Always set the palette registers even if the head is not connected. This will not do anything bad (if the display is disabled, it has no
effect), and since the test was a bit too constraining (I have an LVDS panel on head A), it would prevent setting colors for the CMAP modes. This finally gets the \n demo working from the original binary from BeOS, without any change. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@42523 a95241bf-73f2-0310-859d-f6bbb57e9c96
This commit is contained in:
@@ -1191,10 +1191,8 @@ intel_set_indexed_colors(uint count, uint8 first, uint8 *colors, uint32 flags)
|
||||
uint32 color = colors[0] << 16 | colors[1] << 8 | colors[2];
|
||||
colors += 3;
|
||||
|
||||
if (gInfo->head_mode & HEAD_MODE_A_ANALOG)
|
||||
write32(INTEL_DISPLAY_A_PALETTE + first * sizeof(uint32), color);
|
||||
if (gInfo->head_mode & HEAD_MODE_B_DIGITAL)
|
||||
write32(INTEL_DISPLAY_B_PALETTE + first * sizeof(uint32), color);
|
||||
write32(INTEL_DISPLAY_A_PALETTE + first * sizeof(uint32), color);
|
||||
write32(INTEL_DISPLAY_B_PALETTE + first * sizeof(uint32), color);
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user