The change to edid_checksum is needed because the DisplayID section has another
checksum inside for the actually used part of its 128-byte extension block. The
checksum in this case uses the same algorithm, but for a shorter block. Thus I
added a parameter to specify the size of the block.
I'll address the rest of your comments in a 3-rd version of this patch.
> On 15 Mar 2021, at 10:44, Gerd Hoffmann <kra...@redhat.com> wrote:
>
>> +typedef struct Timings {
>
>> +static void generate_timings(Timings *timings, uint32_t refresh_rate,
>> + uint32_t xres, uint32_t yres)
>
> Adding these should be splitted to a separate patch.
>
>> -static void edid_checksum(uint8_t *edid)
>> +static void edid_checksum(uint8_t *edid, size_t len)
>
> Why this change? Also a good candidate for a separate patch.
>
>> + if (size >= 384) {
>> + did = edid + 256;
>
> "if (size >= 384 && large_screen)" ?
> Also setting did should be next to setting dta.
>
>
> if (did) {
>> + dummy_displayid(did);
>
> init_displayid() ?
>
> Especially if we generate that only in case we actually have a large
> screen so we never have an empty extension section?
>
> take care,
> Gerd
>
