Current VESA vendor-specific block parsing expects real block size to be
the same as the defined struct size, use real offsets in conditionals
instead to add struct fields in future commits.
Signed-off-by: Yaroslav Bolyukin <iam@lach.pw>
---
drivers/gpu/drm/drm_edid.c | 28 ++++++++++++----------------
1 file changed, 12 insertions(+), 16 deletions(-)
diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index 64f7a94dd9e4..a52fd6de9327 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -6544,7 +6544,7 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
if (oui(vesa->oui[0], vesa->oui[1], vesa->oui[2]) != VESA_IEEE_OUI)
return;
- if (sizeof(*vesa) != sizeof(*block) + block->num_bytes) {
+ if (block->num_bytes < 5) {
drm_dbg_kms(connector->dev,
"[CONNECTOR:%d:%s] Unexpected VESA vendor block size\n",
connector->base.id, connector->name);
@@ -6567,24 +6567,20 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
break;
}
- if (!info->mso_stream_count) {
- info->mso_pixel_overlap = 0;
- return;
- }
-
- info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
- if (info->mso_pixel_overlap > 8) {
+ if (info->mso_stream_count) {
+ info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
+ if (info->mso_pixel_overlap > 8) {
+ drm_dbg_kms(connector->dev,
+ "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
+ connector->base.id, connector->name,
+ info->mso_pixel_overlap);
+ info->mso_pixel_overlap = 8;
+ }
drm_dbg_kms(connector->dev,
- "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
+ "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
connector->base.id, connector->name,
- info->mso_pixel_overlap);
- info->mso_pixel_overlap = 8;
+ info->mso_stream_count, info->mso_pixel_overlap);
}
-
- drm_dbg_kms(connector->dev,
- "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
- connector->base.id, connector->name,
- info->mso_stream_count, info->mso_pixel_overlap);
}
static void drm_update_vesa_specific_block(struct drm_connector *connector,
--
2.51.2On Wed, 26 Nov 2025, Yaroslav Bolyukin <iam@lach.pw> wrote:
> Current VESA vendor-specific block parsing expects real block size to be
> the same as the defined struct size, use real offsets in conditionals
> instead to add struct fields in future commits.
>
> Signed-off-by: Yaroslav Bolyukin <iam@lach.pw>
I think this is something we want to backport, since MSO would break
with bigger vendor-specific blocks, and that leads to black screens on
MSO displays.
Cc: stable@vger.kernel.org
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
> ---
> drivers/gpu/drm/drm_edid.c | 28 ++++++++++++----------------
> 1 file changed, 12 insertions(+), 16 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
> index 64f7a94dd9e4..a52fd6de9327 100644
> --- a/drivers/gpu/drm/drm_edid.c
> +++ b/drivers/gpu/drm/drm_edid.c
> @@ -6544,7 +6544,7 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
> if (oui(vesa->oui[0], vesa->oui[1], vesa->oui[2]) != VESA_IEEE_OUI)
> return;
>
> - if (sizeof(*vesa) != sizeof(*block) + block->num_bytes) {
> + if (block->num_bytes < 5) {
> drm_dbg_kms(connector->dev,
> "[CONNECTOR:%d:%s] Unexpected VESA vendor block size\n",
> connector->base.id, connector->name);
> @@ -6567,24 +6567,20 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
> break;
> }
>
> - if (!info->mso_stream_count) {
> - info->mso_pixel_overlap = 0;
> - return;
> - }
> -
> - info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
> - if (info->mso_pixel_overlap > 8) {
> + if (info->mso_stream_count) {
> + info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
> + if (info->mso_pixel_overlap > 8) {
> + drm_dbg_kms(connector->dev,
> + "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
> + connector->base.id, connector->name,
> + info->mso_pixel_overlap);
> + info->mso_pixel_overlap = 8;
> + }
> drm_dbg_kms(connector->dev,
> - "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
> + "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
> connector->base.id, connector->name,
> - info->mso_pixel_overlap);
> - info->mso_pixel_overlap = 8;
> + info->mso_stream_count, info->mso_pixel_overlap);
> }
> -
> - drm_dbg_kms(connector->dev,
> - "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
> - connector->base.id, connector->name,
> - info->mso_stream_count, info->mso_pixel_overlap);
> }
>
> static void drm_update_vesa_specific_block(struct drm_connector *connector,
--
Jani Nikula, Intel
On 2025-11-26 10:13, Jani Nikula wrote:
> On Wed, 26 Nov 2025, Yaroslav Bolyukin <iam@lach.pw> wrote:
>> Current VESA vendor-specific block parsing expects real block size to be
>> the same as the defined struct size, use real offsets in conditionals
>> instead to add struct fields in future commits.
