[drm_hwcomposer] [PATCH] Take Connection state into account. (v2)

[Robert Foss]

Hey Mauro!

Thanks for the v2, I would like to merge this, but the commit message is a 
little bit wonky still :)

Let me clean it up for you, and if you're fine with me adding your S-o-B
I'll push it.

Also, if you want to avoid the slow mailing list back and forth, I would happily 
help out over IRC too. You can find me on freenode with the nick robertfoss.
#dri-devel is also a good channel for this kind of work.


On 1/6/18 12:59 AM, Mauro Rossi wrote:
> Porting of original commit 76fb87e675 of Jim Bish in android-ia master to fdo
> 
> Original commit message:
> "There are various places where we should be really taking connection
> state into account before querying the properties or assuming it
> as primary. This patch fixes them."
> 
> (v2) checks on connection state are applied for both internal and external
> connectors, in order to select the correct primary, as opposed to setting,
> independently from its state, the first connector as primary
> 
> This is essential to avoid following logcat errors on integrated and dedicated GPUs:
> 
> ... 2245  2245 E hwc-drm-resources: Could not find a suitable encoder/crtc for display 2
> ... 2245  2245 E hwc-drm-resources: Failed CreateDisplayPipe 56 with -19
> ... 2245  2245 E hwcomposer-drm: Can't initialize Drm object -19
> 
> Tested with i965 on Sandybridge and nouveau on GT120, GT610


This is what I would expect the commit message to look like:

Take Connection state into account

There are various places where we should be really taking connection
state into account before querying the properties or assuming it
as primary. This patch fixes them.

Checks on connection state are applied for both internal and external
connectors, in order to select the correct primary, as opposed to setting,
independently from its state, the first connector as primary.

This is essential to avoid following logcat errors on integrated and dedicated GPUs:

... 2245  2245 E hwc-drm-resources: Could not find a suitable encoder/crtc for 
display 2
... 2245  2245 E hwc-drm-resources: Failed CreateDisplayPipe 56 with -19
... 2245  2245 E hwcomposer-drm: Can't initialize Drm object -19

Tested with i965 on Sandybridge and nouveau on GT120, GT610

Signed-off-by: Jim Bish <jim.bish at intel.com>
Signed-off-by: Mauro Rossi <Mauro Rossi <issor.oruam at gmail.com>

> ---
>   drmresources.cpp | 9 +++++++--
>   1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/drmresources.cpp b/drmresources.cpp
> index 32dd376..d582cfe 100644
> --- a/drmresources.cpp
> +++ b/drmresources.cpp
> @@ -159,7 +159,7 @@ int DrmResources::Init() {
>   
>     // First look for primary amongst internal connectors
>     for (auto &conn : connectors_) {
> -    if (conn->internal() && !found_primary) {
> +    if (conn->state() == DRM_MODE_CONNECTED && conn->internal() && !found_primary) {
>         conn->set_display(0);
>         found_primary = true;
>       } else {
> @@ -170,7 +170,7 @@ int DrmResources::Init() {
>   
>     // Then look for primary amongst external connectors
>     for (auto &conn : connectors_) {
> -    if (conn->external() && !found_primary) {
> +    if (conn->state() == DRM_MODE_CONNECTED && conn->external() && !found_primary) {
>         conn->set_display(0);
>         found_primary = true;
>       }
> @@ -288,6 +288,11 @@ int DrmResources::TryEncoderForDisplay(int display, DrmEncoder *enc) {
>   
>   int DrmResources::CreateDisplayPipe(DrmConnector *connector) {
>     int display = connector->display();
> +
> +  // skip not connected
> +  if (connector->state() == DRM_MODE_DISCONNECTED)
> +    return 0;
> +
>     /* Try to use current setup first */
>     if (connector->encoder()) {
>       int ret = TryEncoderForDisplay(display, connector->encoder());
>