Re: discover, read-edid, mdetect selection in base-config
On Tue, Jul 23, 2002 at 10:17:17AM +0200, Steinar Bang wrote: > Are read-edid and mdetect useful for configuring XFree86 4.1.0? > Doesn't "XFree86 -configure" do a better job? You tell me. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=77120&repeatmerged=yes http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=124247&repeatmerged=yes A configuration process that segfaults. M, good! -- G. Branden Robinson| Mob rule isn't any prettier just Debian GNU/Linux | because you call your mob a [EMAIL PROTECTED] | government. http://people.debian.org/~branden/ | pgpIAGpz2kmvR.pgp Description: PGP signature
Re: XF4.2 packages think Alt is Meta
Branden Robinson writes: > 1) I personally don't care, since I remap the funny keys on the bottom > row to suit my own strange tastes anyway; > 2) I don't want to subject my users to shock and horror. > 3) I don't want to deviate gratuitously from upstream practice. > I'd appreciate input from the people on this list as to the impact of > XFree86 4.2 on 2) and 3). Are those goals truly in conflict? Is this > only a problem for people with buggy window managers? No, it will most probably affect all users of Emacs and other clones of Emacs as well, since it makes intensive use of the Meta key (and the default behavior was to have the left Windows key as Meta). You might argue that Emacs addicts probably rebind their keys as they see fit, but for the average or occasional user, the change might be disturbing. Maybe the upgrade process should warn the user about the change, and maybe even ask if it should keep the old behavior. My 2c, -- Romain FRANCOISE <[EMAIL PROTECTED]> | I know, you love the song but it's a miracle -- http://orebokech.com/ | not the singer. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: discover, read-edid, mdetect selection in base-config
> Branden Robinson <[EMAIL PROTECTED]>: > On Tue, Jul 23, 2002 at 10:17:17AM +0200, Steinar Bang wrote: >> Are read-edid and mdetect useful for configuring XFree86 4.1.0? >> Doesn't "XFree86 -configure" do a better job? > You tell me. > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=77120&repeatmerged=yes > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=124247&repeatmerged=yes > A configuration process that segfaults. M, good! I would have tried with "maybe upstream's fixed this?", except 4.2.0 died on me two days ago, because it didn't recognize a GeForce4 card...:-) But when it works, which it has usually done for me, it creates good XF86Config files needing a minimum of tinkering (usually finding vsync and hsync for older monitors). -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
test (sorry)
signature.asc Description: This is a digitally signed message part
Connecting to a Beamer
[debian-x, please CC me as I am not subscribed] Hi folks, I need to connect my laptop (Latitude C610) to a not-so-modern beamer, which supports displaying the 1024x768 screen of the desktop machine sitting next to it. However, when I tell it to display my screen, it simply says: Not available. H: 58.44Khz V: 53.5Hz Ok, granted, I am trying to display 1400x1050, but I get the same message when I tune my resolution to 1024x768 (I defined two screens in XF86Config-4, one for the laptop, one for the beamer). Using `startx -- -layout present` startx the 1024x768 version -- I've never managed to get in-X resolution switching via a key combo to work. Anyway, even at 1024x768 I get the above message, and yet I need to connect to the beamer. I don't know what to do. I'd so appreciate if some X-savvy person would help me out a little, you can get my XF86Config-4 (for Branden's 4.2 packages on testing) here: ftp://ftp.madduck.net/pub/temp/XF86Config-4.gz (1203 octets) I'd appreciate any tips! Thanks, -- martin; (greetings from the heart of the sun.) \ echo mailto: !#^."<*>"|tr "<*> mailto:"; [EMAIL PROTECTED] "you raise the blade, you make the change you rearrange me till i'm sane. you lock the door, and throw away the key, there's someone in my head but it's not me." -- pink floyd, 1972 pgpQndBU1M4dj.pgp Description: PGP signature
Re: A Radeon and X question, probably not Debian specific
"Charl P. Botha" <[EMAIL PROTECTED]> writes: > I had a look at your X output and all seems fine. How many FPS does > glxgears yield? You have to run it for at least 5 seconds before it outputs > something to stdout. glxgears, in its default small size is about 500 fps, expanding the window to full screen gets me 25 fps. Xscreensaver's gears hack gets about 5 fps running full screen. On the advice of Michel Dänzer (thanks!) I added the debian dri-trunk stuff, and got that working easily enough with the 1.4.0 20020714 Radeon drm driver, but it didn't make any difference. Everything appears to be working correctly, but gl stuff is very slow. The only error I get is [drm:radeon_freelist_get] *ERROR* returning NULL! output from the drm driver. As best I can tell, xinerama is disabled. Starting X with +xinerama put the xinerama enabled comment in the logfile, and disables dri, so I used -xinerama to be safe, but that makes no difference. If anybody has more advice, I would be interested in hearing it. Thanks for the help so far. Fortunately my next dual athlon is coming with an $18 8mb ATI Mobility AGP card, which should be more than adequate to run the server room display in text mode. -- Jeff Lessem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Connecting to a Beamer
