The driver tries to round up the specified timeout cycles to the next power of 2 value. This should be done defore updating timeout variable. Correcting this here. Reported-by: Nishanth Menon <nm@xxxxxx> Signed-off-by: Lokesh Vutla <lokeshvutla@xxxxxx> Acked-by: Santosh Shilimkar <santosh.shilimkar@xxxxxx> --- Changes since V1: Updated Changelog and $subject Added a comment describing timeout calculation. drivers/memory/emif.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/drivers/memory/emif.c b/drivers/memory/emif.c index df08736..508763c 100644 --- a/drivers/memory/emif.c +++ b/drivers/memory/emif.c @@ -728,13 +728,17 @@ static u32 get_pwr_mgmt_ctrl(u32 freq, struct emif_data *emif, u32 ip_rev) /* Timeout based on DDR frequency */ timeout = freq >= freq_threshold ? timeout_perf : timeout_pwr; - /* The value to be set in register is "log2(timeout) - 3" */ + /* + * The value to be set in register is "log2(timeout) - 3" + * if timeout < 16 load 0 in register + * if timeout is not a power of 2, round to next highest power of 2 + */ if (timeout < 16) { timeout = 0; } else { - timeout = __fls(timeout) - 3; if (timeout & (timeout - 1)) - timeout++; + timeout <<= 1; + timeout = __fls(timeout) - 3; } switch (lpmode) { -- 1.7.9.5 -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html