0 places your window at the origin, rather than the decorations. There is some variability among window managers, though. Look at the source for xawtv fullscreen mode, which places it's window at the origin as you are trying to. Alternatively, you can create the window with override_redirect enabled and bypass the window manager altogether. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
[Xpert]ATI glBitmap rendering bug
Using glBitmap with ATI DRI driver while the depth test is on is resulting in randomly passing/failing each bit, resulting in garbage. There are patterns in the pass/fail that make it look like a depth lookup is occuring, but in a messed up way. Here is a quick demo. Can some other ATI user test this out? I am using a current CVS XFree86 on a Radeon. The 4.2 release was completely unstable for my Radeon. Joe Krahn --- glbitmapbug.c: #include #include #include #include GLboolean depth=GL_FALSE; void gprint(int x, int y, char *fmt, ... ) { va_list ap; char string[256]; int i; va_start(ap, fmt); vsprintf(string, fmt, ap); va_end(ap); glRasterPos2i(x, y); for (i = 0; i < strlen(string); i++) glutBitmapCharacter(GLUT_BITMAP_HELVETICA_18, string[i]); } void key(unsigned char key, int x, int y) { if (key == 27) { exit(0); } if (key=='z'){ depth=!depth; if (depth)printf("Enabling depth test.\n"); else printf("Disabling depth test.\n"); glutPostRedisplay(); } } int width,height; void reshape(int w, int h){ width=w; height=h; glViewport(-0.1,-0.1,w,h); glMatrixMode(GL_PROJECTION_MATRIX); glLoadIdentity(); gluOrtho2D(0,w,h,0); glutPostRedisplay(); } void display(void) { int i; if (depth) glEnable(GL_DEPTH_TEST); else glDisable(GL_DEPTH_TEST); glDrawBuffer(GL_FRONT); glClear(GL_COLOR_BUFFER_BIT|GL_DEPTH_BUFFER_BIT); glColor3f(1,0.5,0.5); gprint (10,20,"Press 'z' to toggle depth test.\n"); for(i=40;i<250;i++) gprint (10,i,"XX"); glFlush(); } int main(int argc, char *argv[]) { glutInit(&argc, argv); glutInitDisplayMode(GLUT_RGBA | GLUT_DOUBLE); glutCreateWindow("Bitmap BUG test"); glutDisplayFunc(display); glutReshapeFunc(reshape); glutKeyboardFunc(key); glutMainLoop(); return 0; } ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]DRI Unresolved Symbols
Kyle Super wrote: > > Hi, > I recently downloaded and installed XFree86 version 4.2.0. I was using > kernel version 2.4.2-2 and had to install the latest DRI r128 binary from > http://dri.sourceforge.net/download.phtml. This went fine and I had no > problems starting the xserver. > > I then updated the kernel to version 2.4.9-31 to fix some mtrr problems. > Once again, I installed the latest DRI r128 binary, but now on xserver > startup r128.o does not load (although the xserver does start). This is what > appeared in the dri.log file after I installed r128.o for the new kernel: > > make: `r128.o' is up to date. > /usr/X11R6/lib > /lib/modules/2.4.9-31custom-May02/kernel/drivers/char/drm/r128.o: unresolved > symbol boot_cpu_data_R3gig_0657d037 > /lib/modules/2.4.9-31custom-May02/kernel/drivers/char/drm/r128.o: unresolved > symbol do_mmap_pgoff_R3gig_1b290628 ... > > Does anyone know how to fix this? > > Thanks, > Kyle Super I'm not sure why the DRI module didn't build, but you should be aware that the dri drivers at dri.sourceforge.net are not the same as from XFree86. DRI development has a completely different CVS, which gets merged to XFree86 occasionally. If you just want to get XFree86 working, use the DRI drivers from the same source tree. Look in: xc/programs/Xserver/hw/xfree86/os-support/linux/drm/kernel Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]X-windows input driver from char device
Yick Lam wrote: > > Hi, > > I am developing a driver for the touchpad designed by our team of > engineer. The hardware is originally a com port device and it works > perfectly. Now the hardware is changed to USB and the driver I > devloped is changed to a character device. > > With the com port, somewhere in the driver, I have: > local->fd = xf86OpenSerial(local->option); > > and when there is a data from serial port, the function > xf86ReadInput(...) will be called. > > However, since now I the driver I developed is a character > device, instead of xf86OpenSerial, I use: > local->fd = xf86open(priv->devfilename, ORDWR|O_NDELAY, 0) > and I am expecting the function xf86ReadInput will be called > when there is data avaible. But this does not work. > > Questions: > 1) Why I get no message from when use xf86open > 2) Has anyone written a driver the reads input from pipe >in Linux so that I can test the protocol with my own >input? > > Yick > Hong Kong, China Input is a bit tricky because it has to go through a SIGIO handler that knows about your input device. I don't remember the details, but the USB handler in the Wacom driver should be a useful example for Linux-only USB. We really need to get USB support at the server level, because most new devices will be USB. As for FIFO support, I have also thought about that as a useful debug/test tool. Socket support would be good where two-way communication is needed, and could be used to connect simulated devices, or even connected over a network to a remote device. But, I don't know of anyone implementing either of these. Another useful tool I've thought about is an input module loader that can run/debug input drivers without loading a whole server. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]TAKE2: A few pointered questions
Leon Brooks wrote: > > Here's another question I've not found answered in a fair bit of web > trolling. Is it possible to have two (or more) independent mouse pointers at > large on the one X display at the same time? If so, how does one (at a > portable X application level) set about distinguishing between events from > each? > > In particular, I want to write simultaneous multiplayer games (both > co-operative and competitive, mostly educational) for X (-: and would like to > run an X server that supports this as it makes testing a good deal easier :-). You can always install additional mice as extension devices. The application has to then open the extension device and draw it's own pointer. It takes more effort to open extension devices. Look at the contributed program xinput as an example. Having two core mice would cause a lot of design issues, and is probably not going to happen any time soon. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Minimizing VC switch time
It seems to me that the VT switching problem is due to text consoles not being smart enough to restore a text mode when thay are restarted. Would that be too hard to implement, due to the variety of (sometimes proprietary) video cards? It seems that a console should be able to do it, because the OS has initialized the video card at some point. A related question: are there plans to incorporate a VT-less server option in XFree86? Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]stereo