On Wed, 2002-07-24 at 19:09, martin f krafft wrote: > [debian-x, please CC me as I am not subscribed] > > Hi folks, > > I need to connect my laptop (Latitude C610) to a not-so-modern beamer, > which supports displaying the 1024x768 screen of the desktop machine > sitting next to it. However, when I tell it to display my screen, it > simply says: > > Not available. H: 58.44Khz V: 53.5Hz > > Ok, granted, I am trying to display 1400x1050, but I get the same > message when I tune my resolution to 1024x768 (I defined two screens > in XF86Config-4, one for the laptop, one for the beamer). Using > `startx -- -layout present` startx the 1024x768 version -- I've never > managed to get in-X resolution switching via a key combo to work. > > Anyway, even at 1024x768 I get the above message, and yet I need to > connect to the beamer. I don't know what to do. I'd so appreciate if > some X-savvy person would help me out a little, you can get my > XF86Config-4 (for Branden's 4.2 packages on testing) here: > > ftp://ftp.madduck.net/pub/temp/XF86Config-4.gz (1203 octets) > > I'd appreciate any tips! Try Option "CRTScreen" in the device section. Without that, the external output will use the parameters for the internal panel. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Unicode compose map?
On Tue, Jul 23, 2002 at 06:21:47PM +0200, Juliusz Chroboczek wrote: > The alternative to the compose mechanism -- using the input method > mechanism -- is just too nasty to consider. End of offtopic.) Is that the same as the 'altgr' key? > I can personally confirm that the non-Greek compose file does contain > entries for the ISO 8859-2 characters needed for Polish; it probably > contains all combinations that you need, but I haven't checked myself. Great. Can you either send me the compose files, or point me to the correct cvs directory? I dont want to check out the entire source tree to get two files :) In the mean time, I also stumbled across an updated version of the debian-utf8 howto: http://melkor.dnp.fmph.uniba.sk/~garabik/debian-utf8/howto.html They mention a compose map by Pablo Saratxaga, but I didnt get a chance to try it out yet. > Note that the above applies to current XFree86 CVS; I don't know what > the status of 4.2.0 is. Note furthermore that if you're working with > UTF-8, you really ought to upgrade your X libraries to 4.2.0 or later. I havent noticed any problems so far, but I'm not using it thought the entire X Window system. My general locale is en_US, and I set it to en_US.UTF-8 only if TERM=xterm|screen. This is mainly because I use gnome, which is unable to display any text if I try to use unicode. Would 4.2.0 libs help with that problem? Thanks, Norbert pgpmRRCRavxvS.pgp Description: PGP signature
Re: A Radeon and X question, probably not Debian specific
On Wed, 2002-07-24 at 20:39, Jeff Lessem wrote: > "Charl P. Botha" <[EMAIL PROTECTED]> writes: > > > I had a look at your X output and all seems fine. How many FPS does > > glxgears yield? You have to run it for at least 5 seconds before it outputs > > something to stdout. > > glxgears, in its default small size is about 500 fps, expanding the > window to full screen gets me 25 fps. Xscreensaver's gears hack gets > about 5 fps running full screen. The size of the latter's window doesn't seem to impact the framerate here, so that sounds like software rendering. > On the advice of Michel Dänzer (thanks!) I added the debian dri-trunk > stuff, and got that working easily enough with the 1.4.0 20020714 > Radeon drm driver, but it didn't make any difference. Everything > appears to be working correctly, but gl stuff is very slow. The only > error I get is [drm:radeon_freelist_get] *ERROR* returning NULL! > output from the drm driver. That on the other hand only appears with direct rendering... Does running a client with LIBGL_DEBUG=verbose give any more interesting info as to whether you're actually using direct rendering? -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: A Radeon and X question, probably not Debian specific
