Hi Jeffrey,
On Fri, Mar 25, 2016 at 01:21:09PM +0800, jeffrey.lin wrote:
> +#define MAX_PACKET_SIZE 60
...
> +#define RAYDIUM_FW_PAGESIZE 128
...
> +
> +static int raydium_i2c_send(struct i2c_client *client,
> + u8 addr, u8 *data, size_t len)
> +{
> + u8 buf[MAX_PACKET_SIZE + 1];
> + int tries = 0;
> +
> + if (len > MAX_PACKET_SIZE)
> + return -EINVAL;
...
> +static int raydium_i2c_fw_write_page(struct i2c_client *client,
> + const void *page)
> +{
> + static const u8 ack_ok[] = { 0x55, 0xAA };
> + u8 buf[2];
> + int retry;
> + int error;
> +
> + for (retry = 0; retry < MAX_FW_UPDATE_RETRIES; retry++) {
> + error = raydium_i2c_send(client, CMD_BOOT_WRT,
> + (u8 *)page, RAYDIUM_FW_PAGESIZE);
> + if (error) {
> + dev_err(&client->dev,
> + "BLDR Write Page failed: %d\n", error);
> + continue;
> + }
Given the above definitions of MAX_PACKET_SIZE and
RAYDIUM_FW_PAGESIZE I do not believe that firmware update is working.
What is the biggest buffer that can be sent to the device? Maybe we
should allocate it dynamically?
I am also wondering about formatting command structure as byte sequence.
Would it be better to define it as:
struct raidium_cmd {
u32 bank;
int len;
u8 cmd[RAIDIUM_MAX_CMD_LEN];
}
What is the longest command that the controller supports?
(Do not resubmit the driver yet, let's discuss first).
Thanks.
--
Dmitry