Stuart Levy wrote: > Has anyone here tried the ATI stereo drivers? We're thinking to > get some of their cards just to be able to use stereo under Linux. I have heard at least one person successfully test the FireGL stereo. However, these cards are quite expensive and you will get stereo much cheaper by buying the Xig server for $99. It supports stereo on almost any supported card, which excludes nVidia due to closed hardware specs. Xig test drivers can be downloaded free. I've tested stereo on ATI and MGA cards, with great success. There is also an incomplete open source MGA driver patch for stereo. So, MGA will likely be the first full stereo support for XFree86. The last option is emulated stereo, using interlaced or above/below formats, for example. I've been working on a wrapper library to do this. I will never be as good as real stereo, but it isn't to hard to get full screen, full resolution stereo using a sync doubler and a double height screen mode. This will actually work pretty well once XFree86 can do runtime resizing. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Direct 2D Access for Streaming Video
Shane Walton wrote: > > Hello, > > Could someone point me in the right direction to the > ability to DMA a video frame from shared memory to the > graphics device? Is there a way to use OpenGL along > with DRI? There is just too much overhead involved > to allow the Xserver to handle a CPU based memory > transfer. Any thoughts/comments are appreciated. > Thank you much. > > Regards, > > = > Shane M. Walton, Software Engineer > Digital System Resources, Inc. > [EMAIL PROTECTED] : [EMAIL PROTECTED] > 703.234.1674 I suspect the best approach would be the XVideo extension. Look, for examplem, at XvPutVideo. If you want to use OpenGL, glDrawPixels will copy a source image to the screen, hopefully optimized in hardware. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]renaming ati drivers
Arkadiusz Miskiewicz wrote: > > Hi, > > I want to include GATOS driver in my XFree86*.rpm and I need > to rename XFree86 ati* drivers to ati_old, r128_old, > radeon_old. > Sounds like th GATOS drivers should be using a different name. You could always make the install easier by just setting the module path in the XF86Config to a place where GATOS drivers live, with symlinks to the rest. Then it all fits in one package with no renaming. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Writting a Driver for LCD /Touch Screen
> Latha wrote: > > Hi, > I want to write an LCD / Touch Screen Driver for Xfree86. > I thought the best approach would be to see a Mouse driver of Xfree86. > But i fail to get the flow of the Xfree86 architecture. It is very disorganized at this time. There are no documents on it's current state, and it still has not stabilized from major 4.0 changes. Do not follow the mouse or keyboard. They are very stable, but special cases and complex. > > I am not able to get the entry points and the interface between the X and >linux-kernel. The only interface to the kernel is opening a serial I/O port. If it is USB, you will need some updates to the server code first. > > Can anyone tell where do i start about with this... > Start from the current CVS code, and find another touchscreen driver with the most recent source code changes -- it will be most likely to be up to date. Not all drivers are. After hacking on this for a bit, send me some Email for further advice. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]A better X mouse cursor acceleration?
Ross Vandegrift wrote: > > [snip fraction accel] > > After playing with this for a while, it feels to me that: > > A) The acceleration isn't noticable with low (normal) speeds > > B) Once you reach a high enough speed, it gets fast too quickly > > > > This sounds like a signature of that polynomial curve to me. > > I played with this for a bit when I read that mail as well. I think the > original poster indicated the acceleration was exponential. I agree it > go too fast very quickly. I'd think about trying some small polynomial > curves and perhaps a logarithmic (though that might be too slow). It sounds like you need in A to increase the overall sensitivity, using the XF86Config Resolution option, or maybe the Sample Rate. In B, the curve is probably just too sharp, or maybe it just needs a limit on the maximum accel rate. Windows acceleration is simpler; there is a sensitivity, then a 2X threshold, then 4X threshold. No fancy accel curve. The main difference from the old X accel method is that only counts above the threshold are scaled up. The one thing that is missing in all equations is time. Everything is count based, which means the values are sample-rate dependent, and you will get big jumps if events are delayed for some reason, like system load. Maybe we can use the current exponential, but limit it to 4X, and normalize for time to a standard mouse sample rate, which I think is 40Hz. Sound good? Joe ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
[Xpert]Relative XInput events broken
X has the ability for a program to get relative or absolute events from a valuator. XFree86 has decided that ALL events are absolute. Devices that send relative data are pre-integrated before that data is stored in the ValuatorClass struct and before an app can see it. This is obviously wrong. Maybe a misunderstanding as the relative flag being a device type, rather than the type of events we want to receive. The pointer coordinates are stored independent of device valuator data. A device sending relative events should have relative events in it's history, even if it is the core pointer. I'm guessing that was how miPointerDeltaCursor used to work. Is there a good reason why miPointerDeltaCursor was eliminated? More broken XInput stuff... Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]A better X mouse cursor acceleration?
The current acceleration is as follows. There are three values, num, denom, and threshold. You can set them all with xset, and don't need XChangePointerControl. I don't get noticeable acceleration with "xset m default". Here's the current XFree86 acceleration: If threshold > 0, the old way is used, which isn't very good. If threshold == 0, the distance moved is multiplied by: (d**(num/denum-1))/2 Install the exponential acceleration with, for example: xset m 18/10 0 num/denom values of about 1.5 to 2 seem to work good. Don't go much higher or the pointer is unstable. I guess this should go into the XF mouse manpage. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]A better X mouse cursor acceleration?
It actually wouldn't be too hard to do mouse acceleration outside of X. Just build it into gpm, and have X read fake pre-accelerated mouse data out of gpm's repeater fifo. Putting it directly into the mouse driver isn't that hard either. However, it will be a bit harder to add run-time adjustment of the acceleration parameters, because XInput doesn't do well at communicating arbitrary information between clients and XI drivers. Since I've been digging through XInput in my spare time lately, I'll look into updating accel support. Meanwhile, it wouldn't hurt to add accel support to gpm. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Help to compile dri kernel module
Steve Kieu wrote: > > Hi, > > I would like to compile the dri kernel module for > XFree86-4.2.0, the kernel 2.4.19-pre2-ac4 doesn'nt > have and XFree86-4.2.0 complain about dri modules and > 3D accel is disabled. So I go to the > dri.sourceforge.net, download the file > linux-drm-4.2.0-kernelsource.tar.gz > extract and make -f Makefile.linux. The error is as > below: > cc -O2 -O2 -mcpu=i686 -march=i686 > -fno-strength-reduce -Wall -Wwrite-strings > -Wpointer-arith -Wcast-align -Wstrict-prototypes > -Wnested-externs -Wpointer-arith -D__KERNEL__ -DMODULE > -fomit-frame-pointer -DCONFIG_AGP -DCONFIG_AGP_MODULE > -DCONFIG_DRM_SIS > -I/lib/modules/2.4.19-pre2-ac4/build/include -c > i810_dma.c -o i810_dma.o > i810_dma.c: In function `i810_free_page': > i810_dma.c:298: structure has no member named `wait' > make: *** [i810_dma.o] Error 1 > sk@localhost:~/src/drm$ > > My cards is intel810. Linux Mandrake 8.1 gcc > gcc-2.96-0.76mdk > > What should I do? > > = > Steve Kieu The DRI source is on a different development tree that is faster moving and less stable than the source at XFree86, which synchronizes with DRI source intermittantly. This is not obvious from web pages or docs of either XFree86 or DRI, and had me confused for a while too. So, DRI source will have more problems and more features. To get the XFree86 version of DRI, get the XFree86 source and got to the following directory: xc/programs/Xserver/hw/xfree86/os-support/linux/drm/kernel That's where the 4.2 kernel stuff is for Linux. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Monitor troubles with nVidia
Mark Vojkovich wrote: > > On Wed, 13 Mar 2002, David E. Young wrote: > > > Greetings. First, the hardware particulars; Athlon 1400, 512mb, nVidia TNT2 > > 16mb, ViewSonic A90f, Redhat 7.2, XFree86 4.x (whatever shipped with Redhat > > 7.2). > > > > Ok. The A90f is brand new and the second one I've tried in the last four > > weeks. First monitor: after a few days, whenever the monitor went into > > power-saving mode it exhibited an audible "whine" from the rear area. This > > while running Windows 2000. I was ready to install Redhat anyhow, so did > > that. Same problem. Also, the "whine" was louder during OS boot while Linux > > was in text mode (640x400). The "whine" would disappear after X started but > > resume whenever the monitor entered standby, off, etc. Contacted ViewSonic; > > they said "bad monitor". Returned it last week to Office Depot (nice people) > > and received another A90f. Started it up, no whine. Cool. Two days later I > > was working along and POP! from the monitor, screen goes blank for a few > > seconds, then image returns fine. Uh oh. Few days gone by and monitor > > working fine. Shut system down for two days or so, rebooted and monitor > > exhibited very loud whine during Linux boot that disappeared after X > > started. Sigh. > > > > WTF, O? The same hardware drove a ViewSonic 17PS for many months and that > > 17PS is six years old. I am *not* overclocking anything, don't fiddle with > > xvidtune or anything like that. And, I've tried the native nVidia drivers > > too. > > This really sounds like a problem with the monitor. > > http://www.google.com/search?hl=en&q=+Viewsonic+%22high+pitched%22 > >There are some suggestions about plugging a monitor into a > different power outlet. Not sure I'd expect that to help though. > > Mark. Definitely bad hardware. It's the high voltage flyback transformer. The frequency is related to the scanline rate. Lower scan rates will come closer to the audible spectrum and be louder. The only thing you can do is try different video modes. That big pop is also from the high voltage transformer. Arcing usually only gets worse. Be prepared to exercise the warranty again Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]unloading a module while X is running
Steve Geinitz wrote: > > hi, i am using a microtouch touch screen with X 4.0.1 > so in my XF86Config-4 under modules I have an entry: > Load "mutouch" and the other necessarystuff to use the > touchscreen, it works great. the problem is that i have written a > gui interface for a touchscreen calibration utility that runs > from X. When i don't load anytouchscreen stuff for X i can't use > the touchscreen but the calibration utility works(since it communicates > directly with the touchscreen controller), herein lies the problem > i want to use the touchscreen before and after the calibration > without restarting X. is there anyway to unload "mutouch"? > i would also really like to accomplish this task because MetroX > can do it. tia. > Currently there is no way to do this in XFree86. You can't even tell X to close the device, and it doesn't even close the port if you switch to another VT. Your only choice is to restart X. You could modify the driver a bit and get it to close the port on DeviceControl(DEVICE_OFF). Then, it will switch of when switching VTs, and you could run the calibration tool in a second server on another VT. Not very nice, but it's the only way to make it work without changing the server. I have recently been digging through the XInput code. I am working on support for run-time device opening (an option of course), where devices are only opened when a client uses them, and X doesn't hang when an absent device is opened. This will fix your problem even without a dynamic loader. I am also updating the sample driver and writing some docs so the next person to attmpt writing a driver won't be as lost as I was, trying to figure it out in the middle of a series of changes from 4.0 to 4.2. I also am planning to add support for the run-time TScalibration device control as supported in MetroX, as well as adding an additional crontrol type for sending generic commands to your driver, so you aren't stuck hacking Feedbacks, etc. Last, and maybe most importantly, MetroX has expressed willingness to donate thier runtime module loader, so I may be able to work that in, but that has further reaching implications than just XInput, and I don't know if other people like the idea yet, so that my take some time. I welcome any feedback. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Radeon driver maintainer
Scott Henson wrote: > > I was wondering who the maintainer for the radeon driver under X is. I > am trying to get my Radeon VE to work and it is really difficult because > from what I have heard the ATI driver is crap. If that isnt true please > forgive me, but I would really like to talk to who ever is maintaining > this driver. Anyone want to help me. Thankyou > -- > -Peace kid > Scott Henson [EMAIL PROTECTED] > You probably won't find out much with a post like that. It sound like you just want to complain about the driver. It is better to post your problem and see if someone can help. With a newer driver like Radeon, you may be better off to download and compile the CVS version. I think all of the XFree86 drivers have some degree of crap in them compared to things like the FreeBSD or Linux kernels. A lot has changed recently in graphics hardware, and new features like DRI and Xvideo extension have been added. The drivers are very complex, require a lot of work, and there's very little funding. Things are stabilizing, but you should be patient. If you want a more stable server, you can always pay money and buy one from MetroX or Xig, which are quite reasonable. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Nvidia driver and nv bug ?