In your message of: 24 Jul 2002 23:22:23 +0200, you write: >The size of the latter's window doesn't seem to impact the framerate here, >so that sounds like software rendering. Xscreensaver's gears does about 50 fps at its default size, 25 fps in a maximized window, and 5 fps in full screen saved mode. I wish it was software rendering, because then I would know what problem I have to fix. Everything seems to be working fine, if I swap the radeon card with a matrox card, and change the driver line in XF86Config-4 from radeon to mga, everything works great. I have not tried the radeon card in another machine, nor have I used it under Windows. 2d performance is fine, and I otherwise have no X related problems. foo:~%env LIBGL_DEBUG=verbose glxgears libGL: XF86DRIGetClientDriverName: 4.0.1 radeon (screen 0) libGL: OpenDriver: trying /usr/X11R6/lib/modules-dri-trunk/dri/radeon_dri.so libGL: OpenDriver: trying /usr/X11R6/lib/modules-dri-trunk/dri/radeon_dri.so drmOpenByBusid: busid is PCI:1:5:0 drmOpenDevice: minor is 0 drmOpenDevice: node name is /dev/dri/card0 drmOpenDevice: open result is 4, (OK) drmOpenByBusid: drmOpenMinor returns 4 drmOpenByBusid: drmGetBusid reports PCI:1:5:0 foo:~%cat /proc/mtrr reg00: base=0x ( 0MB), size= 512MB: write-back, count=1 reg01: base=0xe000 (3584MB), size= 64MB: write-combining, count=2 reg07: base=0xf000 (3840MB), size= 128MB: write-combining, count=2 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: A Radeon and X question, probably not Debian specific
On Thu, 2002-07-25 at 00:17, Jeff Lessem wrote: > In your message of: 24 Jul 2002 23:22:23 +0200, you write: > >The size of the latter's window doesn't seem to impact the framerate here, > >so that sounds like software rendering. > > Xscreensaver's gears does about 50 fps at its default size, 25 fps in > a maximized window, Around 45 fps in both cases here. > and 5 fps in full screen saved mode. There seems to be a bug in the current DRI code which prevents direct rendering under certain circumstances, in particular this one. It's hard to track down, it fails to map the framebuffer - no idea what's up yet. > foo:~%env LIBGL_DEBUG=verbose glxgears > libGL: XF86DRIGetClientDriverName: 4.0.1 radeon (screen 0) > libGL: OpenDriver: trying /usr/X11R6/lib/modules-dri-trunk/dri/radeon_dri.so > libGL: OpenDriver: trying /usr/X11R6/lib/modules-dri-trunk/dri/radeon_dri.so > drmOpenByBusid: busid is PCI:1:5:0 > drmOpenDevice: minor is 0 > drmOpenDevice: node name is /dev/dri/card0 > drmOpenDevice: open result is 4, (OK) > drmOpenByBusid: drmOpenMinor returns 4 > drmOpenByBusid: drmGetBusid reports PCI:1:5:0 Looks good, hmm. I have a similar problem when running with PCI GART on this PowerBook, but that can't be the case for you because it's still disabled on x86. I assume it's related to the L3 cache here, maybe something about AGP for you... For the record, can you try if setting any of RADEON_NO_CODEGEN=t RADEON_NO_VTXFMT=t RADEON_NO_TCL=t makes a difference? -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: A Radeon and X question, probably not Debian specific
In your message of: 25 Jul 2002 00:31:18 +0200, you write: >Around 45 fps in both cases here. It must be limited to about 50, because that is what I get on my Mobility r128 laptop and on the old system with the mga card. >There seems to be a bug in the current DRI code which prevents direct >rendering under certain circumstances, in particular this one. It's hard >to track down, it fails to map the framebuffer - no idea what's up yet. Yeah, I thought it would be some strange bug, or an interaction with the AMD 760 AGP bridge, or something like that. I was just hoping to hear about some special REALLY_GO_FAST=t type thing that I didn't know to set. >For the record, can you try if setting any of > >RADEON_NO_CODEGEN=t >RADEON_NO_VTXFMT=t >RADEON_NO_TCL=t > >makes a difference? I set all of those alone and all together, and there wasn't any difference. Everything still worked, it was just slow. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: A Radeon and X question, probably not Debian specific
On Thu, 2002-07-25 at 01:02, Jeff Lessem wrote: > In your message of: 25 Jul 2002 00:31:18 +0200, you write: > >Around 45 fps in both cases here. > > It must be limited to about 50, because that is what I get on my > Mobility r128 laptop and on the old system with the mga card. Yes, it has a 0.02 sec delay between frames by default. Run it with -delay 0 to get the max framerate. > >There seems to be a bug in the current DRI code which prevents direct > >rendering under certain circumstances, in particular this one. It's hard > >to track down, it fails to map the framebuffer - no idea what's up yet. > > Yeah, I thought it would be some strange bug, or an interaction with > the AMD 760 AGP bridge, or something like that. I was just hoping to > hear about some special REALLY_GO_FAST=t type thing that I didn't know > to set. Let me clarify, you're seeing two separate problems: you only get software rendering with xscreensaver due to a strange DRI bug. The slowness with direct rendering could be a hardware quirk. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Connecting to a Beamer
