erebion/coreboot
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

edid: Accept valid detail blocks without timing descriptor.

Browse source 
The detail block may contain timing descriptor, or other fields like monitor
descriptor, so we should return 1 in detailed_block function when a valid
structure is found, otherwise for any EDID containing monitor descriptor we will
see following error messages:

	EDID block does not conform at all!
		Detailed blocks filled with garbage

BRANCH=none
BUG=chrome-os-partner:25933
TEST=emerge-nyan coreboot chromeos-bootimage;
     Manually executed on a Nayn and not seeing error message like
      "Detailed blocks filled with garbage".
     Also tried EDID from following devices: Link, Peppy. Display panel
     initialization is still functional.

Change-Id: Ib4e91d648741e5b54a558d53a1152273c7341427
Signed-off-by: Hung-Te Lin <hungte@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/193002
This commit is contained in:
Hung-Te Lin 2014-04-03 19:59:37 +08:00 • committed by chrome-internal-fetch
commit a1f212d6aa
1 changed files with 11 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

22
src/lib/edid.c
View file

@ -208,7 +208,7 @@ detailed_block(struct edid *out, unsigned char *x, int in_extension)
* 0x0e is used by EPI: http://www.epi-standard.org/
*/
printk(BIOS_SPEW, "Manufacturer-specified data, tag %d\n", x[3]);
return 0;
return 1;
}
switch (x[3]) {
case 0x10:
@ -216,11 +216,11 @@ detailed_block(struct edid *out, unsigned char *x, int in_extension)
for (i = 5; i < 18; i++)
if (x[i] != 0x00)
has_valid_dummy_block = 0;
return 0;
return 1;
case 0xF7:
/* TODO */
printk(BIOS_SPEW, "Established timings III\n");
return 0;
return 1;
case 0xF8:
{
int valid_cvt = 1; /* just this block */
@ -232,26 +232,26 @@ detailed_block(struct edid *out, unsigned char *x, int in_extension)
for (i = 0; i < 4; i++)
valid_cvt &= detailed_cvt_descriptor(out, x + 6 + (i * 3), (i == 0));
has_valid_cvt &= valid_cvt;
return 0;
return 1;
}
case 0xF9:
/* TODO */
printk(BIOS_SPEW, "Color management data\n");
return 0;
return 1;
case 0xFA:
/* TODO */
printk(BIOS_SPEW, "More standard timings\n");
return 0;
return 1;
case 0xFB:
/* TODO */
printk(BIOS_SPEW, "Color point\n");
return 0;
return 1;
case 0xFC:
printk(BIOS_SPEW, "Monitor name: %s\n",
extract_string(x + 5,
&has_valid_string_termination,
13));
return 0;
return 1;
case 0xFD:
{
int h_max_offset = 0, h_min_offset = 0;
@ -396,7 +396,7 @@ detailed_block(struct edid *out, unsigned char *x, int in_extension)
* Slightly weird to return a global, but I've never seen any
* EDID block wth two range descriptors, so it's harmless.
*/
return 0;
return 1;
}
case 0xFE:
/*
@ -405,11 +405,11 @@ detailed_block(struct edid *out, unsigned char *x, int in_extension)
*/
printk(BIOS_SPEW, "ASCII string: %s\n",
extract_string(x + 5, &has_valid_string_termination, 13));
return 0;
return 1;
case 0xFF:
printk(BIOS_SPEW, "Serial number: %s\n",
extract_string(x + 5, &has_valid_string_termination, 13));
return 0;
return 1;
default:
printk(BIOS_SPEW, "Unknown monitor description type %d\n", x[3]);
return 0;