Skip to content

Commit

Permalink
Fix wrong assumptions in cea_for_each_detailed_block v2
Browse files Browse the repository at this point in the history
The current logic misunderstands the spec about CEA 18byte descriptors.
First, the spec doesn't state "detailed timing descriptors" but "18 byte
descriptors", so any data record could be stored, mixed timings and
other data, just as in the standard EDID.
Second, the lower four bit of byte 3 of the CEA record do not contain
the number of descriptors, but "the total number of DTDs defining native
formats in the whole EDID [...], starting with the first DTD in the DTD
list (which starts in the base EDID block)." A device can of course
support non-native formats.

As such the number can't be used to determine n, and the existing code
will filter non-timing 18byte descriptors anyway.

Signed-off-by: Christian Schmidt <schmidt@digadd.de>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
  • Loading branch information
Christian Schmidt authored and Dave Airlie committed Dec 20, 2011
1 parent a0ab734 commit 4966b2a
Showing 1 changed file with 2 additions and 17 deletions.
19 changes: 2 additions & 17 deletions drivers/gpu/drm/drm_edid.c
Original file line number Diff line number Diff line change
Expand Up @@ -508,25 +508,10 @@ static void
cea_for_each_detailed_block(u8 *ext, detailed_cb *cb, void *closure)
{
int i, n = 0;
u8 rev = ext[0x01], d = ext[0x02];
u8 d = ext[0x02];
u8 *det_base = ext + d;

switch (rev) {
case 0:
/* can't happen */
return;
case 1:
/* have to infer how many blocks we have, check pixel clock */
for (i = 0; i < 6; i++)
if (det_base[18*i] || det_base[18*i+1])
n++;
break;
default:
/* explicit count */
n = min(ext[0x03] & 0x0f, 6);
break;
}

n = (127 - d) / 18;
for (i = 0; i < n; i++)
cb((struct detailed_timing *)(det_base + 18 * i), closure);
}
Expand Down

0 comments on commit 4966b2a

Please sign in to comment.