| 1 |
2009/1/3 Volker Armin Hemmann <volkerarmin@××××××××××.com>: |
| 2 |
> On Samstag 03 Januar 2009, Beso wrote: |
| 3 |
>> 2009/1/2 Volker Armin Hemmann <volker.armin.hemmann@××××××××××××.de>: |
| 4 |
>> > On Freitag 02 Januar 2009, Beso wrote: |
| 5 |
>> >> 2009/1/1 Sebastian Redl <sebastian.redl@×××××××××××.at>: |
| 6 |
>> >> > Hi, |
| 7 |
>> >> > |
| 8 |
>> >> > When my hard-disk made suspicious noises recently, I decided to |
| 9 |
>> >> > replace it - be safe rather than sorry - and use the opportunity to do |
| 10 |
>> >> > a complete, fresh install of my system. The new system is up and |
| 11 |
>> >> > running (OpenOffice troubles aside), with the added advantage that I |
| 12 |
>> >> > finally managed to get to the most recent ATI binary driver version |
| 13 |
>> >> > instead of that stone-age version that happened to not hang my system |
| 14 |
>> >> > under the old setup. |
| 15 |
>> >> > |
| 16 |
>> >> > With the new driver, I want to use its AIGLX capabilities. However, |
| 17 |
>> >> > I'm having trouble getting XComposite to work. I have it enabled in |
| 18 |
>> >> > xorg.conf: |
| 19 |
>> >> > |
| 20 |
>> >> > Section "Extensions" |
| 21 |
>> >> > Option "Composite" "Enable" |
| 22 |
>> >> > EndSection |
| 23 |
>> >> |
| 24 |
>> >> you need also to enable damage and render extension in the extensions |
| 25 |
>> >> section. here's what |
| 26 |
>> >> ity should look like: |
| 27 |
>> > |
| 28 |
>> > NO |
| 29 |
>> > |
| 30 |
>> > he does not. In fact he does not have to enable anything of this. |
| 31 |
>> |
| 32 |
>> when i've last been with fglrx you would need to enable that manually or |
| 33 |
>> they won't start. as i've said it has been about 6-8 months that i haven't |
| 34 |
>> been using fglrx. |
| 35 |
>> |
| 36 |
>> > Here is a close to perfect xorg.conf, with composite/aiglx working: |
| 37 |
>> > |
| 38 |
>> > Section "ServerLayout" |
| 39 |
>> > Identifier "Layout0" |
| 40 |
>> > Screen 0 "aticonfig-Screen[0]-0" 0 0 |
| 41 |
>> > InputDevice "Keyboard0" "CoreKeyboard" |
| 42 |
>> > InputDevice "Mouse0" "CorePointer" |
| 43 |
>> > EndSection |
| 44 |
>> > |
| 45 |
>> > Section "Files" |
| 46 |
>> > EndSection |
| 47 |
>> > |
| 48 |
>> > Section "Module" |
| 49 |
>> > Load "evdev" |
| 50 |
>> > Load "v4l" |
| 51 |
>> > EndSection |
| 52 |
>> > |
| 53 |
>> > Section "InputDevice" |
| 54 |
>> > Identifier "Mouse0" |
| 55 |
>> > Driver "evdev" |
| 56 |
>> > Option "CorePointer" |
| 57 |
>> > Option "Name" "Logitech, Inc. MX610 Laser Cordless Mouse" |
| 58 |
>> > Option "evBits" "+1-2" |
| 59 |
>> > Option "keyBits" "~272-287" |
| 60 |
>> > Option "relBits" "~0-2 ~6 ~8" |
| 61 |
>> > EndSection |
| 62 |
>> |
| 63 |
>> this doesn't work anymore. for the use of evdev (2.0.5 if i remember |
| 64 |
>> well) with new hal versions |
| 65 |
>> you have to put these options inside |
| 66 |
>> /etc/hal/fdi/policy/10-x11-input.fdi or evedev would throw |
| 67 |
>> out warnings and errors and the devices would not have full |
| 68 |
>> functionality. the other option would |
| 69 |
>> be to put the ignore flag inside xorg.conf. i don't see you having |
| 70 |
>> this flag enabled and this |
| 71 |
>> leads me to think that you're not using the 1.5.x xorg-server. |
| 72 |
> |
| 73 |
> hal: |
| 74 |
> 0.5.11-r4 |
| 75 |
> evdev: |
| 76 |
> 2.1.0 |
| 77 |
> xorg-server: |
| 78 |
> 1.5.3 |
| 79 |
> |
| 80 |
> and for some reasons I have full functionality. |
| 81 |
> |
| 82 |
well, i've had some issues with this new way of configuration and spent some |
| 83 |
hours to fix up hal rules. either i have been very unlucky or you've been lucky. |
| 84 |
|
| 85 |
>> |
| 86 |
>> here's what colin guthrie said about this functionality in a |
| 87 |
>> discussion about xorg |
| 88 |
>> allowing non hal devices inside xorg.conf on the xorg mailing list: |
| 89 |
>> |
| 90 |
>> Yes. Some platforms that Xorg is on do not support hal. But that said, |
| 91 |
>> there are lots of things in Gnome/KDE that just don't work without HAL, |
| 92 |
