11 Nov, 2013

2 commits


21 Oct, 2013

5 commits


22 Aug, 2013

5 commits


16 Aug, 2013

2 commits


15 Jul, 2013

1 commit

  • The correct muxing for emi_sel clock should be
    2b'00 - 396M PFD
    2b'01 - PLL3
    2b'10 - AXI clk root
    2b'11 - 352M PFD

    This patch corrects the muxing in the clock driver.

    Signed-off-by: Liu Ying
    Signed-off-by: Shawn Guo
    Acked-by: Dirk Behme

    Liu Ying
     

17 Jun, 2013

7 commits


03 Jun, 2013

1 commit


23 May, 2013

1 commit

  • As pll5_video_div has been introduced to represent the clock
    generated from post-divider for video.
    Instead of pll5_video, pll5_video_div should be proper root clock
    for ldb_di_sel.

    Signed-off-by: Jiada Wang
    Signed-off-by: Shawn Guo

    Jiada Wang
     

12 May, 2013

3 commits

  • There is no clock pll2_pfd9_720m. Instead it should be pll3_pfd0_720m.
    Fix the typo in gpu3d_shader_sels.

    Signed-off-by: Shawn Guo
    Acked-by: Dirk Behme

    Shawn Guo
     
  • According to the i.MX6 Dual/Quad technical reference manual
    (Figure 18-2. Clock Tree - Part 1) the MLB clock is directly
    feed by the AXI_CLK_ROOT. This is called 'axi' in our code.

    Note that the clock of the MLB IP block on the i.MX6 is completely
    independent of the PLL8 (MLB PLL). The MLB PLL isn't responsible
    for feeding the MLB IP block with a clock. Instead, it's used
    internally by the MLB module to sync the bus clock in case the MLB
    6-pin interface is enabled:

    MediaLB Control 0 Register, MLB150_MLBC0[5], MLBPEN:
    1 MediaLB 6-pin interface enabled. MLB PLL and MLB PHY is enabled in this case.

    I.e. the PLL8 MLB PLL has to be handled by the MLB driver and isn't needed
    for clocking the MLB module itself.

    Signed-off-by: Dirk Behme
    CC: Jiada Wang
    Signed-off-by: Shawn Guo

    Dirk Behme
     
  • The periph_clk2_sel mux can be set to pll3, osc/pll1_ref_clk, or osc/
    pll2_burn_in_clk. The periph2_clk2_sel mux can be set to pll3 or pll2.

    Signed-off-by: Philipp Zabel
    Signed-off-by: Shawn Guo

    Philipp Zabel
     

12 Apr, 2013

6 commits


09 Apr, 2013

2 commits

  • According to the recent i.MX6 Quad technical reference manual, mode 0x4 (100b)
    of the CCM_CS2DCR register (address 0x020C402C) bits [11-9] and [14-12] select
    the PLL3 clock, and not the PLL3 PFD1 540M clock. In our code, the PLL3 root
    clock is named 'pll3_usb_otg', select this instead of the 540M clock.

    Signed-off-by: Dirk Behme
    Signed-off-by: Shawn Guo

    Dirk Behme
     
  • While booting from device tree, imx6q used to provide twd clock lookup
    by calling clk_register_clkdev() in clock driver. However, the commit
    bd60345 (ARM: use device tree to get smp_twd clock) forces DT boot to
    look up the clock from device tree. It causes the failure below when
    twd driver tries to get the clock, and hence kernel has to calibrate the
    local timer frequency.

    smp_twd: clock not found -2
    ...
    Calibrating local timer... 396.13MHz.

    Fix the regression by providing twd clock lookup from device tree, and
    remove the unused twd clk_register_clkdev() call from clock driver.

    Signed-off-by: Shawn Guo

    Shawn Guo
     

11 Mar, 2013

1 commit

  • We always boot from PLL1, so let's have pll1_sys in the clks_init_on
    list to have clk prepare/enable use count match the hardware status,
    so that drivers managing pll1_sys like cpufreq can get the use count
    right from the start.

    Reported-by: Dirk Behme
    Signed-off-by: Shawn Guo

    Shawn Guo
     

20 Feb, 2013

1 commit


10 Feb, 2013

1 commit

  • This mxs usbphy is only needs to be on after system boots
    up, and software never needs to control it anymore.
    Meanwhile, usbphy's parent needs to be notified if usb
    is suspend or not. So we design below mxs usbphy usage:

    - usbphy1_gate and usbphy2_gate:
    Their parents are dummy clock, we only needs to enable
    it after system boots up.
    - usbphy1 and usbphy2
    Usage reserved bit for this clock, in that case, the refcount
    will be updated, but without hardware changing.

    Signed-off-by: Peter Chen
    Signed-off-by: Shawn Guo

    Peter Chen
     

30 Jan, 2013

1 commit


29 Jan, 2013

1 commit