Guillaume Membré wrote: > > Hello, > > I have bought a geforce2 TI, and I installed the nvidia driver 2314, it > worked well with opengl games but when I look tv with my bt878 card, my > computer crashed :( (I send a bug report to nvidia) > Before I had a 3dfx3 2000 agp and it worked fine. no problem until I > changed the graphic card. > > So I decided to use the nv drivers included in the Xfree 4.1.0, it > worked perfectly with my tv. no crash. > And when I want to play with quake3, I have so to restart my X with the > nvidia drivers it crashed when the screen switched between the console > to the windows mode. > > I try differents things : > *first to lunch 2 X server in the same times with different config file > but freeze > * so I decided to lunch just one : when I lunch X with nv driver, kill > it, and start X with nvidia it crashes. I have to reboot. I try to start > X with nvidia module (it works), kill it and start X with nv it workes > again, kill it, but when I restart X with nvidia module : crash > > I have no log of what happen when I crashed because I have to reboot > manually the computer :( > > Anyone has an idea ? It's a bug ? > > Thanks a lot > > - Guillaume There's two problems here. First, XFree86 and Nvidia use different kernel modules to get direct hardware access. Loading both kernel modules at the same time sounds like a good way to crash the kernel. You will definitely not be able to run both at the same. Second, they use different and incompatible libGL files. The OpenGL isn't even compatible when doing a remote display to a different server (or at least it used to be...) To switch between server versions is complicated. You can make a script to unload the other kernel modules, swap libGL files (i.e. by redirecting some symlinks), modprobe the new kernel modules, and start the server with the new driver. It is a big hassle, and it isn't likely to change because Nvidia does things very different from XFree86. Joe Krahn Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]XF86Config-4: How to set up for CRT/LCD on an notebook?
Michel Dänzer wrote: > > On Mon, 2002-02-25 at 15:12, Olaf Stetzer wrote: > > > > I have some troubles in setting up my notebook for both > > CRT and LCD output. I tried to find some documentation > > about this issue but did not find anything! The situation > > is as follows: > > > > Both outputs work and I can even switch between LCD, CRT and > > LCD/CRT-mode with Fn-F6 but I don't know how to set up > > the different Screens so that I get a decent Mode for the > > monitor output. I have tried to set up two Monitor and > > Screen Sections and even two ServerLayout Sections for the > > two outputs but it seems that XFree is alway using the > > second of the two ServerLayouts which is the LCD one. > > You can choose a specific ServerLayout with the -layout switch. > But you cannot switch server layouts while it is running. There is also no way to change resolutions while X is running without keeping the virtual screen size the same, so you can't switch to the LCD resolution and still have the whole workspace visible. The ability to resize the screen is being worked on, though. If you want two different resolutions of the same screen at the same time, I think this can only happen if the video card can internally scale the VGA-out to fit the LCD screen. But, I don't know about laptop vidoe hardware. VGA projector manufacturers recommend that you start your laptop after it is plugged in to the projector, and not try to use the LCD display because a lot of laptops can't put out high-resolution while the LCD display is on, no matter what OS you use. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]XFRee86 cvs
Jonnii Henttinen wrote: > > Hello again! > How about building xfree86 cvs under debian linux? Am i able to install > under different location, something like /usr/local than /usr? That way i > should be able to ensure i wouldn't blow up debians package managing system. > --jonni > You can set these in xc/config/cf/site.def: #define ProjectRoot /usr/local/X11R6 #define NothingOutsideProjectRoot YES #define EtcX11Directory ProjectRoot/etc This will keep it out of your package managed dirs. But, there are still complications. Before running X there, you need to make sure your path goes to /usr/local/X11R6 first, and libraries need to come from there, like using LD_LIBRARY_PATH. Here's another method I use: Rename /usr/X11R6 to /usr/X11R6-4.1, then use XFree's lndir to make a symlink clone directory /usr/X11R6-4.2. Configure your new X to install there. Now, you can symlink either one to /usr/X11R6 (and maybe run ldconfig) depending on if you want to run in developer or stable mode. Hmm.. maybe one could set up run level 4 as X-cvs mode... Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]2 Xservers
Elizabeth Ebbert wrote: > > Joe Krahn wrote: > > > Elizabeth Ebbert wrote: > > > > > > I would like to know if it is possible to run two xservers on the > same > > > VT and how to set it up? I have installed XFree86 4.2.0 on > slackware > > > 8.0 platform with a Radeon VE duel head video card. > > > > > Maybe what you really want is two screens from one Xserver. > > That is the normal dual head setup. I think that is working > > for the Radeon now. > > > > Joe Krahn > > ___ > > Xpert mailing list > > [EMAIL PROTECTED] > > http://XFree86.Org/mailman/listinfo/xpert > > No, I already have the one Xserver on dual monitor running, what I am > trying to do is have two Xservers that share the keyboard but have > individual mouse / touchscreen so that two people can work without > stealing > the mouse. I was hoping with the advancement with XFree86 that it was > now > possible. It has to be on the same VT because we are displaying > touchscreen menus that are running on another computer. Any > help/direction > would be appreciated. > What you need is for one sever to run with no VT. Normally, the idea is that the 2nd one gets it's own keyboard and mouse. Xig worked on this for their server, but gave up on it, at least for now, because it can be a bit unstable. Someone hacked a version of XFree86 to do this with a USB mouse/keyboard, but I don't know how stable it is. This could be quite useful for a Kiosk or classroom setup. I think it would be a good thing. As for sharing a keyboard, how do the servers know who wants to be typing? A keyboard switch plus the VT-less server would fix that, if you can get a VT-less servwer running. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]verification on XChangeDeviceControl
