From: Gil Dekel <gildekel@google.com>
To: igt-dev@lists.freedesktop.org
Cc: Gil Dekel <gildekel@google.com>,
Swati Sharma <swati2.sharma@intel.com>,
Karthik B S <karthik.b.s@intel.com>,
Mark Yacoub <markyacoub@google.com>
Subject: [PATCH i-g-t v2 2/3] lib/igt_hdr: Parse DisplayID CTA data blocks
Date: Sat, 14 Nov 2026 19:19:03 -0500 [thread overview]
Message-ID: <20261115001917.2933254-3-gildekel@google.com> (raw)
In-Reply-To: <20261115001917.2933254-1-gildekel@google.com>
Some eDP OLED panels carry DisplayID 2.0 extensions with a CTA data
block (tag 0x81) containing an HDR Static Metadata Data Block instead
of a CTA-861 extension block. igt_is_edid_hdr() only inspects CTA-861
extensions, so it reports such panels as SDR. kms_hdr then runs the
negative invalid-hdr subtest on them, which fails because the panel
accepts the SMPTE ST 2084 EOTF.
Also walk DisplayID extensions and scan each CTA data block (tag 0x81)
with cta_data_blocks_have_hdr(), bounded by the DisplayID section
length before the trailing section checksum.
Extend lib/tests/igt_hdr with synthetic DisplayID CTA test cases for
HDR, SDR-only, and a CTA data block that overflows the DisplayID
section.
Cc: Swati Sharma <swati2.sharma@intel.com>
Cc: Karthik B S <karthik.b.s@intel.com>
Cc: Mark Yacoub <markyacoub@google.com>
Assisted-by: Claude:claude-opus-5.5
Signed-off-by: Gil Dekel <gildekel@google.com>
---
Notes:
v2: Apologies for the v1 bloat from agent-assisted drafting;
replace the raw panel EDID dump with a synthetic
make_displayid_cta_edid() helper.
lib/igt_hdr.c | 40 +++++++++++++++++++++++++++++++++++++--
lib/tests/igt_hdr.c | 46 +++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 84 insertions(+), 2 deletions(-)
diff --git a/lib/igt_hdr.c b/lib/igt_hdr.c
index 8c36a035e..2d2893588 100644
--- a/lib/igt_hdr.c
+++ b/lib/igt_hdr.c
@@ -27,6 +27,7 @@
#define CTA_DATA_BLOCK_OFFSET 4
#define HDR_STATIC_METADATA_BLOCK 0x06
#define USE_EXTENDED_TAG 0x07
+#define DISPLAYID_DATA_BLOCK_CTA 0x81
static bool cta_block(const char *edid_ext)
{
@@ -88,14 +89,45 @@ static bool cta_ext_has_hdr(const struct edid_cea *cea)
cta_data_blocks_have_hdr((const uint8_t *)cea->data, len);
}
+/*
+ * Some eDP OLED panels advertise HDR through a CTA data block (tag 0x81)
+ * inside a DisplayID extension rather than a CTA-861 extension block.
+ */
+static bool displayid_ext_has_hdr(const struct edid_dispid *dispid)
+{
+ /* The section checksum follows the data blocks inside data[]. */
+ int end = min_t(int, dispid->header.num_bytes,
+ sizeof(dispid->data) - 1);
+ const uint8_t *data = (const uint8_t *)dispid->data;
+ int i = 0;
+
+ while (i + (int)sizeof(struct dispid_block_header) <= end) {
+ const struct dispid_block_header *block =
+ (const struct dispid_block_header *)&data[i];
+ int payload = i + sizeof(*block);
+
+ if (payload + block->num_bytes > end)
+ break;
+
+ if (block->tag == DISPLAYID_DATA_BLOCK_CTA &&
+ cta_data_blocks_have_hdr(&data[payload], block->num_bytes))
+ return true;
+
+ i = payload + block->num_bytes;
+ }
+
+ return false;
+}
+
/**
* igt_is_edid_hdr:
* @data: raw EDID, the base block followed by its extension blocks
* @size: size of @data in bytes
*
* Looks for an HDR Static Metadata Block in the CTA extensions of an
- * EDID. A sink is considered HDR-capable when it advertises support for
- * Traditional Gamma HDR or SMPTE ST 2084 EOTFs. Never reads past @size.
+ * EDID, and in the CTA data blocks of its DisplayID extensions. A sink is
+ * considered HDR-capable when it advertises support for Traditional
+ * Gamma HDR or SMPTE ST 2084 EOTFs. Never reads past @size.
*
* Returns: true if the EDID advertises HDR support, false otherwise.