> [debian-x, please CC me as I am not subscribed] > > Hi folks, > > I need to connect my laptop (Latitude C610) to a not-so-modern beamer, > which supports displaying the 1024x768 screen of the desktop machine ... > martin; (greetings from the heart of the sun.) > \ echo mailto: !#^."<*>"|tr "<*> mailto:"; [EMAIL PROTECTED] ... You need to know the beamers HorizSync and VertRefresh. Update your XF86Config-4 accordingly. Regards, /Karl --- Karl HammarAspö Data [EMAIL PROTECTED] Lilla Aspö 2340 +46 173 140 57Networks S-742 94 Östhammar +46 18 26 09 00 Computers Sweden +46 10 270 26 67 Consulting --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: discover, read-edid, mdetect selection in base-config
On Tue, Jul 23, 2002 at 10:17:17AM +0200, Steinar Bang wrote: > Are read-edid and mdetect useful for configuring XFree86 4.1.0? > Doesn't "XFree86 -configure" do a better job? You tell me. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=77120&repeatmerged=yes http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=124247&repeatmerged=yes A configuration process that segfaults. M, good! -- G. Branden Robinson| Mob rule isn't any prettier just Debian GNU/Linux | because you call your mob a [EMAIL PROTECTED] | government. http://people.debian.org/~branden/ | msg03830/pgp0.pgp Description: PGP signature
Re: XF4.2 packages think Alt is Meta
Branden Robinson writes: > 1) I personally don't care, since I remap the funny keys on the bottom > row to suit my own strange tastes anyway; > 2) I don't want to subject my users to shock and horror. > 3) I don't want to deviate gratuitously from upstream practice. > I'd appreciate input from the people on this list as to the impact of > XFree86 4.2 on 2) and 3). Are those goals truly in conflict? Is this > only a problem for people with buggy window managers? No, it will most probably affect all users of Emacs and other clones of Emacs as well, since it makes intensive use of the Meta key (and the default behavior was to have the left Windows key as Meta). You might argue that Emacs addicts probably rebind their keys as they see fit, but for the average or occasional user, the change might be disturbing. Maybe the upgrade process should warn the user about the change, and maybe even ask if it should keep the old behavior. My 2c, -- Romain FRANCOISE <[EMAIL PROTECTED]> | I know, you love the song but it's a miracle -- http://orebokech.com/ | not the singer. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: discover, read-edid, mdetect selection in base-config
> Branden Robinson <[EMAIL PROTECTED]>: > On Tue, Jul 23, 2002 at 10:17:17AM +0200, Steinar Bang wrote: >> Are read-edid and mdetect useful for configuring XFree86 4.1.0? >> Doesn't "XFree86 -configure" do a better job? > You tell me. > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=77120&repeatmerged=yes > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=124247&repeatmerged=yes > A configuration process that segfaults. M, good! I would have tried with "maybe upstream's fixed this?", except 4.2.0 died on me two days ago, because it didn't recognize a GeForce4 card...:-) But when it works, which it has usually done for me, it creates good XF86Config files needing a minimum of tinkering (usually finding vsync and hsync for older monitors). -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
test (sorry)
signature.asc Description: This is a digitally signed message part
Re: A Radeon and X question, probably not Debian specific
"Charl P. Botha" <[EMAIL PROTECTED]> writes: > I had a look at your X output and all seems fine. How many FPS does > glxgears yield? You have to run it for at least 5 seconds before it outputs > something to stdout. glxgears, in its default small size is about 500 fps, expanding the window to full screen gets me 25 fps. Xscreensaver's gears hack gets about 5 fps running full screen. On the advice of Michel Dänzer (thanks!) I added the debian dri-trunk stuff, and got that working easily enough with the 1.4.0 20020714 Radeon drm driver, but it didn't make any difference. Everything appears to be working correctly, but gl stuff is very slow. The only error I get is [drm:radeon_freelist_get] *ERROR* returning NULL! output from the drm driver. As best I can tell, xinerama is disabled. Starting X with +xinerama put the xinerama enabled comment in the logfile, and disables dri, so I used -xinerama to be safe, but that makes no difference. If anybody has more advice, I would be interested in hearing it. Thanks for the help so far. Fortunately my next dual athlon is coming with an $18 8mb ATI Mobility AGP card, which should be more than adequate to run the server room display in text mode. -- Jeff Lessem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Unicode compose map?