I looked into the XFree86 server code, DEVICE_RESOLUTION is the only one. xDeviceTSCalibrationCtl came from MetroX. It looks like only MetroX servers support it. The only hack available is to encapsulate commands in DEVICE_RESOLUTION packets. You can send a series of bytes that are assembled into a 0 terminated string. The bad part is that a get resolution command uses saved values rather than call the device, so you cannot use it to get data back from the device. Joe Krahn "Prasad, Ragini" wrote: > > Thanks for replying. > > xDeviceTSCalibrationCtl seems to match my needs at present pretty well. But > I am not clear , what function to use from my client, as well as what > control type to specify for this undocumented feature. Cause > XChangeDeviceControl does not let me specify anything other than > DEVICE_RESOLUTION. What would I need to use if not XChangeDeviceControl ? Or > how else could i use this undocumented feature. > > I completely agree with you that it gets extremely tough to work with > incomplete implementation of feedback controls. A DEVICE_PRIVATE would take > care of most devices, giving it a private control, and choice to define its > interaction. > > Thanks, > Ragini. > > -Original Message- > From: Joe Krahn [mailto:[EMAIL PROTECTED]] > Sent: Friday, February 22, 2002 8:44 PM > To: Prasad, Ragini > Cc: '[EMAIL PROTECTED]'; '[EMAIL PROTECTED]' > Subject: Re: [Xpert]verification on XChangeDeviceControl > > "Prasad, Ragini" wrote: > > > > Hi, > > > > Is it possible to use XChangeDeviceControl for control type other than > > DEVICE_RESOLUTION. I am writing a calibration tool for touchscreen. > > > > I have added control procedure, for control_proc, to set my touchscreen > > parameters in my X extension driver. > > > > For this control_proc to get invoked and set the params, I need to use > > XChangeDeviceControl in my calibration program. But what should I set my > > control type in XChangeDeviceControl ? > > > > I also see xDeviceTSCalibrationCtl structure defined in the XIproto.h. > This > > gives me a undertanding that there is a way of using XChangeDeviceControl > > function for control type other than resolution. Or is there some other > > function which can also set the control parameters? > > > > Thanks for the help. > > Ragini. > > No, there is not! XChangeDeviceControl is horribly incomplete. It is > something I would very much like to see changed. I can't believe it > has existed so long in an unfinished state. > > (Oh, and I hope you didn't use the XFree86 sample driver...) > > Typically, device drivers resort to an ugly hack, and misuse > FeedBack controls to send data. For example, there is a Spaceball > driver that uses the BellFeedback something like this: > If the percent and duration are zero, then the pitch is a byte of > data to send to the Spaceball. > > What's worse for XFree86 is that the server does not export many > of the defined feedbacks yet, so it's hard to implement a good hack. > Also be warned that SetValuator doesn't work in XFree86. > I worked out a decent hack using DEVICE_RESOLUTION to encapsulate > arbitrary command data. But I'd really like to see XInput get finished. > > xDeviceTSCalibrationCtl is in fact a touch-screen calibration, > but it was designed with one device in mind, so I don't know > how well it will work for you. Also, it's undocumented, of course, > and not part of the X11 standard. But it may work for you if > you just want to get the driver working. > > SGI has implemented (a long time ago) the XSGIDeviceControl command. > It sends two arbitrary command strings, a command name and a command > value. Char strings are great for generic device commands because > there's no endianness problems to worry about in the packet > transport. > > Here are some proposed control types I would like to see: > DEVICE_RESET - guess what this does. > DEVICE_RANGE - valuator limits, similar to DEVICE_RESOLUTION > DEVICE_THRESHOLD - A noise threshold before sending data, > like the null radius in a joystick/Spaceball > DEVICE_ZERO Can be used to auto-rezero, or to manually > set the zero value. > DEVICE_PRIVATE - A device-specific command, sort of like > XSGIDevice control. It could consist of a > private command type as an int (or possibly > as a string) followed by a string of command > dependent data. > > The DEVICE_PRIVATE command is the most essential, because it > could support all of the others while waiting f
Re: [Xpert]lilo config
[EMAIL PROTECTED] wrote: > > In suse 7.2 "lilo.conf" file there is a setting "vga=771". Is there a > document that descibes 771 and other video modes? thanks. > ___ > Xpert mailing list > [EMAIL PROTECTED] > http://XFree86.Org/mailman/listinfo/xpert Look in the Linux source documentation: linux/Documentation/fb The vga= modes can select regular text modes or frambuffer modes, where text is rendered as graphics, for slower but nicer text. I think the fb docs list both the text and framebuffer modes. ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Cyberblade not accelerated
Tony Mobily wrote: > > [By accident, I sent this message to " [EMAIL PROTECTED]" as well... I > hope no-one will be cross at me!] > > Hi everyone, > > I have just bought a laptop to replace my desktop computer. > It's a Toshiba Satellite 1800, and has a CyberBlade video card (X says > "Chipset cyberbladeXPAi1 found". > > With the option "ShadowFB" things get a bit better, but not that much > better. > > Do you have any idea? Does anyone have the same problem? > > Any help is very much appreciated - I spent all my money on this laptop > and... :-) > There is some new accelerated support in 4.2.0 that may work for you, but laptops are always the last to become stable because the hardware has been tweaked for each specific laptop design. Even if it works, there's no DRI support, so 3D stuff will be slow, sometimes unuseably slow. In general, if you want to run Linux, you should always check for hardware support before buying. SOMETHING TO CONSIDER: You just bought a laptop with Windows installed, because Microsoft has made it almost illegal to sell a computer without WIndows, right? SO, Microsoft takes there statistics of what percent of users run Windows, and they put you in that list, even though you have erased WIndows, all as part of the evil FUD system to get the few Linux supporters to decide it's a waste of time. Am I wrong? Is there a way to unregister a computer as running Windoze? Or are the statistics I've read correct in saying that 95% of all users run Windoze? Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]2 Xservers
Elizabeth Ebbert wrote: > > I would like to know if it is possible to run two xservers on the same > VT and how to set it up? I have installed XFree86 4.2.0 on slackware > 8.0 platform with a Radeon VE duel head video card. > Maybe what you really want is two screens from one Xserver. That is the normal dual head setup. I think that is working for the Radeon now. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]verification on XChangeDeviceControl