*/
@@ -116,6 +148,10 @@ bool igt_is_edid_hdr(const void *data, size_t size)
if (ext->tag == EDID_EXT_CEA && cta_ext_has_hdr(&ext->data.cea))
return true;
+
+ if (ext->tag == EDID_EXT_DISPLAYID &&
+ displayid_ext_has_hdr(&ext->data.dispid))
+ return true;
}
return false;
diff --git a/lib/tests/igt_hdr.c b/lib/tests/igt_hdr.c
index 628d4b2e6..c24f7c94b 100644
--- a/lib/tests/igt_hdr.c
+++ b/lib/tests/igt_hdr.c
@@ -33,6 +33,37 @@ static void make_cta_edid(uint8_t edid[2 * EDID_BLOCK_SIZE], uint8_t eotf)
cta[7] = 0x01; /* static metadata descriptor type 1 */
}
+/*
+ * Base block plus one DisplayID 2.0 extension carrying a non-CTA data
+ * block followed by a CTA data block (tag 0x81) whose HDR Static Metadata
+ * Data Block advertises @eotf.
+ */
+static void make_displayid_cta_edid(uint8_t edid[2 * EDID_BLOCK_SIZE],
+ uint8_t eotf, uint8_t cta_block_len)
+{
+ uint8_t *ext = edid + EDID_BLOCK_SIZE;
+
+ memset(edid, 0, 2 * EDID_BLOCK_SIZE);
+ memcpy(edid, edid_header, sizeof(edid_header));
+ edid[126] = 1; /* extension count */
+
+ ext[0] = 0x70; /* DisplayID extension tag */
+ ext[1] = 0x20; /* DisplayID 2.0 revision */
+ ext[2] = 0x0b; /* section payload size: 4-byte block + 7-byte CTA block */
+
+ /* Preceding non-CTA DisplayID data block (tag 0x20, 1 payload byte). */
+ ext[5] = 0x20;
+ ext[7] = 0x01;
+
+ /* DisplayID CTA data block (tag 0x81) containing a 4-byte HDR block. */
+ ext[9] = 0x81;
+ ext[11] = cta_block_len;
+ ext[12] = 0xe3; /* extended tag data block, 3 payload bytes */
+ ext[13] = 0x06; /* HDR Static Metadata Data Block */
+ ext[14] = eotf;
+ ext[15] = 0x01; /* static metadata descriptor type 1 */
+}
+
int igt_main()
{
uint8_t edid[2 * EDID_BLOCK_SIZE];
@@ -51,4 +82,19 @@ int igt_main()
make_cta_edid(edid, 0x05);
igt_assert(!igt_is_edid_hdr(edid, EDID_BLOCK_SIZE));
}
+
+ igt_subtest("displayid-cta-hdr") {
+ make_displayid_cta_edid(edid, 0x05, 4);
+ igt_assert(igt_is_edid_hdr(edid, sizeof(edid)));
+ }
+
+ igt_subtest("displayid-cta-sdr-only") {
+ make_displayid_cta_edid(edid, 0x01, 4);
+ igt_assert(!igt_is_edid_hdr(edid, sizeof(edid)));
+ }
+
+ igt_subtest("displayid-cta-block-overflow") {
+ make_displayid_cta_edid(edid, 0x05, 0x7f);
+ igt_assert(!igt_is_edid_hdr(edid, sizeof(edid)));
+ }
}
--
Gil Dekel, Software Engineer, Google / ChromeOS Display and Graphics
next prev parent reply other threads:[~2026-11-16 5:21 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-14 22:29 [PATCH i-g-t 0/3] lib/igt_hdr: Find HDR in DisplayID EDIDs Gil Dekel
2026-11-14 22:29 ` [PATCH i-g-t 1/3] lib/igt_hdr: Add a testable igt_is_edid_hdr() Gil Dekel
2026-11-14 22:29 ` [PATCH i-g-t 2/3] lib/igt_hdr: Parse DisplayID CTA data blocks Gil Dekel
2026-11-14 22:29 ` [PATCH i-g-t 3/3] tests/kms_hdr: Use /dev/kmsg in invalid-hdr Gil Dekel
2026-11-15 0:19 ` [PATCH i-g-t v2 0/3] lib/igt_hdr: Find HDR in DisplayID EDIDs Gil Dekel
2026-11-15 0:19 ` [PATCH i-g-t v2 1/3] lib/igt_hdr: Add a testable igt_is_edid_hdr() Gil Dekel
2026-11-15 0:19 ` Gil Dekel [this message]
2026-11-15 0:19 ` [PATCH i-g-t v2 3/3] tests/kms_hdr: Use /dev/kmsg in invalid-hdr Gil Dekel
2026-11-16 7:51 ` ✓ Xe.CI.BAT: success for lib/igt_hdr: Find HDR in DisplayID EDIDs Patchwork
2026-11-16 8:07 ` ✓ i915.CI.BAT: " Patchwork
2026-11-16 9:07 ` ✗ Xe.CI.FULL: failure " Patchwork
2026-11-16 9:40 ` [PATCH i-g-t 0/3] " Jani Nikula
2026-11-17 3:13 ` Gil Dekel
2026-11-16 11:26 ` ✗ i915.CI.Full: failure for " Patchwork
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261115001917.2933254-3-gildekel@google.com \
--to=gildekel@google.com \
--cc=igt-dev@lists.freedesktop.org \
--cc=karthik.b.s@intel.com \
--cc=markyacoub@google.com \
--cc=swati2.sharma@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox