Hi Punnaiah,
> +/**
> + * pl353_smc_set_cycles - Set memory timing parameters
> + * @dev: Pointer to the device struct
> + * @t0: t_rc read cycle time
> + * @t1: t_wc write cycle time
> + * @t2: t_rea/t_ceoe output enable assertion delay
> + * @t3: t_wp write enable deassertion delay
> + * @t4: t_clr/t_pc page cycle time
> + * @t5: t_ar/t_ta ID read time/turnaround time
> + * @t6: t_rr busy to RE timing
> + *
> + * Sets NAND chip specific timing parameters.
> + */
> +void pl353_smc_set_cycles(struct device *dev, u32 t0, u32 t1, u32 t2, u32 t3,
> + u32 t4, u32 t5, u32 t6)
> +{
> + struct pl353_smc_data *pl353_smc = dev_get_drvdata(dev);
> +
> + t0 &= PL353_SMC_SET_CYCLES_T0_MASK;
> + t1 = (t1 & PL353_SMC_SET_CYCLES_T1_MASK) <<
> + PL353_SMC_SET_CYCLES_T1_SHIFT;
> + t2 = (t2 & PL353_SMC_SET_CYCLES_T2_MASK) <<
> + PL353_SMC_SET_CYCLES_T2_SHIFT;
> + t3 = (t3 & PL353_SMC_SET_CYCLES_T3_MASK) <<
> + PL353_SMC_SET_CYCLES_T3_SHIFT;
> + t4 = (t4 & PL353_SMC_SET_CYCLES_T4_MASK) <<
> + PL353_SMC_SET_CYCLES_T4_SHIFT;
> + t5 = (t5 & PL353_SMC_SET_CYCLES_T5_MASK) <<
> + PL353_SMC_SET_CYCLES_T5_SHIFT;
> + t6 = (t6 & PL353_SMC_SET_CYCLES_T6_MASK) <<
> + PL353_SMC_SET_CYCLES_T6_SHIFT;
Right now, if the caller (i.e. the PL353 NAND driver) passes in a value for any
of the timing parameters that is too large, it will be truncated with the
corresponding mask and then saved, resulting in unexpected behavior. I saw
this when the PL353 NAND driver tried to set t_rc and t_wc to 17 cycles, which
does not fit in the 4-bit fields for those values -- it got clipped to 4 bits
and stored as 1 cycle, which made communication with the NAND chip fail and
time out.
It would be good to sanity-check each timing here, and rather than returning
void, return a return code indicating success or failure (e.g. -EINVAL for
invalid timings).
Note also that truncation isn't the only thing to check for here. The PL350
r2p1 datasheet lists a "Minimum permitted value" for each of these timing
values.
> +
> + t0 |= t1 | t2 | t3 | t4 | t5 | t6;
> +
> + writel(t0, pl353_smc->base + PL353_SMC_SET_CYCLES_OFFS);
> + writel(PL353_SMC_DC_UPT_NAND_REGS, pl353_smc->base +
> + PL353_SMC_DIRECT_CMD_OFFS);
> +}
> +EXPORT_SYMBOL_GPL(pl353_smc_set_cycles);
> +
Thanks,
Ben
--
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
