[prev in list] [next in list] [prev in thread] [next in thread]
List: freedesktop-igt-dev
Subject: Re: [igt-dev] [PATCH v2 1/2] igt_kms: add hotspot plane property
From: Albert Esteve <aesteve () redhat ! com>
Date: 2023-07-31 13:36:31
Message-ID: CADSE00JJTDDQ=nxjv0oFDMDYivAp6EE8WU0Q1B94-MfNUtL1TQ () mail ! gmail ! com
[Download RAW message or body]
On Mon, Jul 31, 2023 at 1:27 PM Albert Esteve <aesteve@redhat.com> wrote:
>
>
> On Mon, Jul 31, 2023 at 10:49 AM Modem, Bhanuprakash <
> bhanuprakash.modem@intel.com> wrote:
>
>>
>>
>> On Mon-31-07-2023 01:42 pm, Albert Esteve wrote:
>> >
>> >
>> >
>> > On Mon, Jul 31, 2023 at 8:30 AM Modem, Bhanuprakash
>> > <bhanuprakash.modem@intel.com <mailto:bhanuprakash.modem@intel.com>>
>> wrote:
>> >
>> > Hello,
>> >
>> > On Fri-28-07-2023 08:07 pm, Albert Esteve wrote:
>> > > Introduce LOCAL_DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT
>> > > capability definition until patched drm.h uapi header
>> > > is upstreamed and lands on IGT.
>> > >
>> > > CURSOR_PLANE_HOTSPOT capability is enabled
>> > > conditionally, and then tracked in the igt_display
>> > > struct.
>> > >
>> > > Add HOTSPOT_X and HOTSPOT_Y properties
>> > > to the atomic_plane_properties struct.
>> > >
>> > > Signed-off-by: Albert Esteve <aesteve@redhat.com
>> > <mailto:aesteve@redhat.com>>
>> > > ---
>> > > lib/igt_kms.c | 10 ++++++++++
>> > > lib/igt_kms.h | 11 ++++++++++-
>> > > 2 files changed, 20 insertions(+), 1 deletion(-)
>> > >
>> > > diff --git a/lib/igt_kms.c b/lib/igt_kms.c
>> > > index f2b0eed57..73119d0e5 100644
>> > > --- a/lib/igt_kms.c
>> > > +++ b/lib/igt_kms.c
>> > > @@ -601,6 +601,8 @@ const char * const
>> > igt_plane_prop_names[IGT_NUM_PLANE_PROPS] = {
>> > > [IGT_PLANE_CRTC_Y] = "CRTC_Y",
>> > > [IGT_PLANE_CRTC_W] = "CRTC_W",
>> > > [IGT_PLANE_CRTC_H] = "CRTC_H",
>> > > + [IGT_PLANE_HOTSPOT_X] = "HOTSPOT_X",
>> > > + [IGT_PLANE_HOTSPOT_Y] = "HOTSPOT_Y",
>> > > [IGT_PLANE_FB_ID] = "FB_ID",
>> > > [IGT_PLANE_CRTC_ID] = "CRTC_ID",
>> > > [IGT_PLANE_IN_FENCE_FD] = "IN_FENCE_FD",
>> > > @@ -2296,6 +2298,11 @@ static void igt_plane_reset(igt_plane_t
>> > *plane)
>> > > if (igt_plane_has_prop(plane, IGT_PLANE_SCALING_FILTER))
>> > > igt_plane_set_prop_enum(plane,
>> > IGT_PLANE_SCALING_FILTER, "Default");
>> > >
>> > > + if (igt_plane_has_prop(plane, IGT_PLANE_HOTSPOT_X))
>> > > + igt_plane_set_prop_value(plane,
>> > IGT_PLANE_HOTSPOT_X, 0);
>> > > + if (igt_plane_has_prop(plane, IGT_PLANE_HOTSPOT_Y))
>> > > + igt_plane_set_prop_value(plane,
>> > IGT_PLANE_HOTSPOT_Y, 0);
>> > > +
>> > > igt_plane_clear_prop_changed(plane, IGT_PLANE_IN_FENCE_FD);
>> > > plane->values[IGT_PLANE_IN_FENCE_FD] = ~0ULL;
>> > > plane->gem_handle = 0;
>> > > @@ -2679,6 +2686,9 @@ void igt_display_require(igt_display_t
>> > *display, int drm_fd)
>> > > drmSetClientCap(drm_fd, DRM_CLIENT_CAP_UNIVERSAL_PLANES,
>> 1);
>> > > if (drmSetClientCap(drm_fd, DRM_CLIENT_CAP_ATOMIC, 1) == 0)
>> > > display->is_atomic = 1;
>> > > +
>> > > + if (drmSetClientCap(drm_fd,
>> > LOCAL_DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT, 1) == 0)
>> > > + display->set_hotspots = 1;
>> >
>> > Any special reason to add this to display struct?
