9da505d1f9
The i.MX frame-buffer read operation should be faster for all configurations then drawing each individual character again in response to scroll events. The nonstandard fields allows to configure frame-buffer special options flags for different display configurations by board specific initialization code. One of such specific options is reversed order of pixels in each individual byte. i.MX frame-buffer seems to be designed for big-endian use first. The byte order is correctly configured for little-endian ordering, but if 1, 2 or 4 bits per pixel are used, pixels ordering is incompatible to Linux generic frame-buffer drawing functions. The patch "Allow generic BitBLT functions to work with swapped pixel order in bytes" introduces required functionality into FBDEV core. The pixels ordering selection has to be enabled at compile time CONFIG_FB_CFB_REV_PIXELS_IN_BYTE and for each display configuration which requires it by flag FB_NONSTD_REV_PIX_IN_B in "nonstd" field of info structure. This patch provides way for board specific code to select this option. Signed-off-by: Pavel Pisa <pisa@cmp.felk.cvut.cz> Signed-off-by: Antonino Daplas <adaplas@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> |
||
|---|---|---|
| .. | ||
| debug-macro.S | ||
| dma.h | ||
| entry-macro.S | ||
| gpio.h | ||
| hardware.h | ||
| imx-dma.h | ||
| imx-regs.h | ||
| imx-uart.h | ||
| imxfb.h | ||
| io.h | ||
| irqs.h | ||
| memory.h | ||
| mmc.h | ||
| mx1ads.h | ||
| spi_imx.h | ||
| system.h | ||
| timex.h | ||
| uncompress.h | ||
| vmalloc.h | ||