On Fri 01 Jun 13:26 PDT 2018, Vikash Garodia wrote: > +static void venus_reset_hw(struct venus_core *core) > +{ > + void __iomem *reg_base = core->base; > + > + writel(0, reg_base + WRAPPER_FW_START_ADDR); > + writel(VENUS_FW_MEM_SIZE, reg_base + WRAPPER_FW_END_ADDR); > + writel(0, reg_base + WRAPPER_CPA_START_ADDR); > + writel(VENUS_FW_MEM_SIZE, reg_base + WRAPPER_CPA_END_ADDR); > + writel(0x0, reg_base + WRAPPER_CPU_CGC_DIS); > + writel(0x0, reg_base + WRAPPER_CPU_CLOCK_CONFIG); > + > + /* Make sure all register writes are committed. */ > + mb(); wmb() doesn't wait until the writes are completed, it simply ensures that any writes before it are performed before any writes after it. If you really want to ensure that these configs has hit the hardware before you sleep, read back the value of the WRAPPER_CPU_CLOCK_CONFIG register. > + > + /* > + * Need to wait 10 cycles of internal clocks before bringing ARM9 > + * out of reset. > + */ > + udelay(1); > + > + /* Bring Arm9 out of reset */ > + writel_relaxed(0, reg_base + WRAPPER_A9SS_SW_RESET); There's no harm in using writel() here... > +} Regards, Bjorn -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html