>> > IMHO, it is better to handle it at test level.
>> >
>> >
>> > Hi Bhanu,
>> >
>> > Mostly for keeping it consistent with the `is_atomic` capability.
>> >
>> > Are you suggesting moving the `drmSetClientCap` to the test?
>>
>> Yes.
>>
>> > I am ok with changing this, just trying to understand what would
>> > be the preferred way here.
>>
>> Either way (adding to display struct or checking at test level) is fine.
>>
>> Generally to maintain the re-usability (ex: checking for the same
>> capability multiple times) we used to cache the capabilities to display
>> struct. If multiple tests (or multiple places in your test) want to use
>> this hotspot capability, you can add it to the display struct.
>>
>
> ATM it is only this test. If there is a need for more tests,
> we can move capability setting back to display, but for the time being I
> will follow
> your advice and move it to the test.
>
> Something like...?
> igt_subtest("modeset-atomic-cursor-hotspot") {
> igt_require(drmSetClientCap(display.drm_fd,
> LOCAL_DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT, 1) == 0);
> modeset_atomic_cursor_hotspot(&display);
> igt_assert_eq(drmSetClientCap(display.drm_fd,
> LOCAL_DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT, 0), 0);
> }
>
>
Hi Bhanu,
As I was testing this, I realized that if I delay setting the capability,
the
`igt_display_require` function will not find the cursor planes as they get
disabled
with the kernel patch that I am testing:
https://www.spinics.net/lists/stable/msg667033.html
Specifically, we need to set the capability before:
plane_resources = drmModeGetPlaneResources(display->drm_fd);
... or else the test will be skipped.
Therefore, I need to keep it where it is now, and then I can also keep
storing
the capability in the display struct.
>
>> - Bhanu
>>
>> >
>> > Thanks!
>> >
>> >
>> > - Bhanu
>> >
>> > >
>> > > plane_resources =
>> drmModeGetPlaneResources(display->drm_fd);
>> > > igt_assert(plane_resources);
>> > > diff --git a/lib/igt_kms.h b/lib/igt_kms.h
>> > > index 1b6988c17..cf0a6d82b 100644
>> > > --- a/lib/igt_kms.h
>> > > +++ b/lib/igt_kms.h
>> > > @@ -41,6 +41,12 @@
>> > >
>> > > /* Low-level helpers with kmstest_ prefix */
>> > >
>> > > +/**
>> > > + * Clients which do set cursor hotspot and treat the cursor
>> plane
>> > > + * like a mouse cursor should set this property.
>> > > + */
>> > > +#define LOCAL_DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT 6
>> > > +
>> > > /**
>> > > * pipe:
>> > > * @PIPE_NONE: Invalid pipe, used for disconnecting a output
>> > from a pipe.
