On 9/30/19 7:13 PM, Breno Leitao wrote:
> There are some weird monitors that returns invalid data, as zeroed
> Horizontal/Vertical Active/Blanking.
>
> This causes edid-decode to crash with a divsion by error exception. This
> simple patch avoids so, checking for the divisor before proceeding.
>
> Signed-off-by: Breno Leitao <leitao@xxxxxxxxxx>
> ---
> edid-decode.c | 11 +++++++++++
> 1 file changed, 11 insertions(+)
>
> diff --git a/edid-decode.c b/edid-decode.c
> index 7442f8a..4b2cef8 100644
> --- a/edid-decode.c
> +++ b/edid-decode.c
> @@ -1022,6 +1022,17 @@ static int detailed_block(const unsigned char *x,
> int in_extension)
This should have been on one line, it looks like your mailer is wrapping lines.
> break;
> }
>
> + if ((ha + hbl) == 0 ||
> + (va + vbl) == 0) {
I'd use '!ha || !hbl || !va || !vbl' here.
> + printf("Invalid data. Refusing to continue.\n"
I'd say: "Invalid Detailing Timings:" here.
> + "Horizontal Active %4d\n"
> + "Horizontal Blanking %4d\n"
> + "Vertical Active %4d\n"
> + "Vertical Blanking %4d\n",
> + ha, hbl, va, vbl);
> + return 0;
> + }
> +
> pixclk_khz = (x[0] + (x[1] << 8)) * 10;
> refresh = (pixclk_khz * 1000) / ((ha + hbl) * (va + vbl));
> printf("Detailed mode: Clock %.3f MHz, %d mm x %d mm\n"
>
I'm happy to take this if you post a v2 with these changes!
Regards,
Hans
