R5xx Mobilities and Rv6x0: PLL Fixes
Hi all, I finally pushed some fixes to the PLL setup and PLL generation code, and now all problems should be gone. If you previously had any problems with the PLL and experienced modes which your monitor wouldn't sync to or which would be wobbling horizontally, then this commit should fix this. These things were specifically noticable on the radeon mobilities of the r500 generation, as these had the pll spread spectrum turned on. And also, many of the generated modes simply weren't acceptable for rv6x0 generation hardware, with a redesigned and pickier PLL. Please give this new code a whirl, and if you still see some dotclock issues with this, tell me about it asap so i can go and find and fix them. Thanks, Luc Verhaegen. SUSE/Novell X Driver Developer. -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
Am Montag, den 22.10.2007, 23:56 +0200 schrieb Luc Verhaegen:
Hi all,
I finally pushed some fixes to the PLL setup and PLL generation code, and now all problems should be gone.
If you previously had any problems with the PLL and experienced modes which your monitor wouldn't sync to or which would be wobbling horizontally, then this commit should fix this.
These things were specifically noticable on the radeon mobilities of the r500 generation, as these had the pll spread spectrum turned on. And also, many of the generated modes simply weren't acceptable for rv6x0 generation hardware, with a redesigned and pickier PLL.
Please give this new code a whirl, and if you still see some dotclock issues with this, tell me about it asap so i can go and find and fix them.
Thanks,
Hi Luc, I synced to the newest git but this new patch gives me just a blank (black) screen when switching to other (text) consoles. By reverting just this patch on my local clone, I have back the ability to switch consoles. The garbled screen after a suspend/resume to ram is the same on both versions - but that is a different issue. System here is a thinkpad z61p conntest says: Checking connectors on 0x71C4, 0x17AA, 0x202B (@01:00:00): Load Detection: RHD_OUTPUT_NONE HotPlug: RHD_HPD_NONE DDC: RHD_DDC_2 LVDS Info: 18bits, dual link, LDI Panel found. Power Timing: 0x8CA, 0x00F, 0x04, 0x19, 0x0A8 Macro: 0x0CD20407, Clock Pattern: 0x0063 lspci says: 01:00.0 VGA compatible controller: ATI Technologies Inc M56GL [Mobility FireGL V5200] xorg.log attached If you need anything else - just let me know. keep up the great work Henry
Luc Verhaegen. SUSE/Novell X Driver Developer.
On Tue, Oct 23, 2007 at 07:42:47PM +0200, henry ritzlmayr wrote:
Am Montag, den 22.10.2007, 23:56 +0200 schrieb Luc Verhaegen:
Hi all,
I finally pushed some fixes to the PLL setup and PLL generation code, and now all problems should be gone.
If you previously had any problems with the PLL and experienced modes which your monitor wouldn't sync to or which would be wobbling horizontally, then this commit should fix this.
These things were specifically noticable on the radeon mobilities of the r500 generation, as these had the pll spread spectrum turned on. And also, many of the generated modes simply weren't acceptable for rv6x0 generation hardware, with a redesigned and pickier PLL.
Please give this new code a whirl, and if you still see some dotclock issues with this, tell me about it asap so i can go and find and fix them.
Thanks,
Hi Luc,
I synced to the newest git but this new patch gives me just a blank (black) screen when switching to other (text) consoles. By reverting just this patch on my local clone, I have back the ability to switch consoles.
The garbled screen after a suspend/resume to ram is the same on both versions - but that is a different issue.
System here is a thinkpad z61p conntest says:
Checking connectors on 0x71C4, 0x17AA, 0x202B (@01:00:00): Load Detection: RHD_OUTPUT_NONE HotPlug: RHD_HPD_NONE DDC: RHD_DDC_2 LVDS Info: 18bits, dual link, LDI Panel found. Power Timing: 0x8CA, 0x00F, 0x04, 0x19, 0x0A8 Macro: 0x0CD20407, Clock Pattern: 0x0063
lspci says: 01:00.0 VGA compatible controller: ATI Technologies Inc M56GL [Mobility FireGL V5200]
xorg.log attached
If you need anything else - just let me know.
keep up the great work Henry
Hrm, this is pretty weird. I have just tested bot vga console and vesa on an M56 (71C5), and it is perfectly happy here. The culpritt can only be the PLL spread spectrum that is now saved and restored. I have attached a patch for extra debugging, and a patch to possibly fix this issue. Please apply both and test again. Thanks, Luc Verhaegen. SUSE/Novell X Driver Developer.
