1 |
-----BEGIN PGP SIGNED MESSAGE----- |
2 |
Hash: SHA1 |
3 |
|
4 |
On Wed, 10 Aug 2005, Donnie Berkholz wrote: |
5 |
|
6 |
> --[PinePGP]--------------------------------------------------[begin]-- |
7 |
> Ferris McCormick wrote: |
8 |
> | I wrote the note very fast, and it was not too clear. With sunffb, the |
9 |
> | kernel code for current 2.6.x (x > 6) is broken, and davem has taken dri |
10 |
> | support out of the xorg sunffb driver (to paraphrase him, you can't do |
11 |
> | both drm and correct font rendering). |
12 |
> |
13 |
> OK, it might be nice to bring alanh in on this conversation since he |
14 |
> recently fixed up the ffb DRI code, also to retest with the new code. |
15 |
> |
16 |
|
17 |
Probably. Kernel seems to be changing so quickly now that as soon as the |
18 |
ffb DRI code gets fixed, the kernel changes out from under it. So when we |
19 |
get to 2.6.13-rcxx, I have problems building it and have nothing to test |
20 |
it on. (I am really not interested in kernel-2.4.30+, because one of |
21 |
these days davem really will fix kernel-2.6.82 (or something) for sparc, |
22 |
and we can start burying the 2.4.xx series. I am speaking for myself, |
23 |
here, but I don't think sparc is hanging on to kernel-2.4.xx because we |
24 |
are in love with it.) |
25 |
|
26 |
> And as far as davem taking out DRI from the DDX code, that seems |
27 |
easy |
28 |
> enough to re-enable and test again -- just #if 0 changed back to XF86DRI. |
29 |
> |
30 |
> donnie@supernova |
31 |
> /usr/local/share/xorg/modular/driver/xf86-video-sunffb/src $ grep -i |
32 |
> xf86dri * |
33 |
> ffb_dri.c:#define _XF86DRI_SERVER_ |
34 |
> ffb_driver.c:/*#ifdef XF86DRI*/ |
35 |
> ffb_driver.c:#if 0 /*def XF86DRI*/ |
36 |
> ffb_driver.c:#if 0 /*def XF86DRI*/ |
37 |
> ffb_driver.c:#if 0 /*def XF86DRI*/ |
38 |
> ffb.h:#ifdef XF86DRI |
39 |
> ffb.h:#ifdef XF86DRI |
40 |
> ffb.h:#ifdef XF86DRI |
41 |
> ffb.h:#ifdef XF86DRI |
42 |
> |
43 |
|
44 |
Yes, it's easy enough to re-enable. I haven't tried it yet, though, |
45 |
because I have figured davem had a good reason to disable it, and I |
46 |
generally trust his knowledge on such matters. So discovering what |
47 |
that reason is has been pretty low priority. (I don't know if anyone |
48 |
else has tested or not. The question comes up every now and then, |
49 |
and so far my response has been (tongue in cheek) along the |
50 |
lines of your observation along with a request "If you try this, we'd all |
51 |
like to know how it goes.") |
52 |
|
53 |
> | With mach64, the problem on sparc has been the hardware itself, if I |
54 |
> | remember correctly. That is, kernel, xorg are fine, but the mach64 card |
55 |
> | used on U5/U10 is memory-deficient. Someone who has looked at it more |
56 |
> | recently than I have will correct this. (There is a mach64-based sparc |
57 |
> | frame buffer which should be OK for DRI, but I've never seen one. I'll |
58 |
> | chase it down in the framebuffer handbook a little later.) |
59 |
> |
60 |
> My u5 works just fine with DRI at 800x600 resolution, 16-bit color. Any |
61 |
> higher than this and you're right, not enough memory. |
62 |
|
63 |
That is new information for me; I had been led to believe otherwise, and |
64 |
I have not been in a position to verify for myself. Thanks. |
65 |
|
66 |
> |
67 |
> | Changing DRI_DIRS is exactly what I did. |
68 |
> |
69 |
> Not really... |
70 |
> |
71 |
> DRIVER_DIRS != DRI_DIRS |
72 |
> |
73 |
|
74 |
Yes. I answered you without reading carefully. |
75 |
|
76 |
> e.g., DRIVER_DIRS = dri, DRI_DIRS = mach64 ffb |
77 |
> |
78 |
|
79 |
You are correct, of course. Actually, there should probably be a |
80 |
corresponding dri driver for each sparc-enabled frame buffer, which makes |
81 |
it something like "mach64 ffb savage ..." |
82 |
|
83 |
The actual ebuild will have to be a bit more subtle than my immediate |
84 |
patch-to-enable-assembly+get-to-testable for the following reason: |
85 |
|
86 |
It always makes sense to enable glx (Mesa) whether there is DRI support or |
87 |
not; some applications can run adequately well using the Mesa-indirect |
88 |
approach, and some graphics cards --- e.g., Elite == afb --- don't allow |
89 |
dri at all. That is what (for sparc, at least) USE=glx should control. |
90 |
On some systems, however, like your U5+mach64, it is beneficial to build |
91 |
the xxxx_dri.so module. That is what the USE=dri flag (on sparc) |
92 |
should control. So, USE=dri without also USE=glx does not make any sense |
93 |
on sparc, but USE="glx -dri" is required for, say, Elite-graphics systems |
94 |
for which one does want libGL built (Like my SB1000+afb system, which is |
95 |
fast enough to do reasonable glx-like things (e.g., blender), but can't do |
96 |
dri at all). Further, there is no way for the ebuild to determine for |
97 |
itself just what case it is addressing. |
98 |
|
99 |
So, ultimately, the mesa ebuild should work as you have it if it is given |
100 |
USE="dri glx", but it should build sparc-specific modules. However, it it |
101 |
gets USE="-dri glx", it should arrange to build libGL stand-alone, because |
102 |
the user is saying in effect "I do want mesa/openGL installed, but I am |
103 |
unable to support DRI", and mesa can be built that way. This requires, I |
104 |
think, the ebuild to determine the HOSTCONFIG for sparc based on whether |
105 |
or not it has USE=dri, and to go on and set a sensible DRI_DIRS for the |
106 |
USE=dri case. In all cases, the CFLAGS, SPARC_ASM, etc. should be set up |
107 |
as they are now. |
108 |
|
109 |
I hope this isn't too incoherent; this has been a hectic day. |
110 |
Regards, |
111 |
Ferris |
112 |
|
113 |
- -- |
114 |
Ferris McCormick (P44646, MI) <fmccor@g.o> |
115 |
Developer, Gentoo Linux (sparc, devrel) |
116 |
-----BEGIN PGP SIGNATURE----- |
117 |
Version: GnuPG v1.4.1 (GNU/Linux) |
118 |
|
119 |
iD8DBQFC+ptFQa6M3+I///cRAja3AJ93r592HPu8k9riqsvhJEDGz+ZntACgjqOv |
120 |
Q13+BsRVaZ2v4xoa4Pvi/B0= |
121 |
=nCQQ |
122 |
-----END PGP SIGNATURE----- |
123 |
-- |
124 |
gentoo-dev@g.o mailing list |