#61 black strike in 640x480

[Pinky]

in resolution 640x480 (only, 1024x768 and 800x600 work)
is top of screen "under" black strike. It is cca 1/3 of
scree height. via driver(0.1.33.2) from modular X.org work.

onboard VT8378 [S3 UniChrome] (rev 01)

[Luc Verhaegen]

Logged In: YES
user_id=220099

Attach a log of an X run with -logverbose 8 please.

[Pinky]

Logged In: YES
user_id=1206973

tha seme for 32,24,16 or 8 bpp

"1024x768" -work
"832x624" -1/10 top of scree black strike
"800x600" -work
"720x400" -1/3 black strike
"640x480" -1/3 black strike
"640x400" -1/3 black strike
"640x350" -1/3 black strike

[Luc Verhaegen]

Logged In: YES
user_id=220099

Log.

[Pinky]

Logged In: YES
user_id=1206973

Log file

[Pinky]

X log file

[Luc Verhaegen]

Logged In: YES
user_id=220099

OK.

First, try the cvs version of the driver, you are still
using 0.2.3. I've fixed a Vertical Blanking bug there a few
weeks ago.

If the issue remains,

Keep your mach64/rage as the primary VGA device; you've set
your BIOS to initialise the mach64 first, and i guess that
your unichrome being non-Primary (and thus seeing different
BIOS initialisation), is why a bit in VBlank is set that
shouldn't be set, and that we're not handling ourselves.

Then put "832x624" as the only mode in your Modes directive
(in the Display subsection of your Screen section). And add
Option "PrintVGARegs" "True" in your Device section.

This, together with -logverbose 8, should get me a full
register dump.

[Pinky]

Logged In: YES
user_id=1206973

in res. 1152x864 are present exactly 4 "black" lines.

[Luc Verhaegen]

Logged In: YES
user_id=220099

If you think that just stating "this resolution is broken"
helps, then you're wrong. If i had been able to reproduce
this myself, it would've been fixed already.

Do try reading my comments from time to time.

[Pinky]

Logged In: YES
user_id=1206973

sory I miss you previous message >????< :)

I cannot CVS checkout due tue firewall(all my own tunnels
are now closed), I need some time to go trought.

[Pinky]

Logged In: YES
user_id=1206973

bug is still present althought ATI card is removed...,
I can not checkout from CVS but nmap show cvs port open:

[user@somewhere ~]$ cvs
-d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/unichrome
login
Logging in to
:pserver:anonymous@cvs.sourceforge.net:2401/cvsroot/unichrome
CVS password:
cvs [login aborted]: end of file from server (consult above
messages if any)

[Pinky]

Logged In: YES
user_id=1206973

It work, thank You very much.

but i must do:
-----------------
--- xf86-video-unichrome-0.2.4.orig/src/via_video.c
2006-03-25 05:00:05.000000000 +0100
+++ xf86-video-unichrome-0.2.4/src/via_video.c 2006-03-30
20:47:18.000000000 +0200
@@ -344,11 +344,11 @@
blit.mem_addr = aligned_buf;
blit.mem_stride = blit.line_length;
/* not used anyway. */
-#if defined(VIA_DRMH_MAJOR) || defined(VIA_DRM_DRIVER_MAJOR)
+/*#if defined(VIA_DRMH_MAJOR) || defined(VIA_DRM_DRIVER_MAJOR)
blit.flags = 0;
#else
blit.bouncebuffer = 0;
-#endif
+#endif*/
blit.to_fb = 1;

/* shedule this copy */
------------------------------
othervise:
----------------
via_video.c: In function `ViaSwovCopyDMA':
via_video.c:348: error: structure has no member named `flags'
make: *** [via_video.lo] Error 1
---------------
and disable (comment)
#define UNWRAP_LIBC 1
int config.h
because:
---------
In file included from via_cursor.c:40:
/usr/include/string.h:293: error: conflicting types for
'xf86memmove'
/usr/include/string.h:44: error: previous declaration of
'xf86memmove' was here

[Luc Verhaegen]

Logged In: YES
user_id=220099

Thanks.