On Tue, Jul 23, 2002 at 06:21:47PM +0200, Juliusz Chroboczek wrote: > The alternative to the compose mechanism -- using the input method > mechanism -- is just too nasty to consider. End of offtopic.) Is that the same as the 'altgr' key? > I can personally confirm that the non-Greek compose file does contain > entries for the ISO 8859-2 characters needed for Polish; it probably > contains all combinations that you need, but I haven't checked myself. Great. Can you either send me the compose files, or point me to the correct cvs directory? I dont want to check out the entire source tree to get two files :) In the mean time, I also stumbled across an updated version of the debian-utf8 howto: http://melkor.dnp.fmph.uniba.sk/~garabik/debian-utf8/howto.html They mention a compose map by Pablo Saratxaga, but I didnt get a chance to try it out yet. > Note that the above applies to current XFree86 CVS; I don't know what > the status of 4.2.0 is. Note furthermore that if you're working with > UTF-8, you really ought to upgrade your X libraries to 4.2.0 or later. I havent noticed any problems so far, but I'm not using it thought the entire X Window system. My general locale is en_US, and I set it to en_US.UTF-8 only if TERM=xterm|screen. This is mainly because I use gnome, which is unable to display any text if I try to use unicode. Would 4.2.0 libs help with that problem? Thanks, Norbert msg03835/pgp0.pgp Description: PGP signature
Connecting to a Beamer
[debian-x, please CC me as I am not subscribed] Hi folks, I need to connect my laptop (Latitude C610) to a not-so-modern beamer, which supports displaying the 1024x768 screen of the desktop machine sitting next to it. However, when I tell it to display my screen, it simply says: Not available. H: 58.44Khz V: 53.5Hz Ok, granted, I am trying to display 1400x1050, but I get the same message when I tune my resolution to 1024x768 (I defined two screens in XF86Config-4, one for the laptop, one for the beamer). Using `startx -- -layout present` startx the 1024x768 version -- I've never managed to get in-X resolution switching via a key combo to work. Anyway, even at 1024x768 I get the above message, and yet I need to connect to the beamer. I don't know what to do. I'd so appreciate if some X-savvy person would help me out a little, you can get my XF86Config-4 (for Branden's 4.2 packages on testing) here: ftp://ftp.madduck.net/pub/temp/XF86Config-4.gz (1203 octets) I'd appreciate any tips! Thanks, -- martin; (greetings from the heart of the sun.) \ echo mailto: !#^."<*>"|tr "<*> mailto:"; net@madduck "you raise the blade, you make the change you rearrange me till i'm sane. you lock the door, and throw away the key, there's someone in my head but it's not me." -- pink floyd, 1972 msg03836/pgp0.pgp Description: PGP signature
Re: Connecting to a Beamer
On Wed, 2002-07-24 at 19:09, martin f krafft wrote: > [debian-x, please CC me as I am not subscribed] > > Hi folks, > > I need to connect my laptop (Latitude C610) to a not-so-modern beamer, > which supports displaying the 1024x768 screen of the desktop machine > sitting next to it. However, when I tell it to display my screen, it > simply says: > > Not available. H: 58.44Khz V: 53.5Hz > > Ok, granted, I am trying to display 1400x1050, but I get the same > message when I tune my resolution to 1024x768 (I defined two screens > in XF86Config-4, one for the laptop, one for the beamer). Using > `startx -- -layout present` startx the 1024x768 version -- I've never > managed to get in-X resolution switching via a key combo to work. > > Anyway, even at 1024x768 I get the above message, and yet I need to > connect to the beamer. I don't know what to do. I'd so appreciate if > some X-savvy person would help me out a little, you can get my > XF86Config-4 (for Branden's 4.2 packages on testing) here: > > ftp://ftp.madduck.net/pub/temp/XF86Config-4.gz (1203 octets) > > I'd appreciate any tips! Try Option "CRTScreen" in the device section. Without that, the external output will use the parameters for the internal panel. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: A Radeon and X question, probably not Debian specific
