On Thu, May 25, 2017 at 2:51 AM, Angus Ainslie <angus@xxxxxxxx> wrote: > On 2017-05-24 10:34, Boris Brezillon wrote: >> >> AHB BIST gate is actually controlled with bit 7. >> >> This bug was detected while trying to use the NAND controller which is >> using the DMA engine to transfer data to the NAND. >> Since the ahb_bist_clk gate bit conflicts with the ahb_dma_clk gate bit, >> the core was disabling the DMA engine clock as part of its 'disable >> unused clks' procedure, which was causing all DMA transfers to fail after >> this point. >> >> Fixes: 5e73761786d6 ("clk: sunxi-ng: Add sun5i CCU driver") >> Cc: stable@xxxxxxxxxxxxxxx >> Reported-by: Angus Ainslie <angus@xxxxxxxx>compatible = >> "nextthing,chip-pro", "nextthing,gr8"; >> Signed-off-by: Boris Brezillon <boris.brezillon@frecompatible = >> "nextthing,chip-pro", "nextthing,gr8";e-electrons.com> >> --- >> drivers/clk/sunxi-ng/ccu-sun5i.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/drivers/clk/sunxi-ng/ccu-sun5i.c >> b/drivers/clk/sunxi-ng/ccu-sun5i.c >> index 5c476f966a72..5372bf8be5e6 100644 >> --- a/drivers/clk/sunxi-ng/ccu-sun5i.c >> +++ b/drivers/clk/sunxi-ng/ccu-sun5i.c >> @@ -243,7 +243,7 @@ static SUNXI_CCU_GATE(ahb_ss_clk, "ahb-ss", >> "ahb", >> static SUNXI_CCU_GATE(ahb_dma_clk, "ahb-dma", "ahb", >> 0x060, BIT(6), 0); >> static SUNXI_CCU_GATE(ahb_bist_clk, "ahb-bist", "ahb", >> - 0x060, BIT(6), 0); >> + 0x060, BIT(7), 0); >> static SUNXI_CCU_GATE(ahb_mmc0_clk, "ahb-mmc0", "ahb", >> 0x060, BIT(8), 0); >> static SUNXI_CCU_GATE(ahb_mmc1_clk, "ahb-mmc1", "ahb", > > > > The patch works perfectly. Using that I was able to remove the > clk_ignore_unsed and CLK_IS_CRITICAL. > Is that a Tested-by? :) ChenYu