>>
>> Signed-off-by: Yaroslav Bolyukin <iam@lach.pw>
>
> I think this is something we want to backport, since MSO would break
> with bigger vendor-specific blocks, and that leads to black screens on
> MSO displays.
>
Not sure why would we want to backport that if we don't backport the
other changes, old kernels will just have the broken implementation,
which in reality affects almost no body, given that there were no bug
reports
> Cc: stable@vger.kernel.org
> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
>
>> ---
>> drivers/gpu/drm/drm_edid.c | 28 ++++++++++++----------------
>> 1 file changed, 12 insertions(+), 16 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
>> index 64f7a94dd9e4..a52fd6de9327 100644
>> --- a/drivers/gpu/drm/drm_edid.c
>> +++ b/drivers/gpu/drm/drm_edid.c
>> @@ -6544,7 +6544,7 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
>> if (oui(vesa->oui[0], vesa->oui[1], vesa->oui[2]) != VESA_IEEE_OUI)
>> return;
>>
>> - if (sizeof(*vesa) != sizeof(*block) + block->num_bytes) {
>> + if (block->num_bytes < 5) {
>> drm_dbg_kms(connector->dev,
>> "[CONNECTOR:%d:%s] Unexpected VESA vendor block size\n",
>> connector->base.id, connector->name);
>> @@ -6567,24 +6567,20 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
>> break;
>> }
>>
>> - if (!info->mso_stream_count) {
>> - info->mso_pixel_overlap = 0;
>> - return;
>> - }
>> -
>> - info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>> - if (info->mso_pixel_overlap > 8) {
>> + if (info->mso_stream_count) {
>> + info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>> + if (info->mso_pixel_overlap > 8) {
>> + drm_dbg_kms(connector->dev,
>> + "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
>> + connector->base.id, connector->name,
>> + info->mso_pixel_overlap);
>> + info->mso_pixel_overlap = 8;
>> + }
>> drm_dbg_kms(connector->dev,
>> - "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
>> + "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
>> connector->base.id, connector->name,
>> - info->mso_pixel_overlap);
>> - info->mso_pixel_overlap = 8;
>> + info->mso_stream_count, info->mso_pixel_overlap);
>> }
>> -
>> - drm_dbg_kms(connector->dev,
>> - "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
>> - connector->base.id, connector->name,
>> - info->mso_stream_count, info->mso_pixel_overlap);
>> }
>>
>> static void drm_update_vesa_specific_block(struct drm_connector *connector,
>
On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote: > On 2025-11-26 10:13, Jani Nikula wrote: >> On Wed, 26 Nov 2025, Yaroslav Bolyukin <iam@lach.pw> wrote: >>> Current VESA vendor-specific block parsing expects real block size to be >>> the same as the defined struct size, use real offsets in conditionals >>> instead to add struct fields in future commits. >>> >>> Signed-off-by: Yaroslav Bolyukin <iam@lach.pw> >> >> I think this is something we want to backport, since MSO would break >> with bigger vendor-specific blocks, and that leads to black screens on >> MSO displays. >> > > Not sure why would we want to backport that if we don't backport the > other changes, old kernels will just have the broken implementation, > which in reality affects almost no body, given that there were no bug > reports The failure mode is: Someone buys a new shiny laptop with eDP MSO, with bigger vendor block, and won't get a picture on screen. BR, Jani. -- Jani Nikula, Intel