"Prasad, Ragini" wrote: > > Hi, > > Is it possible to use XChangeDeviceControl for control type other than > DEVICE_RESOLUTION. I am writing a calibration tool for touchscreen. > > I have added control procedure, for control_proc, to set my touchscreen > parameters in my X extension driver. > > For this control_proc to get invoked and set the params, I need to use > XChangeDeviceControl in my calibration program. But what should I set my > control type in XChangeDeviceControl ? > > I also see xDeviceTSCalibrationCtl structure defined in the XIproto.h. This > gives me a undertanding that there is a way of using XChangeDeviceControl > function for control type other than resolution. Or is there some other > function which can also set the control parameters? > > Thanks for the help. > Ragini. No, there is not! XChangeDeviceControl is horribly incomplete. It is something I would very much like to see changed. I can't believe it has existed so long in an unfinished state. (Oh, and I hope you didn't use the XFree86 sample driver...) Typically, device drivers resort to an ugly hack, and misuse FeedBack controls to send data. For example, there is a Spaceball driver that uses the BellFeedback something like this: If the percent and duration are zero, then the pitch is a byte of data to send to the Spaceball. What's worse for XFree86 is that the server does not export many of the defined feedbacks yet, so it's hard to implement a good hack. Also be warned that SetValuator doesn't work in XFree86. I worked out a decent hack using DEVICE_RESOLUTION to encapsulate arbitrary command data. But I'd really like to see XInput get finished. xDeviceTSCalibrationCtl is in fact a touch-screen calibration, but it was designed with one device in mind, so I don't know how well it will work for you. Also, it's undocumented, of course, and not part of the X11 standard. But it may work for you if you just want to get the driver working. SGI has implemented (a long time ago) the XSGIDeviceControl command. It sends two arbitrary command strings, a command name and a command value. Char strings are great for generic device commands because there's no endianness problems to worry about in the packet transport. Here are some proposed control types I would like to see: DEVICE_RESET - guess what this does. DEVICE_RANGE - valuator limits, similar to DEVICE_RESOLUTION DEVICE_THRESHOLD - A noise threshold before sending data, like the null radius in a joystick/Spaceball DEVICE_ZERO Can be used to auto-rezero, or to manually set the zero value. DEVICE_PRIVATE - A device-specific command, sort of like XSGIDevice control. It could consist of a private command type as an int (or possibly as a string) followed by a string of command dependent data. The DEVICE_PRIVATE command is the most essential, because it could support all of the others while waiting for the 'standard' controls to be finalized. Right now, the whole XFree86 XInput system is in a rather unfinished state, ever since the 4.x transition. My plan is to update the sample driver, and have it include the entire set of feedbacks, and write some guidelines for other device-driver writers. After that, we can add on a DEVICE_PRIVATE control. There may be some politic int the process of adding to the X11 protocol, but this is badly needed IMHO. What do you think? Any suggestions? Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]A big confusion over here - XFree 4.2 - and rage 128
Michel Dänzer wrote: > > On Wed, 2002-02-20 at 00:17, Gustavo Homem wrote: > > > > I am having serious trouble with XFree. I have a rage 128 video card. I was > > using XFree 4.1 from mdk 8.1 and the Xv extension was malfunctioning for DVDs > > (was ok for other movies). It would eat too much CPU so a lot of frames were > > dropped (tested with xine, vlc and mplayer), for example xine would take > > around 45 % and the X server about 45% using Xv output. > > > > I upgraded to XFree 4.2 rpms from mandrake in hope to have this problem > > solved since it was working perfectly on XFree 4.03 from mdk 8.0. > > I can only guess why 4.0.3 worked better than 4.1.0; I guess it's > because the r128 driver in 4.0.3 didn't properly support 420 video (or > was it 422? I forget :) yet. > > > X complained then about drm modules beiing old, so I downloaded the latest > > one from dri.sf.net , compiled and instaled the module at /lib/modules/ > > > > The result was as follows with dri enabled: > > > > 1 - Now bttv no longer freezes the system if used with dri enabled (hooray) > > 2 - The KDM login screen is completely messed (widgets are garbaged, not > > drawn properly) > > I have fixed this in DRI CVS and submitted the fix to XFree86. > > Both 1) and 2) are due to the fact that the r128 driver does 2D > acceleration differently with DRI enabled now. I'm sure you've also > noticed the better 2D performance. :) > > > 3 - This is the weird part: Xv is messed, the colors don't match the real > > ones, they are completely changed and there is an horizontal strip with a > > diferent color variation with xine and a lot of garbage with mplayer, X > > freezes a short time after switching to and from fullscreen (except for the > > mouse). Strangely during the time before X freezes there are no framedrops > > but they come back with dri disabled. > > > > What is exactly the relation of Xvideo and dri ? > > They are conceptually independent, however as of 4.2.0, the r128 driver > tries to take advantage of the DRM to accelerate Xv image transfer. > Unfortunately, this doesn't work well with some chips, and there are > also problems coordinating this with the new 2D acceleration yet. > > Until these issues are sorted out, use of the DRM for Xv should be an > option which defaults to off. I'll submit a patch to that effect soon. > Attached is a hack which disables it completely in source for now. I > hope this is useful for people who are bitten by this. > > -- > Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer > XFree86 and DRI project member / CS student, Free Software enthusiast > ___ > Xpert mailing list > [EMAIL PROTECTED] > http://XFree86.Org/mailman/listinfo/xpert Michel, I don't see an attachment. Did you only send it to Gustavo? I would like to try the patch/hack to disable DRM for Xv. Updating to the latest CVS fixed some of my problems. I just tried xine, and got a system lockup. Mplayer worked, but colors were wrong. It looks like a lot of R128 users will have to avoid using a vanilla XFree86-4.2. That's unfortunate because xawtv with DRI in XF4.1 locks up the system. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Blanking and Unblanking screen
