Re: [PATCH] hw:m25p80: Add STATE_STANDBY command state

[Francisco Iglesias]

Hi Dan,

On [2022 Jun 14] Tue 13:02:05, Dan Zhang wrote:
> HW normally will switch it to stand by mode when receive incorrect
> command.
> i.e. Macronix MX66L1G45G data sheet section 8 DEVICE OPERATION described
> ```
> 2. When an incorrect command is written to this device, it enters
> standby mode and stays in standby mode until the next CS# falling edge.
> In standby mode, This device's SO pin should be High-Z.
> ```
> Add STATE_STANDBY CMDState and let the device ignore all input and keep
> SO as HIZ (output 1)
> 
> Signed-off-by: Dan Zhang <dz4list@gmail.com>
> ---
> A usage of this new state can be aborting in HPM checking 
> or unknown command code received.
> 
>  hw/block/m25p80.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/hw/block/m25p80.c b/hw/block/m25p80.c
> index b6bd430a99..9f89773b11 100644
> --- a/hw/block/m25p80.c
> +++ b/hw/block/m25p80.c
> @@ -423,6 +423,7 @@ typedef enum {
>      STATE_COLLECTING_DATA,
>      STATE_COLLECTING_VAR_LEN_DATA,
>      STATE_READING_DATA,
> +    STATE_STANDBY,

Should we enter the state also? (Otherwise we will be adding code that is
unused).

>  } CMDState;
>  
>  typedef enum {
> @@ -1472,6 +1473,9 @@ static uint32_t m25p80_transfer8(SSIPeripheral *ss, 
> uint32_t tx)
>                            s->cur_addr, (uint8_t)tx);
>  
>      switch (s->state) {
> +    case STATE_STANDBY:
> +        r = 0xFFFFFFFF; /* StandBy state SO shall be HiZ */

0xFF should be enough here (since we are dealing with 8 bits, e.g.
m25p80_transfer8). More safe is probably to return 0 though and see this as if
a pulldown was connected to the line instead (this because r has been default
to 0 and was the most likely return value before in this situation). If you
would agree we can remove above line.

Thanks,
Best regards,
Francisco Iglesias

> +        break;
>  
>      case STATE_PAGE_PROGRAM:
>          trace_m25p80_page_program(s, s->cur_addr, (uint8_t)tx);
> -- 
> 2.34.3
>