On 2025-11-26 15:29, Jani Nikula wrote: > On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote: > The failure mode is: Someone buys a new shiny laptop with eDP MSO, with > bigger vendor block, and won't get a picture on screen. I understand that, I'm only confused because this does seem to be as suitable for backport as this fixed DSC bpp thing is: Someone buys a VR headset with fixed DSC bpp (Bigscreen Beyond/Bigscreen Beyond 2/Vive Pro 2), and the system is unable to detect them. Unless eDP MSO with vendor specific data block extended with unset fixed DSC bpp value is much more popular in the wild, that is. I have however failed to discover any devices which have this value present other than the three mentioned VR headsets. > BR, > Jani. > >
On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote: > On 2025-11-26 15:29, Jani Nikula wrote: >> On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote: >> The failure mode is: Someone buys a new shiny laptop with eDP MSO, with >> bigger vendor block, and won't get a picture on screen. > > I understand that, I'm only confused because this does seem to be as > suitable for backport as this fixed DSC bpp thing is: > > Someone buys a VR headset with fixed DSC bpp (Bigscreen Beyond/Bigscreen > Beyond 2/Vive Pro 2), and the system is unable to detect them. > > Unless eDP MSO with vendor specific data block extended with unset fixed > DSC bpp value is much more popular in the wild, that is. I have however > failed to discover any devices which have this value present other than > the three mentioned VR headsets. The difference is between an existing and enabled feature working vs. a new feature working. eDP MSO is expected to work. Also, IIUC you can still use the VR headsets albeit with a lower resolution. For eDP MSO you simply get nothing, and even beginning to debug the issue is problematic, since that is quite possibly the only display on the device. BR, Jani. -- Jani Nikula, Intel
On 2025-11-26 15:47, Jani Nikula wrote: > On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote: >> On 2025-11-26 15:29, Jani Nikula wrote: >>> On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote: >>> The failure mode is: Someone buys a new shiny laptop with eDP MSO, with >>> bigger vendor block, and won't get a picture on screen. >> >> I understand that, I'm only confused because this does seem to be as >> suitable for backport as this fixed DSC bpp thing is: >> >> Someone buys a VR headset with fixed DSC bpp (Bigscreen Beyond/Bigscreen >> Beyond 2/Vive Pro 2), and the system is unable to detect them. >> >> Unless eDP MSO with vendor specific data block extended with unset fixed >> DSC bpp value is much more popular in the wild, that is. I have however >> failed to discover any devices which have this value present other than >> the three mentioned VR headsets. > > The difference is between an existing and enabled feature working vs. a > new feature working. eDP MSO is expected to work. Also, IIUC you can > still use the VR headsets albeit with a lower resolution. For eDP MSO > you simply get nothing, and even beginning to debug the issue is > problematic, since that is quite possibly the only display on the > device. > You can use Vive Pro 2 with a lower resolution, but there is no lower resolution option present for BSB/BSB2, their support is always borked without this patch, and they are more popular in the linux community. > > BR, > Jani. > >
On Wed, 26 Nov 2025, Jani Nikula <jani.nikula@linux.intel.com> wrote:
> On Wed, 26 Nov 2025, Yaroslav Bolyukin <iam@lach.pw> wrote:
>> Current VESA vendor-specific block parsing expects real block size to be
>> the same as the defined struct size, use real offsets in conditionals
>> instead to add struct fields in future commits.
>>
>> Signed-off-by: Yaroslav Bolyukin <iam@lach.pw>
>
> I think this is something we want to backport, since MSO would break
> with bigger vendor-specific blocks, and that leads to black screens on
> MSO displays.
>
> Cc: stable@vger.kernel.org
> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Oops, I'll take that back.