>> > > @@ -318,7 +324,9 @@ enum igt_atomic_plane_properties {
>> > > IGT_PLANE_ZPOS,
>> > > IGT_PLANE_FB_DAMAGE_CLIPS,
>> > > IGT_PLANE_SCALING_FILTER,
>> > > - IGT_NUM_PLANE_PROPS
>> > > + IGT_PLANE_HOTSPOT_X,
>> > > + IGT_PLANE_HOTSPOT_Y,
>> > > + IGT_NUM_PLANE_PROPS
>> > > };
>> > >
>> > > /**
>> > > @@ -448,6 +456,7 @@ struct igt_display {
>> > > igt_pipe_t *pipes;
>> > > bool has_cursor_plane;
>> > > bool is_atomic;
>> > > + bool set_hotspots;
>> > > bool first_commit;
>> > >
>> > > uint64_t *modifiers;
>> >
>>
>>
[Attachment #3 (text/html)]
<div dir="ltr"><div dir="ltr"><div><div dir="ltr" class="gmail_signature"><div \
dir="ltr"><br></div></div></div></div><br><div class="gmail_quote"><div dir="ltr" \
class="gmail_attr">On Mon, Jul 31, 2023 at 1:27 PM Albert Esteve <<a \
href="mailto:aesteve@redhat.com">aesteve@redhat.com</a>> \
wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px \
0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div \
dir="ltr"><div><div dir="ltr" class="gmail_signature"><div \
dir="ltr"><br></div></div></div></div><br><div class="gmail_quote"><div dir="ltr" \
class="gmail_attr">On Mon, Jul 31, 2023 at 10:49 AM Modem, Bhanuprakash <<a \
href="mailto:bhanuprakash.modem@intel.com" \
target="_blank">bhanuprakash.modem@intel.com</a>> wrote:<br></div><blockquote \
class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid \
rgb(204,204,204);padding-left:1ex"><br> <br>
On Mon-31-07-2023 01:42 pm, Albert Esteve wrote:<br>
> <br>
> <br>
> <br>
> On Mon, Jul 31, 2023 at 8:30 AM Modem, Bhanuprakash <br>
> <<a href="mailto:bhanuprakash.modem@intel.com" \
target="_blank">bhanuprakash.modem@intel.com</a> <mailto:<a \
href="mailto:bhanuprakash.modem@intel.com" \
target="_blank">bhanuprakash.modem@intel.com</a>>> wrote:<br> > <br>
> Hello,<br>
> <br>
> On Fri-28-07-2023 08:07 pm, Albert Esteve wrote:<br>
> > Introduce LOCAL_DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT<br>
> > capability definition until patched drm.h uapi header<br>
> > is upstreamed and lands on IGT.<br>
> ><br>
> > CURSOR_PLANE_HOTSPOT capability is enabled<br>
> > conditionally, and then tracked in the igt_display<br>
> > struct.<br>
> ><br>
> > Add HOTSPOT_X and HOTSPOT_Y properties<br>
> > to the atomic_plane_properties struct.<br>
> ><br>
> > Signed-off-by: Albert Esteve <<a \
href="mailto:aesteve@redhat.com" target="_blank">aesteve@redhat.com</a><br> > \
<mailto:<a href="mailto:aesteve@redhat.com" \
target="_blank">aesteve@redhat.com</a>>><br> > > ---<br>
> > lib/igt_kms.c | 10 ++++++++++<br>
> > lib/igt_kms.h | 11 ++++++++++-<br>
> > 2 files changed, 20 insertions(+), 1 deletion(-)<br>
> ><br>
> > diff --git a/lib/igt_kms.c b/lib/igt_kms.c<br>
> > index f2b0eed57..73119d0e5 100644<br>
> > --- a/lib/igt_kms.c<br>
> > +++ b/lib/igt_kms.c<br>
> > @@ -601,6 +601,8 @@ const char * const<br>
> igt_plane_prop_names[IGT_NUM_PLANE_PROPS] = {<br>
> > [IGT_PLANE_CRTC_Y] = "CRTC_Y",<br>
> > [IGT_PLANE_CRTC_W] = "CRTC_W",<br>
> > [IGT_PLANE_CRTC_H] = "CRTC_H",<br>
> > + [IGT_PLANE_HOTSPOT_X] = "HOTSPOT_X",<br>
> > + [IGT_PLANE_HOTSPOT_Y] = "HOTSPOT_Y",<br>
> > [IGT_PLANE_FB_ID] = "FB_ID",<br>
> > [IGT_PLANE_CRTC_ID] = "CRTC_ID",<br>
> > [IGT_PLANE_IN_FENCE_FD] = "IN_FENCE_FD",<br>
> > @@ -2296,6 +2298,11 @@ static void igt_plane_reset(igt_plane_t<br>
> *plane)<br>
> > if (igt_plane_has_prop(plane, \
IGT_PLANE_SCALING_FILTER))<br> > > \
igt_plane_set_prop_enum(plane,<br> > IGT_PLANE_SCALING_FILTER, \
"Default");<br> > ><br>
> > + if (igt_plane_has_prop(plane, IGT_PLANE_HOTSPOT_X))<br>
> > + igt_plane_set_prop_value(plane,<br>
> IGT_PLANE_HOTSPOT_X, 0);<br>
> > + if (igt_plane_has_prop(plane, IGT_PLANE_HOTSPOT_Y))<br>
> > + igt_plane_set_prop_value(plane,<br>
> IGT_PLANE_HOTSPOT_Y, 0);<br>
> > +<br>
> > igt_plane_clear_prop_changed(plane, \
IGT_PLANE_IN_FENCE_FD);<br> > > \
plane->values[IGT_PLANE_IN_FENCE_FD] = ~0ULL;<br> > > \