On Wed, 2002-07-24 at 20:39, Jeff Lessem wrote: > "Charl P. Botha" <[EMAIL PROTECTED]> writes: > > > I had a look at your X output and all seems fine. How many FPS does > > glxgears yield? You have to run it for at least 5 seconds before it outputs > > something to stdout. > > glxgears, in its default small size is about 500 fps, expanding the > window to full screen gets me 25 fps. Xscreensaver's gears hack gets > about 5 fps running full screen. The size of the latter's window doesn't seem to impact the framerate here, so that sounds like software rendering. > On the advice of Michel Dänzer (thanks!) I added the debian dri-trunk > stuff, and got that working easily enough with the 1.4.0 20020714 > Radeon drm driver, but it didn't make any difference. Everything > appears to be working correctly, but gl stuff is very slow. The only > error I get is [drm:radeon_freelist_get] *ERROR* returning NULL! > output from the drm driver. That on the other hand only appears with direct rendering... Does running a client with LIBGL_DEBUG=verbose give any more interesting info as to whether you're actually using direct rendering? -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: A Radeon and X question, probably not Debian specific
In your message of: 24 Jul 2002 23:22:23 +0200, you write: >The size of the latter's window doesn't seem to impact the framerate here, >so that sounds like software rendering. Xscreensaver's gears does about 50 fps at its default size, 25 fps in a maximized window, and 5 fps in full screen saved mode. I wish it was software rendering, because then I would know what problem I have to fix. Everything seems to be working fine, if I swap the radeon card with a matrox card, and change the driver line in XF86Config-4 from radeon to mga, everything works great. I have not tried the radeon card in another machine, nor have I used it under Windows. 2d performance is fine, and I otherwise have no X related problems. foo:~%env LIBGL_DEBUG=verbose glxgears libGL: XF86DRIGetClientDriverName: 4.0.1 radeon (screen 0) libGL: OpenDriver: trying /usr/X11R6/lib/modules-dri-trunk/dri/radeon_dri.so libGL: OpenDriver: trying /usr/X11R6/lib/modules-dri-trunk/dri/radeon_dri.so drmOpenByBusid: busid is PCI:1:5:0 drmOpenDevice: minor is 0 drmOpenDevice: node name is /dev/dri/card0 drmOpenDevice: open result is 4, (OK) drmOpenByBusid: drmOpenMinor returns 4 drmOpenByBusid: drmGetBusid reports PCI:1:5:0 foo:~%cat /proc/mtrr reg00: base=0x ( 0MB), size= 512MB: write-back, count=1 reg01: base=0xe000 (3584MB), size= 64MB: write-combining, count=2 reg07: base=0xf000 (3840MB), size= 128MB: write-combining, count=2 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: A Radeon and X question, probably not Debian specific
On Thu, 2002-07-25 at 00:17, Jeff Lessem wrote: > In your message of: 24 Jul 2002 23:22:23 +0200, you write: > >The size of the latter's window doesn't seem to impact the framerate here, > >so that sounds like software rendering. > > Xscreensaver's gears does about 50 fps at its default size, 25 fps in > a maximized window, Around 45 fps in both cases here. > and 5 fps in full screen saved mode. There seems to be a bug in the current DRI code which prevents direct rendering under certain circumstances, in particular this one. It's hard to track down, it fails to map the framebuffer - no idea what's up yet. > foo:~%env LIBGL_DEBUG=verbose glxgears > libGL: XF86DRIGetClientDriverName: 4.0.1 radeon (screen 0) > libGL: OpenDriver: trying /usr/X11R6/lib/modules-dri-trunk/dri/radeon_dri.so > libGL: OpenDriver: trying /usr/X11R6/lib/modules-dri-trunk/dri/radeon_dri.so > drmOpenByBusid: busid is PCI:1:5:0 > drmOpenDevice: minor is 0 > drmOpenDevice: node name is /dev/dri/card0 > drmOpenDevice: open result is 4, (OK) > drmOpenByBusid: drmOpenMinor returns 4 > drmOpenByBusid: drmGetBusid reports PCI:1:5:0 Looks good, hmm. I have a similar problem when running with PCI GART on this PowerBook, but that can't be the case for you because it's still disabled on x86. I assume it's related to the L3 cache here, maybe something about AGP for you... For the record, can you try if setting any of RADEON_NO_CODEGEN=t RADEON_NO_VTXFMT=t RADEON_NO_TCL=t makes a difference? -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: A Radeon and X question, probably not Debian specific