>
>> ---
>> drivers/gpu/drm/drm_edid.c | 28 ++++++++++++----------------
>> 1 file changed, 12 insertions(+), 16 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
>> index 64f7a94dd9e4..a52fd6de9327 100644
>> --- a/drivers/gpu/drm/drm_edid.c
>> +++ b/drivers/gpu/drm/drm_edid.c
>> @@ -6544,7 +6544,7 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
>> if (oui(vesa->oui[0], vesa->oui[1], vesa->oui[2]) != VESA_IEEE_OUI)
>> return;
>>
>> - if (sizeof(*vesa) != sizeof(*block) + block->num_bytes) {
>> + if (block->num_bytes < 5) {
>> drm_dbg_kms(connector->dev,
>> "[CONNECTOR:%d:%s] Unexpected VESA vendor block size\n",
>> connector->base.id, connector->name);
>> @@ -6567,24 +6567,20 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
>> break;
>> }
>>
>> - if (!info->mso_stream_count) {
>> - info->mso_pixel_overlap = 0;
This is no longer cleared for !info->mso_stream_count.
Perhaps the code could be reorganized to handle it better.
>> - return;
>> - }
>> -
>> - info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>> - if (info->mso_pixel_overlap > 8) {
>> + if (info->mso_stream_count) {
>> + info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>> + if (info->mso_pixel_overlap > 8) {
>> + drm_dbg_kms(connector->dev,
>> + "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
>> + connector->base.id, connector->name,
>> + info->mso_pixel_overlap);
>> + info->mso_pixel_overlap = 8;
>> + }
>> drm_dbg_kms(connector->dev,
>> - "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
>> + "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
>> connector->base.id, connector->name,
>> - info->mso_pixel_overlap);
>> - info->mso_pixel_overlap = 8;
>> + info->mso_stream_count, info->mso_pixel_overlap);
>> }
>> -
>> - drm_dbg_kms(connector->dev,
>> - "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
>> - connector->base.id, connector->name,
>> - info->mso_stream_count, info->mso_pixel_overlap);
>> }
>>
>> static void drm_update_vesa_specific_block(struct drm_connector *connector,
--
Jani Nikula, Intel
On 2025-11-26 15:08, Jani Nikula wrote:
> On Wed, 26 Nov 2025, Jani Nikula <jani.nikula@linux.intel.com> wrote:
>> On Wed, 26 Nov 2025, Yaroslav Bolyukin <iam@lach.pw> wrote:
>>> Current VESA vendor-specific block parsing expects real block size to be
>>> the same as the defined struct size, use real offsets in conditionals
>>> instead to add struct fields in future commits.
>>>
>>> Signed-off-by: Yaroslav Bolyukin <iam@lach.pw>
>>
>> I think this is something we want to backport, since MSO would break
>> with bigger vendor-specific blocks, and that leads to black screens on
>> MSO displays.
>>
>> Cc: stable@vger.kernel.org
>> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
>
> Oops, I'll take that back.
>
>>
>>> ---
>>> drivers/gpu/drm/drm_edid.c | 28 ++++++++++++----------------
>>> 1 file changed, 12 insertions(+), 16 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
>>> index 64f7a94dd9e4..a52fd6de9327 100644
>>> --- a/drivers/gpu/drm/drm_edid.c
>>> +++ b/drivers/gpu/drm/drm_edid.c
>>> @@ -6544,7 +6544,7 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
>>> if (oui(vesa->oui[0], vesa->oui[1], vesa->oui[2]) != VESA_IEEE_OUI)
>>> return;
>>>
>>> - if (sizeof(*vesa) != sizeof(*block) + block->num_bytes) {
>>> + if (block->num_bytes < 5) {
>>> drm_dbg_kms(connector->dev,
>>> "[CONNECTOR:%d:%s] Unexpected VESA vendor block size\n",
>>> connector->base.id, connector->name);
>>> @@ -6567,24 +6567,20 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
>>> break;
>>> }
>>>
>>> - if (!info->mso_stream_count) {
>>> - info->mso_pixel_overlap = 0;
>
> This is no longer cleared for !info->mso_stream_count.
>
> Perhaps the code could be reorganized to handle it better.
