27 Apr, 2018

2 commits


24 Sep, 2016

1 commit


04 Dec, 2015

1 commit


30 Nov, 2015

1 commit

  • We need to access reg stp_rep9, but not stp_rep[(9 - 1) / 2].
    If using "__raw_writel(0, DI_STP_REP(disp, 9))", this will exceeds
    the size of stp_rep array.

    Signed-off-by: Peng Fan
    Acked-by: Liu Ying
    Cc: Stefano Babic
    Cc: Anatolij Gustschin

    Peng Fan
     

25 Nov, 2015

1 commit

  • Simplify if/else code, since if channel equals to MEM_BG_SYNC or
    MEM_FG_SYNC, we have value 5 for 'dc_chan'.

    Signed-off-by: Peng Fan
    Cc: Anatolij Gustschin
    Cc: Stefano Babic
    Reviewed-by: Stefano Babic

    Peng Fan
     

25 Oct, 2014

1 commit


16 Oct, 2014

2 commits

  • The ipu display insists on having a lower_margin smaller
    then 2. If this is not the case it will attempt to force
    it and adjust the pixclk accordingly. This multiplies pixclk
    in Hz with the width and height, since this is typically
    a * 10^7 * b * 10^2 * c * 10^2 this will overflow the
    uint_32 and make things even worse. Since this is a
    bootloader and the adjustment is neglectible, just force
    it to two and warn about it.

    Cc: Stefano Babic
    Signed-off-by: Jeroen Hofstee

    Jeroen Hofstee
     
  • - fix debug pixel clk display and add unit
    - fix some comments

    Cc: Stefano Babic
    Signed-off-by: Jeroen Hofstee

    Jeroen Hofstee
     

11 Aug, 2014

2 commits

  • Since rgb2ycbcr_coeff and friends are declared const, but assigned
    to a void pointer, clang will warn that the const is implicity casted
    away. If the pointer is changed to void const * gcc will warn when it
    is implicitly casted to a const int array. Just add a correctly
    typed pointer instead to prevent these casts and hence the warnings.

    Cc: Troy Kisky
    Cc: Stefano Babic
    Signed-off-by: Jeroen Hofstee

    Jeroen Hofstee
     
  • Instead of waiting for DC triple buffer to be cleared, this patch
    changes to wait for a relevant DP sync flow end interrupt to come
    when disabling sync BG flows. In this way, we align the implement
    to the freescale internal IPUv3 driver. After applying this patch,
    an uboot hang up issue at the arch_preboot_os() stage, where we
    disable a relevant ipu display channel, is not observed any more on
    some MX6DL platforms.

    Signed-off-by: Liu Ying

    Liu Ying
     

04 Dec, 2013

1 commit


30 Oct, 2013

1 commit

  • Currently the HDMI splash screen image quality on mx6solo does not show a
    very stable image.

    By comparing the IPU driver from U-boot with the one from FSL 4.1.0 BSP,
    we can see that there is an inverted logic for setting the DI_GEN_POL_CLK bit.

    >From FSL BSP [1] we have:

    if (!sig.clk_pol)
    di_gen |= DI_GEN_POLARITY_DISP_CLK;

    Applying the same logic into U-boot fixes the HDMI image stability.

    [1] git.freescale.com/git/cgit.cgi/imx/linux-2.6-imx.git/tree/drivers/mxc/ipu3/ipu_disp.c?h=imx_3.0.35_4.1.0

    Signed-off-by: Fabio Estevam
    Tested-by: Eric Nelson
    Acked-by: Eric Nelson
    Acked-by: Stefano Babic

    Fabio Estevam
     

24 Jul, 2013

1 commit


10 Jul, 2012

1 commit


02 Oct, 2011

1 commit

  • ipu_disp.c: In function ‘ipu_disp_set_global_alpha’:
    ipu_disp.c:1237:11: warning: variable ‘flow’ set but not used
    [-Wunused-but-set-variable]
    ipu_disp.c: In function ‘ipu_disp_set_color_key’:
    ipu_disp.c:1302:16: warning: variable ‘flow’ set but not used
    [-Wunused-but-set-variable]

    Signed-off-by: Marek Vasut
    Cc: Stefano Babic

    Marek Vasut
     

22 Oct, 2010

1 commit

  • The patch is a porting of the IPU Linux driver
    developed by Freescale to have framebuffer
    functionalities in u-boot. The port is based on
    kernel 2.6.31 commit cc4fe714041805997b601fe8e5dd585d8a99297f,
    as delivered by Freescale [i.MX BSP].
    Most features are dropped from the original driver and
    only LCD support is the goal of this porting.

    Signed-off-by: Stefano Babic

    Stefano Babic