Am Mittwoch, den 24.10.2007, 16:25 +0200 schrieb Luc Verhaegen:
On Tue, Oct 23, 2007 at 07:42:47PM +0200, henry ritzlmayr wrote:
Am Montag, den 22.10.2007, 23:56 +0200 schrieb Luc Verhaegen:
Hi all,
I finally pushed some fixes to the PLL setup and PLL generation code, and now all problems should be gone.
If you previously had any problems with the PLL and experienced modes which your monitor wouldn't sync to or which would be wobbling horizontally, then this commit should fix this.
These things were specifically noticable on the radeon mobilities of the r500 generation, as these had the pll spread spectrum turned on. And also, many of the generated modes simply weren't acceptable for rv6x0 generation hardware, with a redesigned and pickier PLL.
Please give this new code a whirl, and if you still see some dotclock issues with this, tell me about it asap so i can go and find and fix them.
Thanks,
Hi Luc,
I synced to the newest git but this new patch gives me just a blank (black) screen when switching to other (text) consoles. By reverting just this patch on my local clone, I have back the ability to switch consoles.
The garbled screen after a suspend/resume to ram is the same on both versions - but that is a different issue.
System here is a thinkpad z61p conntest says:
Checking connectors on 0x71C4, 0x17AA, 0x202B (@01:00:00): Load Detection: RHD_OUTPUT_NONE HotPlug: RHD_HPD_NONE DDC: RHD_DDC_2 LVDS Info: 18bits, dual link, LDI Panel found. Power Timing: 0x8CA, 0x00F, 0x04, 0x19, 0x0A8 Macro: 0x0CD20407, Clock Pattern: 0x0063
lspci says: 01:00.0 VGA compatible controller: ATI Technologies Inc M56GL [Mobility FireGL V5200]
xorg.log attached
If you need anything else - just let me know.
keep up the great work Henry
Hrm, this is pretty weird. I have just tested bot vga console and vesa on an M56 (71C5), and it is perfectly happy here.
The culpritt can only be the PLL spread spectrum that is now saved and restored.
I have attached a patch for extra debugging, and a patch to possibly fix this issue. Please apply both and test again.
Thanks,
I applied both patches. radeonhd_pll_ss_mask.diff had to be done manually! Symptoms stay the same. Switching to a console gives only a blank(black) screen with backlight on. I also tried to sync to the latest git to reapply your patches against it, but the current git is not compile-able: configure.ac:189: required file `man/Makefile.in' not found Xorg.0.log is attached - did you want me to run the with -logverbose 7? Cheers, Henry
Luc Verhaegen. SUSE/Novell X Driver Developer.
On Wed, Oct 24, 2007 at 05:28:23PM +0200, henry ritzlmayr wrote:
I applied both patches. radeonhd_pll_ss_mask.diff had to be done manually!
hrm, weird, what was wrong there?
Symptoms stay the same. Switching to a console gives only a blank(black) screen with backlight on.
Right, so no fix, let me attach another one which might fix things here.
I also tried to sync to the latest git to reapply your patches against it, but the current git is not compile-able:
configure.ac:189: required file `man/Makefile.in' not found
rerun autogen.sh This will rerun automake.
Xorg.0.log is attached - did you want me to run the with -logverbose 7?
Cheers, Henry
For the debugging, yes :) Thanks, Luc Verhaegen. SUSE/Novell X Driver Developer.
Am Mittwoch, den 24.10.2007, 17:52 +0200 schrieb Luc Verhaegen:
On Wed, Oct 24, 2007 at 05:28:23PM +0200, henry ritzlmayr wrote:
I applied both patches. radeonhd_pll_ss_mask.diff had to be done manually!
hrm, weird, what was wrong there?
I didn´t sync my git clone here after the problem started - so I guess I didn´t have the same rhd_pll.c to patch where you diffed it.
Symptoms stay the same. Switching to a console gives only a blank(black) screen with backlight on.
Right, so no fix, let me attach another one which might fix things here.
This one did the trick. My consoles are back again. ---> big thanks.
I also tried to sync to the latest git to reapply your patches against it, but the current git is not compile-able:
configure.ac:189: required file `man/Makefile.in' not found
rerun autogen.sh This will rerun automake.
Thanks - but I tried that already - failed. Anyway with commit 533e86624eb0dfa4c32a22312e14b9750a1136db Hans Ulrich provided the missing files - so I am up to date again.