It defaults to zero due to drm_reset_display_info()
>>> - return;
>>> - }
>>> -
>>> - info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>>> - if (info->mso_pixel_overlap > 8) {
>>> + if (info->mso_stream_count) {
>>> + info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>>> + if (info->mso_pixel_overlap > 8) {
>>> + drm_dbg_kms(connector->dev,
>>> + "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
>>> + connector->base.id, connector->name,
>>> + info->mso_pixel_overlap);
>>> + info->mso_pixel_overlap = 8;
>>> + }
>>> drm_dbg_kms(connector->dev,
>>> - "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
>>> + "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
>>> connector->base.id, connector->name,
>>> - info->mso_pixel_overlap);
>>> - info->mso_pixel_overlap = 8;
>>> + info->mso_stream_count, info->mso_pixel_overlap);
>>> }
>>> -
>>> - drm_dbg_kms(connector->dev,
>>> - "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
>>> - connector->base.id, connector->name,
>>> - info->mso_stream_count, info->mso_pixel_overlap);
>>> }
>>>
>>> static void drm_update_vesa_specific_block(struct drm_connector *connector,
>
On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote:
> On 2025-11-26 15:08, Jani Nikula wrote:
>> On Wed, 26 Nov 2025, Jani Nikula <jani.nikula@linux.intel.com> wrote:
>>> On Wed, 26 Nov 2025, Yaroslav Bolyukin <iam@lach.pw> wrote:
>>>> Current VESA vendor-specific block parsing expects real block size to be
>>>> the same as the defined struct size, use real offsets in conditionals
>>>> instead to add struct fields in future commits.
>>>>
>>>> Signed-off-by: Yaroslav Bolyukin <iam@lach.pw>
>>>
>>> I think this is something we want to backport, since MSO would break
>>> with bigger vendor-specific blocks, and that leads to black screens on
>>> MSO displays.
>>>
>>> Cc: stable@vger.kernel.org
>>> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
>>
>> Oops, I'll take that back.
>>
>>>
>>>> ---
>>>> drivers/gpu/drm/drm_edid.c | 28 ++++++++++++----------------
>>>> 1 file changed, 12 insertions(+), 16 deletions(-)
>>>>
>>>> diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
>>>> index 64f7a94dd9e4..a52fd6de9327 100644
>>>> --- a/drivers/gpu/drm/drm_edid.c
>>>> +++ b/drivers/gpu/drm/drm_edid.c
>>>> @@ -6544,7 +6544,7 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
>>>> if (oui(vesa->oui[0], vesa->oui[1], vesa->oui[2]) != VESA_IEEE_OUI)
>>>> return;
>>>>
>>>> - if (sizeof(*vesa) != sizeof(*block) + block->num_bytes) {
>>>> + if (block->num_bytes < 5) {
>>>> drm_dbg_kms(connector->dev,
>>>> "[CONNECTOR:%d:%s] Unexpected VESA vendor block size\n",
>>>> connector->base.id, connector->name);
>>>> @@ -6567,24 +6567,20 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
>>>> break;
>>>> }
>>>>
>>>> - if (!info->mso_stream_count) {
>>>> - info->mso_pixel_overlap = 0;
>>
>> This is no longer cleared for !info->mso_stream_count.
>>
>> Perhaps the code could be reorganized to handle it better.
>
> It defaults to zero due to drm_reset_display_info()
Yes, and the code above the context initializes it from the vendor
block.
>
>>>> - return;
>>>> - }
>>>> -
>>>> - info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>>>> - if (info->mso_pixel_overlap > 8) {
>>>> + if (info->mso_stream_count) {
>>>> + info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>>>> + if (info->mso_pixel_overlap > 8) {
>>>> + drm_dbg_kms(connector->dev,
>>>> + "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
>>>> + connector->base.id, connector->name,
>>>> + info->mso_pixel_overlap);
>>>> + info->mso_pixel_overlap = 8;
>>>> + }
>>>> drm_dbg_kms(connector->dev,
>>>> - "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
>>>> + "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
>>>> connector->base.id, connector->name,
>>>> - info->mso_pixel_overlap);
>>>> - info->mso_pixel_overlap = 8;
>>>> + info->mso_stream_count, info->mso_pixel_overlap);
>>>> }
>>>> -
>>>> - drm_dbg_kms(connector->dev,
>>>> - "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
>>>> - connector->base.id, connector->name,
>>>> - info->mso_stream_count, info->mso_pixel_overlap);
>>>> }
>>>>
>>>> static void drm_update_vesa_specific_block(struct drm_connector *connector,
>>
>
--
Jani Nikula, Intel
On 2025-11-26 15:59, Jani Nikula wrote:
> On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote:
>> On 2025-11-26 15:08, Jani Nikula wrote:
>>> On Wed, 26 Nov 2025, Jani Nikula <jani.nikula@linux.intel.com> wrote:
>>>> On Wed, 26 Nov 2025, Yaroslav Bolyukin <iam@lach.pw> wrote:
>>>>> Current VESA vendor-specific block parsing expects real block size to be
>>>>> the same as the defined struct size, use real offsets in conditionals
>>>>> instead to add struct fields in future commits.
>>>>>
>>>>> Signed-off-by: Yaroslav Bolyukin <iam@lach.pw>
>>>>
>>>> I think this is something we want to backport, since MSO would break
>>>> with bigger vendor-specific blocks, and that leads to black screens on
>>>> MSO displays.
>>>>
>>>> Cc: stable@vger.kernel.org
>>>> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
>>>
>>> Oops, I'll take that back.
>>>
>>>>
>>>>> ---
>>>>> drivers/gpu/drm/drm_edid.c | 28 ++++++++++++----------------
>>>>> 1 file changed, 12 insertions(+), 16 deletions(-)
>>>>>
>>>>> diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
>>>>> index 64f7a94dd9e4..a52fd6de9327 100644
>>>>> --- a/drivers/gpu/drm/drm_edid.c
>>>>> +++ b/drivers/gpu/drm/drm_edid.c
>>>>> @@ -6544,7 +6544,7 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
>>>>> if (oui(vesa->oui[0], vesa->oui[1], vesa->oui[2]) != VESA_IEEE_OUI)
>>>>> return;
>>>>>
>>>>> - if (sizeof(*vesa) != sizeof(*block) + block->num_bytes) {
>>>>> + if (block->num_bytes < 5) {
>>>>> drm_dbg_kms(connector->dev,
>>>>> "[CONNECTOR:%d:%s] Unexpected VESA vendor block size\n",
>>>>> connector->base.id, connector->name);
>>>>> @@ -6567,24 +6567,20 @@ static void drm_parse_vesa_specific_block(struct drm_connector *connector,
>>>>> break;
>>>>> }
>>>>>
>>>>> - if (!info->mso_stream_count) {
>>>>> - info->mso_pixel_overlap = 0;
>>>
>>> This is no longer cleared for !info->mso_stream_count.
>>>
>>> Perhaps the code could be reorganized to handle it better.
>>
>> It defaults to zero due to drm_reset_display_info()
>
> Yes, and the code above the context initializes it from the vendor
> block.
>
It is wrapped in `if (info->mso_stream_count)` so we get
/* From drm_reset_display_info */
info->mso_pixel_overlap = 0;
/* This part was removed, as we don't want the early return here anymore
due to extended structure length checks
if (!info->mso_stream_count) {
info->mso_pixel_overlap = 0;
return;
} */
/* This code is right below your comment */
if (info->mso_stream_count) {
info->mso_pixel_overlap = ...;
...