In your message of: 25 Jul 2002 00:31:18 +0200, you write: >Around 45 fps in both cases here. It must be limited to about 50, because that is what I get on my Mobility r128 laptop and on the old system with the mga card. >There seems to be a bug in the current DRI code which prevents direct >rendering under certain circumstances, in particular this one. It's hard >to track down, it fails to map the framebuffer - no idea what's up yet. Yeah, I thought it would be some strange bug, or an interaction with the AMD 760 AGP bridge, or something like that. I was just hoping to hear about some special REALLY_GO_FAST=t type thing that I didn't know to set. >For the record, can you try if setting any of > >RADEON_NO_CODEGEN=t >RADEON_NO_VTXFMT=t >RADEON_NO_TCL=t > >makes a difference? I set all of those alone and all together, and there wasn't any difference. Everything still worked, it was just slow. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: A Radeon and X question, probably not Debian specific
On Thu, 2002-07-25 at 01:02, Jeff Lessem wrote: > In your message of: 25 Jul 2002 00:31:18 +0200, you write: > >Around 45 fps in both cases here. > > It must be limited to about 50, because that is what I get on my > Mobility r128 laptop and on the old system with the mga card. Yes, it has a 0.02 sec delay between frames by default. Run it with -delay 0 to get the max framerate. > >There seems to be a bug in the current DRI code which prevents direct > >rendering under certain circumstances, in particular this one. It's hard > >to track down, it fails to map the framebuffer - no idea what's up yet. > > Yeah, I thought it would be some strange bug, or an interaction with > the AMD 760 AGP bridge, or something like that. I was just hoping to > hear about some special REALLY_GO_FAST=t type thing that I didn't know > to set. Let me clarify, you're seeing two separate problems: you only get software rendering with xscreensaver due to a strange DRI bug. The slowness with direct rendering could be a hardware quirk. -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Connecting to a Beamer
> [debian-x, please CC me as I am not subscribed] > > Hi folks, > > I need to connect my laptop (Latitude C610) to a not-so-modern beamer, > which supports displaying the 1024x768 screen of the desktop machine ... > martin; (greetings from the heart of the sun.) > \ echo mailto: !#^."<*>"|tr "<*> mailto:"; net@madduck ... You need to know the beamers HorizSync and VertRefresh. Update your XF86Config-4 accordingly. Regards, /Karl --- Karl HammarAspö Data [EMAIL PROTECTED] Lilla Aspö 2340 +46 173 140 57Networks S-742 94 Östhammar +46 18 26 09 00 Computers Sweden +46 10 270 26 67 Consulting --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: A Radeon and X question, probably not Debian specific
> Yeah, I thought it would be some strange bug, or an interaction with > the AMD 760 AGP bridge, or something like that. I was just hoping > to hear about some special REALLY_GO_FAST=t type thing that I didn't > know to set. I had exactly the same problem with a Radeon VE. It worked just fine in flatland (except when you dragged a window. Then it drew garbage across the screen until you stopped, but that was a separate problem). Nothing I did could get it to do OpenGL. It never produced an error, even with everything at maximum log spewage. On a modestly swift AMD system, glxgears produced an insanely high number, but Quake III gasped for half a frame per second. After many months of research, I was able to solve the problem like so: Section "Device" Identifier "geforce" Driver "nvidia" VideoRam65536 BusID "PCI:1:0:0" EndSection Of course, this required some modification of the hardware configuration. That's probably not the answer you're looking for either. ^_^ Russell -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