Xorg.0.log is attached - did you want me to run the with -logverbose 7?
Cheers, Henry
For the debugging, yes :)
Since the problem is solved - do you still need the logs? Cheers, Henry -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
On Thu, Oct 25, 2007 at 08:16:00AM +0200, henry ritzlmayr wrote:
Am Mittwoch, den 24.10.2007, 17:52 +0200 schrieb Luc Verhaegen:
hrm, weird, what was wrong there?
I didn??t sync my git clone here after the problem started - so I guess I didn??t have the same rhd_pll.c to patch where you diffed it.
Ah, of course. This was a diff against the commited tree, not a diff against the previous patch.
Right, so no fix, let me attach another one which might fix things here.
This one did the trick. My consoles are back again.
Ah, smashing. Will commit in a bit.
---> big thanks.
No, thank you for testing this. I wasn't seeing this here, so there was no use testing this here.
Thanks - but I tried that already - failed.
Anyway with commit 533e86624eb0dfa4c32a22312e14b9750a1136db Hans Ulrich provided the missing files - so I am up to date again.
Right, not created by autoconf.
Since the problem is solved - do you still need the logs?
Nope, i don't need this anymore indeed. Luc Verhaegen. SUSE/Novell X Driver Developer. -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
On 23.10.2007 03:56:18, Luc Verhaegen wrote:
Hi all, I finally pushed some fixes to the PLL setup and PLL generation code, and now all problems should be gone. If you previously had any problems with the PLL and experienced modes which your monitor wouldn't sync to or which would be wobbling horizontally, then this commit should fix this. These things were specifically noticable on the radeon mobilities of the r500 generation, as these had the pll spread spectrum turned on. And also, many of the generated modes simply weren't acceptable for rv6x0 generation hardware, with a redesigned and pickier PLL. Please give this new code a whirl, and if you still see some dotclock issues with this, tell me about it asap so i can go and find and fix them.
Doesn't look like working for me. Though the progress is obvious. Now it doesn't reject the LVDS. Here are two logs: w/ and w/o external panel. Unfortunately I don't know how to turn off LVDS in Driver section. I tried Option "MonitorLayout" "CRT,NONE" -- gave nothing to me. I think, if I could turn LVDS off, external panel would show me something... Who knows...
Thanks,
Thank YOU! -- Ildar Mulyukov, free SW designer/programmer ================================================ email: ildar@users.sourceforge.net home: http://tuganger.narod.ru/ ALT Linux Sisyphus ================================================
Allright, hackers, today I've got my Radeon X1200 (RS690) working for the 1st time.. This happened with external panel plugged off. THANK YOU VERY MUCH. I hope it to be more stable then fglrx driver hanging my PC once or twice a day. Now let me announce the list of things I miss with my new driver: 1. External panel still doesn't work. Xorg log is attached, just in case. 2. glxinfo says: Xlib: extension "GLX" missing on display ":0.0" 3. xvinfo says no adaptors present (XV extension) 4. Dualhead/xinerama is very much desired. I don't mean that the features will be "fixed" ASAP, but still I hope to see them someday. It's probably a good idea to make bugzilla entries for these features. With bugzilla it's very convenient to track interesting things, and on changes quickly fetch/compile/test. (Mailing list is not very helpful for this purpose). Thank you very much for the great work. Best regards, -- Ildar Mulyukov, free SW designer/programmer ================================================ email: ildar@users.sourceforge.net home: http://tuganger.narod.ru/ ALT Linux Sisyphus ================================================
On Sun, Oct 28, 2007 at 04:39:20PM +0600, Ildar Mulyukov wrote:
2. glxinfo says: Xlib: extension "GLX" missing on display ":0.0"
I don't think this is related to the radeonhd driver at all. Verify with vesa or fbdev driver. Best regards, Stefan Public Key available ------------------------------------------------------ Stefan Dirsch (Res. & Dev.) SUSE LINUX Products GmbH Tel: 0911-740 53 0 Maxfeldstraße 5 FAX: 0911-740 53 479 D-90409 Nürnberg http://www.suse.de Germany ----------------------------------------------------------------- SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg) ----------------------------------------------------------------- -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
On 29.10.2007 16:47:56, Stefan Dirsch wrote:
On Sun, Oct 28, 2007 at 04:39:20PM +0600, Ildar Mulyukov wrote:
2. glxinfo says: Xlib: extension "GLX" missing on display ":0.0"
I don't think this is related to the radeonhd driver at all. Verify with vesa or fbdev driver.
This is 100% true. I've accidentally commented out "Load GLX". Now I've put it back. But in fact GL is very slow. glxgears shows ~200FPS. Looks very much like software rendering. So the question persists: GL is not working. Thanks! Regards, Ildar. -- Ildar Mulyukov, free SW designer/programmer ================================================ email: ildar@users.sourceforge.net home: http://tuganger.narod.ru/ ALT Linux Sisyphus ================================================ -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
On Oct 29, 07 17:14:37 +0600, Ildar Mulyukov wrote:
On 29.10.2007 16:47:56, Stefan Dirsch wrote:
On Sun, Oct 28, 2007 at 04:39:20PM +0600, Ildar Mulyukov wrote:
2. glxinfo says: Xlib: extension "GLX" missing on display ":0.0"
I don't think this is related to the radeonhd driver at all. Verify with vesa or fbdev driver.
This is 100% true. I've accidentally commented out "Load GLX". Now I've put it back. But in fact GL is very slow. glxgears shows ~200FPS. Looks very much like software rendering.
LOL.
Of course this is software rendering. As announced already a number of
times we DON'T HAVE ANY DOCS on hardware acceleration.
Actually, commenting out 'Load "GLX"' is good, because with some xserver
versions you could crash the system (due to some internal server issues).
Matthias
--
Matthias Hopf
hi Sorry, probably that was answered many times already, but ... cut ...
LOL. Of course this is software rendering. As announced already a number of times we DON'T HAVE ANY DOCS on hardware acceleration.
And these docs are not about acceleration also ? http://www.x.org/docs/AMD/42589_rv630_rrg_1.01o.pdf http://www.x.org/docs/AMD/RRG-216M56-03oOEM.pdf (I do not understand anything in graphic chips specs, sorry) And there are no any hope to see GLX (on x1400 in my case) in near future ? (I am using FreeBSD so frglx is not a option for me)
Matthias
-- Vladimir B. Grebenschikov vova@fbsd.ru -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
On Oct 30, 07 12:08:27 +0300, Vladimir Grebenschikov wrote:
Of course this is software rendering. As announced already a number of times we DON'T HAVE ANY DOCS on hardware acceleration.
And these docs are not about acceleration also ? http://www.x.org/docs/AMD/42589_rv630_rrg_1.01o.pdf http://www.x.org/docs/AMD/RRG-216M56-03oOEM.pdf
No.
And there are no any hope to see GLX (on x1400 in my case) in near future ? (I am using FreeBSD so frglx is not a option for me)
Depends on how you define near future, and how things progress here. A
full OpenGL implementation takes at least a year. glxgears might run
earlier. Also we can only start when we have docs, and 2D acceleration
will come first. Don't expect the docs this year any more.
Matthias
--
Matthias Hopf
On Oct 28, 07 16:39:20 +0600, Ildar Mulyukov wrote:
2. glxinfo says: Xlib: extension "GLX" missing on display ":0.0" 3. xvinfo says no adaptors present (XV extension) 4. Dualhead/xinerama is very much desired.
I don't mean that the features will be "fixed" ASAP, but still I hope to see them someday. It's probably a good idea to make bugzilla entries for these features. With bugzilla it's very convenient to track interesting
Nope. These are not bugs, these are missing features. And we're well
aware of that. Bugzilla entries for these don't make any sense.
Matthias
--
Matthias Hopf
On 29.10.2007 17:40:27, Matthias Hopf wrote:
On Oct 28, 07 16:39:20 +0600, Ildar Mulyukov wrote:
2. glxinfo says: Xlib: extension "GLX" missing on display ":0.0" 3. xvinfo says no adaptors present (XV extension) 4. Dualhead/xinerama is very much desired. I don't mean that the features will be "fixed" ASAP, but still I hope to see them someday. It's probably a good idea to make bugzilla entries for these features. With bugzilla it's very convenient to track interesting
Nope. These are not bugs, these are missing features. And we're well aware of that. Bugzilla entries for these don't make any sense.
OK I listed above just to be sure I don't miss something. Thanks. -- Ildar Mulyukov, free SW designer/programmer ================================================ email: ildar@users.sourceforge.net home: http://tuganger.narod.ru/ ALT Linux Sisyphus ================================================ -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
participants (6)
-
henry ritzlmayr -
Ildar Mulyukov -
Luc Verhaegen -
Matthias Hopf -
Stefan Dirsch -
Vladimir Grebenschikov