>> so more and more you will need it for correct operation. Is there any |
| 93 |
>> reason you have a problem using hal? |
| 94 |
> |
| 95 |
> there are many reasons. Like hal breaking every version. hal config being a |
| 96 |
> pain in the ass. keyboard with non us layouts are almost unusable with hal - |
| 97 |
> except somebody hhacks a fdi together. |
| 98 |
> In short: hal is a POS and forcing it down everybody's throat does not help. |
| 99 |
> |
| 100 |
well, this discussion went one long on the mailing list. i've been |
| 101 |
following it a little |
| 102 |
and i tend to agree with xorg devs: if there's one way of handling |
| 103 |
devices outside |
| 104 |
or the configuration and with something that just recognizes it and |
| 105 |
automatically |
| 106 |
plugs it on without the user to do anything is far better. i agree |
| 107 |
that the layout config |
| 108 |
isn't working, or at least not in a simple way (there has been a long |
| 109 |
discussion about |
| 110 |
this issue too on the mailing list), but you alreay know that the |
| 111 |
keyboard config |
| 112 |
is set to us by default at x login. setting the right layout in your |
| 113 |
desktop env then |
| 114 |
fixes localization inside the desktop environment. |
| 115 |
|
| 116 |
>> |
| 117 |
>> At the moment, Xorg will automatically disabled kbd/mouse devices unless |
| 118 |
>> you set the AllowEmptyInput flag in the server. |
| 119 |
> |
| 120 |
> not for me. |
| 121 |
> |
| 122 |
i've had to force it manually to have it loading the options. maybe |
| 123 |
your's working |
| 124 |
because you don't use the other particular settings that a driver |
| 125 |
might have, like |
| 126 |
synaptics touchpad gestures. well, again, either you've been lucky or i've been |
| 127 |
unlucky. sometimes xorg is just doing strange things. |
| 128 |
|
| 129 |
>> |
| 130 |
>> if you use xaa you'll also most probably will need also the |
| 131 |
>> Option "XAANoOffscreenPixmaps" "True" |
| 132 |
>> |
| 133 |
>> to eliminate xaa corruption on the screen. i don't know if this issue |
| 134 |
>> has already |
| 135 |
> |
| 136 |
> there are no xaa corruption. |
| 137 |
> |
| 138 |
i'm glad to hear that the bug of the pixel corruption around the pointer and |
| 139 |
around moving windows has been fixed. as i've said when i was still using |
| 140 |
fglrx it had this corruption issue (sometimes also video was getting corrupted) |
| 141 |
that i've had to set the removal of offscreen pixmaps to true. i know it was |
| 142 |
happening mostly with igp boards, like the ones in laptops. |
| 143 |
|
| 144 |
>> >>you'll also need to enable texturedrender in your device section to |
| 145 |
>> >>enable the Xv. i've been not |
| 146 |
>> >>using fglrx for some months now (the xf86-video-ati is quite enough |
| 147 |
>> >>for my needs) and i don't |
| 148 |
>> >>remember all the hacks that are available for fglrx driver. look at |
| 149 |
>> >>the phoronix forums for more |
| 150 |
>> >>detailed configuration infos about fglrx. |
| 151 |
>> > |
| 152 |
>> > no, he does not. |
| 153 |
>> |
| 154 |
>> what does he do not? to add configuration options or to see phoronix |
| 155 |
>> forums?! is not very clear from your answer. |
| 156 |
> |
| 157 |
> he does not need to add that option. On installation the driver will put the |
| 158 |
> right option into amdpcsdb - and everything there overrides xorg.conf. |
| 159 |
> |
| 160 |
so there's now an override in the amd personal configuration file. at the time |
| 161 |
i've been using fglrx this file had just appeared. by the way, it |
| 162 |
seems that this |
| 163 |
file is being overwritten after each update. is it the one for the |
| 164 |
permanent fglrx |
| 165 |
configuration or is there another file for this? i think it's a little |
| 166 |
stupid to |
| 167 |
overwrite this file every time, removing old settings, so that the end user to |
| 168 |
have to run again the amd config center. |
| 169 |
|
| 170 |
-- |
| 171 |
dott. ing. beso |
Archives