}
/* I can reinsert the value reset explicitly here, if that's what you
suggesting to do:
else {
info->mso_pixel_overlap = 0;
}*/
>
>>
>>>>> - return;
>>>>> - }
>>>>> -
>>>>> - info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>>>>> - if (info->mso_pixel_overlap > 8) {
>>>>> + if (info->mso_stream_count) {
>>>>> + info->mso_pixel_overlap = FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>>>>> + if (info->mso_pixel_overlap > 8) {
>>>>> + drm_dbg_kms(connector->dev,
>>>>> + "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
>>>>> + connector->base.id, connector->name,
>>>>> + info->mso_pixel_overlap);
>>>>> + info->mso_pixel_overlap = 8;
>>>>> + }
>>>>> drm_dbg_kms(connector->dev,
>>>>> - "[CONNECTOR:%d:%s] Reserved MSO pixel overlap value %u\n",
>>>>> + "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
>>>>> connector->base.id, connector->name,
>>>>> - info->mso_pixel_overlap);
>>>>> - info->mso_pixel_overlap = 8;
>>>>> + info->mso_stream_count, info->mso_pixel_overlap);
>>>>> }
>>>>> -
>>>>> - drm_dbg_kms(connector->dev,
>>>>> - "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap %u\n",
>>>>> - connector->base.id, connector->name,
>>>>> - info->mso_stream_count, info->mso_pixel_overlap);
>>>>> }
>>>>>
>>>>> static void drm_update_vesa_specific_block(struct drm_connector *connector,
>>>
>>
>
On 2025-11-26 16:06, Yaroslav wrote:
>
>
> On 2025-11-26 15:59, Jani Nikula wrote:
>> On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote:
>>> On 2025-11-26 15:08, Jani Nikula wrote:
>>>> On Wed, 26 Nov 2025, Jani Nikula <jani.nikula@linux.intel.com> wrote:
>>>>> On Wed, 26 Nov 2025, Yaroslav Bolyukin <iam@lach.pw> wrote:
>>>>>> Current VESA vendor-specific block parsing expects real block size
>>>>>> to be
>>>>>> the same as the defined struct size, use real offsets in conditionals
>>>>>> instead to add struct fields in future commits.
>>>>>>
>>>>>> Signed-off-by: Yaroslav Bolyukin <iam@lach.pw>
>>>>>
>>>>> I think this is something we want to backport, since MSO would break
>>>>> with bigger vendor-specific blocks, and that leads to black screens on
>>>>> MSO displays.
>>>>>
>>>>> Cc: stable@vger.kernel.org
>>>>> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
>>>>
>>>> Oops, I'll take that back.
>>>>
>>>>>
>>>>>> ---
>>>>>> drivers/gpu/drm/drm_edid.c | 28 ++++++++++++----------------
>>>>>> 1 file changed, 12 insertions(+), 16 deletions(-)
>>>>>>
>>>>>> diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
>>>>>> index 64f7a94dd9e4..a52fd6de9327 100644
>>>>>> --- a/drivers/gpu/drm/drm_edid.c
>>>>>> +++ b/drivers/gpu/drm/drm_edid.c
>>>>>> @@ -6544,7 +6544,7 @@ static void
>>>>>> drm_parse_vesa_specific_block(struct drm_connector *connector,
>>>>>> if (oui(vesa->oui[0], vesa->oui[1], vesa->oui[2]) !=
>>>>>> VESA_IEEE_OUI)
>>>>>> return;
>>>>>> - if (sizeof(*vesa) != sizeof(*block) + block->num_bytes) {
>>>>>> + if (block->num_bytes < 5) {
>>>>>> drm_dbg_kms(connector->dev,
>>>>>> "[CONNECTOR:%d:%s] Unexpected VESA vendor block
>>>>>> size\n",
>>>>>> connector->base.id, connector->name);
>>>>>> @@ -6567,24 +6567,20 @@ static void
>>>>>> drm_parse_vesa_specific_block(struct drm_connector *connector,
>>>>>> break;
>>>>>> }
>>>>>> - if (!info->mso_stream_count) {
>>>>>> - info->mso_pixel_overlap = 0;
>>>>
>>>> This is no longer cleared for !info->mso_stream_count.
>>>>
>>>> Perhaps the code could be reorganized to handle it better.
>>>
>>> It defaults to zero due to drm_reset_display_info()
>>
>> Yes, and the code above the context initializes it from the vendor
>> block.
>>
>
> It is wrapped in `if (info->mso_stream_count)` so we get
>
> /* From drm_reset_display_info */
> info->mso_pixel_overlap = 0;
>
> /* This part was removed, as we don't want the early return here anymore
> due to extended structure length checks
> if (!info->mso_stream_count) {
> info->mso_pixel_overlap = 0;
> return;
> } */
>
>
> /* This code is right below your comment */
> if (info->mso_stream_count) {
> info->mso_pixel_overlap = ...;
> ...
> }
>
> /* I can reinsert the value reset explicitly here, if that's what you
> suggesting to do:
>
> else {
> info->mso_pixel_overlap = 0;
> }*/
>
>
For reference, repository state at the point after this patch:
https://github.com/CertainLach/linux-1/blob/01b0097171ebe85ae124a7b535dce0b46a1dd8e2/drivers/gpu/drm/drm_edid.c#L6561-L6593
Commit:
https://github.com/CertainLach/linux-1/commit/01b0097171ebe85ae124a7b535dce0b46a1dd8e2
(Note that the fix for DP/EDP mistake is already here, but nothing else
was changed in the github version of this patch)
>
>>
>>>
>>>>>> - return;
>>>>>> - }
>>>>>> -
>>>>>> - info->mso_pixel_overlap =
>>>>>> FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>>>>>> - if (info->mso_pixel_overlap > 8) {
>>>>>> + if (info->mso_stream_count) {
>>>>>> + info->mso_pixel_overlap =
>>>>>> FIELD_GET(DISPLAYID_VESA_MSO_OVERLAP, vesa->mso);
>>>>>> + if (info->mso_pixel_overlap > 8) {
>>>>>> + drm_dbg_kms(connector->dev,
>>>>>> + "[CONNECTOR:%d:%s] Reserved MSO pixel overlap
>>>>>> value %u\n",
>>>>>> + connector->base.id, connector->name,
>>>>>> + info->mso_pixel_overlap);
>>>>>> + info->mso_pixel_overlap = 8;
>>>>>> + }
>>>>>> drm_dbg_kms(connector->dev,
>>>>>> - "[CONNECTOR:%d:%s] Reserved MSO pixel overlap
>>>>>> value %u\n",
>>>>>> + "[CONNECTOR:%d:%s] MSO stream count %u, pixel
>>>>>> overlap %u\n",
>>>>>> connector->base.id, connector->name,
>>>>>> - info->mso_pixel_overlap);
>>>>>> - info->mso_pixel_overlap = 8;
>>>>>> + info->mso_stream_count, info->mso_pixel_overlap);
>>>>>> }
>>>>>> -
>>>>>> - drm_dbg_kms(connector->dev,
>>>>>> - "[CONNECTOR:%d:%s] MSO stream count %u, pixel overlap
>>>>>> %u\n",
>>>>>> - connector->base.id, connector->name,
>>>>>> - info->mso_stream_count, info->mso_pixel_overlap);
>>>>>> }
>>>>>> static void drm_update_vesa_specific_block(struct drm_connector
>>>>>> *connector,
>>>>
>>>
>>
On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote: > On 2025-11-26 16:06, Yaroslav wrote: >> >> >> On 2025-11-26 15:59, Jani Nikula wrote: >>> On Wed, 26 Nov 2025, Yaroslav <iam@0la.ch> wrote: >>>> On 2025-11-26 15:08, Jani Nikula wrote: >>>>> On Wed, 26 Nov 2025, Jani Nikula <jani.nikula@linux.intel.com> wrote: >>>>>> On Wed, 26 Nov 2025, Yaroslav Bolyukin <iam@lach.pw> wrote: >>>>>>> Current VESA vendor-specific block parsing expects real block size >>>>>>> to be >>>>>>> the same as the defined struct size, use real offsets in conditionals >>>>>>> instead to add struct fields in future commits. >>>>>>> >>>>>>> Signed-off-by: Yaroslav Bolyukin <iam@lach.pw> >>>>>> >>>>>> I think this is something we want to backport, since MSO would break >>>>>> with bigger vendor-specific blocks, and that leads to black screens on >>>>>> MSO displays. >>>>>> >>>>>> Cc: stable@vger.kernel.org >>>>>> Reviewed-by: Jani Nikula <jani.nikula@intel.com> >>>>> >>>>> Oops, I'll take that back. Yeah, my bad, I'll take back me taking that back. Original Reviewed-by stands. Sorry for the noise. BR, Jani. -- Jani Nikula, Intel
© 2016 - 2025 Red Hat, Inc.