Andy Isaacson wrote: > > On Tue, Feb 19, 2002 at 09:07:03AM -0800, Jonathan Marks wrote: > > On Sun, 2002-02-17 at 15:53, Mark Vojkovich wrote: > > > On 17 Feb 2002, Jonathan Marks wrote: > > > > I have an app that displays images and videos in fullscreen. Often when > > > > killing the one program and starting the next there is a flicker and a > > > > glimpse of the desktop before the next image/video starts. > > > > > > > > I wonder if there is a way to blank the screen, for the period that I > > > > kill the one program and start the other, and then unblank the screen. > > > > > > See the man page on XSetScreenSaver and friends. One of those might > > > do what you want. > > > > Thanks for the prompt response. This application is a bunch of shell > > and perl scripts. I was wondering if there is a way to do this without > > resorting to writing C code. Is there a way to blank and unblank the > > whole screen in X from the command line? > > Is it really going to be better for you for the screen to go black > between images than for it to flash to the desktop? > > If this is really what you want, you can do an "xset s activate" before > killing off the first image-viewer, then "xset s reset" to turn the > screen back on. > > Assuming this is a slideshow of some kind, you might want to do > "xset s off; xset -dpms" beforehand, to prevent the screen going black > on you. > > -andy > ___ > Xpert mailing list > [EMAIL PROTECTED] > http://XFree86.Org/mailman/listinfo/xpert I think what you need is not to black the screen, but to have a black backdrop, that is always there to catch moments of no image. That way, you minimize the amount of blank time. For example, this will work: display -title backdrop -geometry 1024x768+0+0\! XC:black If you set the window manager not to decorate the window, by matching the title. ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]A big confusion over here - XFree 4.2 - and rage 128
I'm using XF4.2+Rage128 with problems as well. I also am happy to see xawtv+DRI work! I think I had some Xv problems before switching to the current CVS, but I had bigger problems. Text video consoles became inaccessible frequently, and occasional 2D acceleration bugs, usually when moving a long scrollbar. The 2D bugs used to lockup X (KB+mouse dead). I also got random hangs when closing a window, but then CPU use goes to 100% and pointer is alive, so it's a different problem. I did another CVS update two days ago and recompiled. There are now no more lockups, but VT text terminals still die frequently, and 2D drawing bugs still happen. Maybe this is related to your drawing bugs. DO you see this in your logs when DRI is enabled? [drm:r128_cce_indirect] *ERROR* process 2865 using buffer owned by 0 Do you ever lose the VT text consoles? Joe Krahn Gustavo Homem wrote: > > Sorry, the first time I sent this message I wasn't subscribed so I guess it > didn't arrive. > > Thanks > Gustavo > > --- > Hello everyone: > > I am having serious trouble with XFree. I have a rage 128 video card. I was > using XFree 4.1 from mdk 8.1 and the Xv extension was malfunctioning for DVDs > (was ok for other movies). It would eat too much CPU so a lot of frames were > dropped (tested with xine, vlc and mplayer), for example xine would take > around 45 % and the X server about 45% using Xv output. > > I upgraded to XFree 4.2 rpms from mandrake in hope to have this problem > solved since it was working perfectly on XFree 4.03 from mdk 8.0. > > X complained then about drm modules beiing old, so I downloaded the latest > one from dri.sf.net , compiled and instaled the module at /lib/modules/ > > The result was as follows with dri enabled: > > 1 - Now bttv no longer freezes the system if used with dri enabled (hooray) > 2 - The KDM login screen is completely messed (widgets are garbaged, not > drawn properly) > 3 - This is the weird part: Xv is messed, the colors don't match the real > ones, they are completely changed and there is an horizontal strip with a > diferent color variation with xine and a lot of garbage with mplayer, X > freezes a short time after switching to and from fullscreen (except for the > mouse). Strangely during the time before X freezes there are no framedrops > but they come back with dri disabled. > > What is exactly the relation of Xvideo and dri ? I never needed to have dri > enabled to have smooth XVideo before (it wasn't even available at first), my > computer has played DVDs with Xvideo since the first xine versions with gatos > drivers. Did this change with XFree 4.1 ? I also don't know what is causing > the 1,2,3 problems and I'd be glad if someone could help. > > Thanks a lot > Gustavo Homem ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Radeon 8500 3D support
Thomas Bruckschlegel wrote: > > > Date: Mon, 18 Feb 2002 11:25:16 +0100 > > From: "Anders Saaby" <[EMAIL PROTECTED]> > > To: <[EMAIL PROTECTED]> > > Subject: [Xpert]Radeon 8500 3D support > > > > Hi! > > > > Have anybody any idea about when we could expect OpenGL support for the > > ATI Radeon 8500 card in Xfree? Are anybody working on it? > > > > Secondly: if 3d support is too far avay in the future, then does anybody > > know if xigraphics Summit driver is a useable alternative? > > > > Thanks in advance! > > > > Anders Saaby > I did some benchmarks on > http://www.tommti-systems.de/main-Dateien/news010202_150202.html > They support OpenGL 1.2 and I had some crashes during my tests. $129 isn't > also that cheat > I couldn't find the benchmarks there. It must be somehwere in arkiv. Was this with their newest update, which fixed several bugs? I tested their driver with a Radeon SDR and an MGA-550. I found some bugs (especially with stereo) which are fixed in the newest version. My impression was that it has many fewer bugs than XFree86, but also has fewer features available, but I didn't do thorough testing, judst tried a few programs that I am interested in using. Xig will give you updates, but if you buy a different video card, you'll have to buy another server. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Desktop selection problem
"Rafeek, P (MED)" wrote: > > Hi, > How can we bring up a window (from some other server) in a workspace > other than the active workspace? The "workspace" is a window manager creation, not part of X, so their is no way to tell X which desktop to use. However, the window manager might have a "matched windows" feature, so that it can be configured to recognize an incoming window and place it on another desktop. For example, you could set it so that all windows where the title starts with "[2]" would go to desktop 2. Once the window is placed, you can go back and remove the title prefix. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Diabling Ctrl-Alt-F1 Keysequences Under X
To lock VT switching, try the following simple code. This is nice because root can re-enable switching to get a terminal if needed. Compile the code below. Run it in the first console with: vtlock lock /dev/tty Next, enter startx -- vt1 You now have X on vt1, non-switchable. For X on the usual vt7, lock it with: vtlock lock /dev/tty7 /* vtlock.c -- A simple virtual termnial [un]locker */ #include #include #include int main(int argc, char *argv[]){ int fd,lock; if (argc<2){ fprintf(stderr,"Usage: %s lock|unlock \n" "\tlock or unlock the given tty.\n", argv[0]); exit(-1); } if (!(fd=open(argv[2],O_RDONLY))) { fprintf(stderr,"Error: unable to open tty %s\n",argv[2]); exit(-1); } lock=!strcmp(argv[1],"lock"); printf("Ioctl VT_%sLOCKSWITCH returned %d\n",lock?"":"UN", ioctl(fd,lock?VT_LOCKSWITCH:VT_UNLOCKSWITCH)); return(0); } ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]Diabling Ctrl-Alt-F1 Keysequences Under X
Andy Sparrow wrote: > > > Maybe this could be an alternate solution, can i force X to run on tty1? So > > if a Ctrl-Alt-F1 occured nothing would happen since they are already on 1? > > Hmm, actually, I don't see why you couldn't disable all the gettys & run > [xwgk]dm as a display manager, which will take the first "free" VTY, if none > is specified. > > So, just run a display manager out of a startup script at boottime, without > specifying a VTY for it to run on, and it'll start on the first VTY. > > Which, thinking about it, might be why no-one implemented 'DontSwitch'. > Because it's not necessary - as soon as X is a one-way trip, you may as well > not have a text mode console at all because you have no way of getting back to > it once X is running - so don't run a text console, just start a display > manager instead :) This will result in no available text consoles, but you can still use Crtl-Alt-Fn to switch to a blank, unused VT. He doesn't want that to be available. I think all that is needed is a simple utility to lock the VT. The same code would be used in XFree86, but it doesn't need to be in XFree86 to work: ioctl (fd, VT_LOCKSWITCH, 1); /* lock the switch */ ioctl (fd, VT_UNLOCKSWITCH, 1); /* unlock it again */ Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]4.2.0 R128 hanging
Michel Dänzer wrote: > > On Sam, 2002-02-16 at 16:50, Joe Krahn wrote: > > > > > > > Here's an example backtrace, that seems to me like somthing > > > > deeper than XFree86 code being at fault: > > > > > > > > (gdb) bt > > > > #0 0x40139012 in __writev (fd=20, vector=0xb6fc, count=1) > > > > at ../sysdeps/unix/sysv/linux/writev.c:51 > > > > #1 0x080d5ad4 in _XSERVTransSocketWritev () > > > > #2 0x080d61cc in _XSERVTransWritev () > > > > #3 0x080d0248 in StandardFlushClient () > > > > #4 0x080cea8f in CloseDownConnection () > > > > #5 0x080b2f69 in CloseDownClient () > > > > #6 0x080addbd in Dispatch () > > > > #7 0x080bd6d8 in main () > > > > #8 0x4006e627 in __libc_start_main (main=0x80bd1f0 , argc=1, > > > > ubp_av=0xbc04, init=0x806cc5c <_init>, fini=0x816f94c <_fini>, > > > > rtld_fini=0x4000dcd4 <_dl_fini>, stack_end=0xbbfc) > > > > at ../sysdeps/generic/libc-start.c:129 > > > > > > A backtrace from the server would probably be more interesting, or does > > > only the client hang? > > This is a backtrace from the server, > > D'oh, should have noticed. :) > > > but it's not the XFree86 gdb, so it won't catch module bugs. > > > > I tried installing RedHat's RawHide XFree86-4.2, and the same crash > > occurs. I'll re-install mine, which has debug symbols, and try > > the XFree86 gdb on it. > > Modules aren't even involved here; the hang occurs in a glibc function > so it seems the problem isn't in the X server at all. > > -- > Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer > XFree86 and DRI project member / CS student, Free Software enthusiast I was thinking it looked like a glibc bug. The other thing that happens is that vt switching gets messed up. I can no longer switch to a termnial. I'm stuck at vt 7. So, maybe this is what happens if vt access gets cut off? If so, this might even be kernel bug. But, why is only R128 triggering it? I was hoping somebody else has seen the above backtrace and could give me some suggestions. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
Re: [Xpert]4.2.0 R128 hanging
Michel Dänzer wrote: > > On Sam, 2002-02-16 at 04:04, Joe Krahn wrote: > > I've compiled XF 4.2.0 on RedHat 7.2. I get random hangs, > > sometimes when moving long scrollbars, sometimes when closing > > a window. It's on an Athlon, VIA chipset, using RedHat's > > gcc 2.96. Maybe I should use kgcc instead? > > If I disable DRI and add UseCCEFor2D, it still happens. > > Option "UseCCEFor2D" is no more. With DRI enabled, the CCE is always > used for 2D acceleration. > > > Here's an example backtrace, that seems to me like somthing > > deeper than XFree86 code being at fault: > > > > (gdb) bt > > #0 0x40139012 in __writev (fd=20, vector=0xb6fc, count=1) > > at ../sysdeps/unix/sysv/linux/writev.c:51 > > #1 0x080d5ad4 in _XSERVTransSocketWritev () > > #2 0x080d61cc in _XSERVTransWritev () > > #3 0x080d0248 in StandardFlushClient () > > #4 0x080cea8f in CloseDownConnection () > > #5 0x080b2f69 in CloseDownClient () > > #6 0x080addbd in Dispatch () > > #7 0x080bd6d8 in main () > > #8 0x4006e627 in __libc_start_main (main=0x80bd1f0 , argc=1, > > ubp_av=0xbc04, init=0x806cc5c <_init>, fini=0x816f94c <_fini>, > > rtld_fini=0x4000dcd4 <_dl_fini>, stack_end=0xbbfc) > > at ../sysdeps/generic/libc-start.c:129 > > A backtrace from the server would probably be more interesting, or does > only the client hang? This is a backtrace from the server, but it's not the XFree86 gdb, so it won't catch module bugs. I tried installing RedHat's RawHide XFree86-4.2, and the same crash occurs. I'll re-install mine, which has debug symbols, and try the XFree86 gdb on it. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
[Xpert]4.2.0 R128 hanging
I've compiled XF 4.2.0 on RedHat 7.2. I get random hangs, sometimes when moving long scrollbars, sometimes when closing a window. It's on an Athlon, VIA chipset, using RedHat's gcc 2.96. Maybe I should use kgcc instead? If I disable DRI and add UseCCEFor2D, it still happens. Is anyone else using 4.2, successfully or not, with R128 and RedHat 7.2? Here's an example backtrace, that seems to me like somthing deeper than XFree86 code being at fault: (gdb) bt #0 0x40139012 in __writev (fd=20, vector=0xb6fc, count=1) at ../sysdeps/unix/sysv/linux/writev.c:51 #1 0x080d5ad4 in _XSERVTransSocketWritev () #2 0x080d61cc in _XSERVTransWritev () #3 0x080d0248 in StandardFlushClient () #4 0x080cea8f in CloseDownConnection () #5 0x080b2f69 in CloseDownClient () #6 0x080addbd in Dispatch () #7 0x080bd6d8 in main () #8 0x4006e627 in __libc_start_main (main=0x80bd1f0 , argc=1, ubp_av=0xbc04, init=0x806cc5c <_init>, fini=0x816f94c <_fini>, rtld_fini=0x4000dcd4 <_dl_fini>, stack_end=0xbbfc) at ../sysdeps/generic/libc-start.c:129 I updated from CVS, still hangs. But, there is a glibc update from RedHat which I haven't applied yet. I'll try that with kgcc, and maybe with gcc3. Joe Krahn ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert