Allow a longer time for the EC to reboot. Also use a constant for the hash delay time, so it is clear what it is for.
Signed-off-by: Simon Glass sjg@chromium.org ---
drivers/misc/cros_ec.c | 24 +++++++++++++++++------- 1 file changed, 17 insertions(+), 7 deletions(-)
diff --git a/drivers/misc/cros_ec.c b/drivers/misc/cros_ec.c index 013c67e0464..ce5fa5bee35 100644 --- a/drivers/misc/cros_ec.c +++ b/drivers/misc/cros_ec.c @@ -44,6 +44,10 @@ enum { CROS_EC_CMD_TIMEOUT_MS = 5000, /* Timeout waiting for a synchronous hash to be recomputed */ CROS_EC_CMD_HASH_TIMEOUT_MS = 2000, + + /* Wait 10 ms between attempts to check if EC's hash is ready */ + CROS_EC_HASH_CHECK_DELAY_MS = 10, + };
#define INVALID_HCMD 0xFF @@ -502,9 +506,10 @@ static int cros_ec_wait_on_hash_done(struct udevice *dev,
start = get_timer(0); while (hash->status == EC_VBOOT_HASH_STATUS_BUSY) { - mdelay(50); /* Insert some reasonable delay */ + mdelay(CROS_EC_HASH_CHECK_DELAY_MS);
p->cmd = EC_VBOOT_HASH_GET; + if (ec_command(dev, EC_CMD_VBOOT_HASH, 0, p, sizeof(*p), hash, sizeof(*hash)) < 0) return -1; @@ -622,18 +627,23 @@ int cros_ec_reboot(struct udevice *dev, enum ec_reboot_cmd cmd, uint8_t flags) return -1;
if (!(flags & EC_REBOOT_FLAG_ON_AP_SHUTDOWN)) { + ulong start; + /* * EC reboot will take place immediately so delay to allow it * to complete. Note that some reboot types (EC_REBOOT_COLD) * will reboot the AP as well, in which case we won't actually * get to this point. */ - /* - * TODO(rspangler@chromium.org): Would be nice if we had a - * better way to determine when the reboot is complete. Could - * we poll a memory-mapped LPC value? - */ - udelay(50000); + mdelay(50); + start = get_timer(0); + while (cros_ec_hello(dev, NULL)) { + if (get_timer(start) > 3000) { + log_err("EC did not return from reboot\n"); + return -ETIMEDOUT; + } + mdelay(5); + } }
return 0;