plane->gem_handle = 0;<br> > > @@ -2679,6 +2686,9 @@ void \
igt_display_require(igt_display_t<br> > *display, int drm_fd)<br>
> > drmSetClientCap(drm_fd, DRM_CLIENT_CAP_UNIVERSAL_PLANES, \
1);<br> > > if (drmSetClientCap(drm_fd, \
DRM_CLIENT_CAP_ATOMIC, 1) == 0)<br> > > \
display->is_atomic = 1;<br> > > +<br>
> > + if (drmSetClientCap(drm_fd,<br>
> LOCAL_DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT, 1) == 0)<br>
> > + display->set_hotspots = 1;<br>
> <br>
> Any special reason to add this to display struct?<br>
> IMHO, it is better to handle it at test level.<br>
> <br>
> <br>
> Hi Bhanu,<br>
> <br>
> Mostly for keeping it consistent with the `is_atomic` capability.<br>
> <br>
> Are you suggesting moving the `drmSetClientCap` to the test?<br>
<br>
Yes.<br>
<br>
> I am ok with changing this, just trying to understand what would<br>
> be the preferred way here.<br>
<br>
Either way (adding to display struct or checking at test level) is fine.<br>
<br>
Generally to maintain the re-usability (ex: checking for the same <br>
capability multiple times) we used to cache the capabilities to display <br>
struct. If multiple tests (or multiple places in your test) want to use <br>
this hotspot capability, you can add it to the display \
struct.<br></blockquote><div><br></div><div>ATM it is only this test. If there is a \
need for more tests,</div><div>we can move capability setting back to display, but \
for the time being I will follow</div><div>your advice and move it to the \
test.</div><div><br></div><div>Something \
like...?</div><div>igt_subtest("modeset-atomic-cursor-hotspot") \
{<br></div><div> igt_require(drmSetClientCap(display.drm_fd, \
LOCAL_DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT, 1) == 0);<br> \
modeset_atomic_cursor_hotspot(&display);<br> \
igt_assert_eq(drmSetClientCap(display.drm_fd, \
LOCAL_DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT, 0), 0);<br> }<br></div><div> \
</div></div></div></blockquote><div><br></div><div>Hi \
Bhanu,</div><div><br></div><div>As I was testing this, I realized that if I delay \
setting the capability, the</div><div>`igt_display_require` function will not find \
the cursor planes as they get disabled</div><div>with the kernel patch that I am \
testing: <a href="https://www.spinics.net/lists/stable/msg667033.html">https://www.spinics.net/lists/stable/msg667033.html</a></div><div><br></div><div>Specifically, \
we need to set the capability before:</div><div>plane_resources = \
drmModeGetPlaneResources(display->drm_fd);<br></div><div><br></div><div>... or \
else the test will be skipped.</div><div><br></div><div>Therefore, I need to keep it \
where it is now, and then I can also keep storing</div><div>the capability in the \
display struct.</div><div> </div><blockquote class="gmail_quote" style="margin:0px \
0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div \
dir="ltr"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px \
0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> <br>
- Bhanu<br>
<br>
> <br>
> Thanks!<br>
> <br>
> <br>
> - Bhanu<br>
> <br>
> ><br>
> > plane_resources = \
drmModeGetPlaneResources(display->drm_fd);<br> > > \
igt_assert(plane_resources);<br> > > diff --git a/lib/igt_kms.h \
b/lib/igt_kms.h<br> > > index 1b6988c17..cf0a6d82b 100644<br>
> > --- a/lib/igt_kms.h<br>
> > +++ b/lib/igt_kms.h<br>
> > @@ -41,6 +41,12 @@<br>
> ><br>
> > /* Low-level helpers with kmstest_ prefix */<br>
> ><br>
> > +/**<br>
> > + * Clients which do set cursor hotspot and treat the cursor \
plane<br> > > + * like a mouse cursor should set this property.<br>
> > + */<br>
> > +#define LOCAL_DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT 6<br>
> > +<br>
> > /**<br>
> > * pipe:<br>
> > * @PIPE_NONE: Invalid pipe, used for disconnecting a \
output<br> > from a pipe.<br>
> > @@ -318,7 +324,9 @@ enum igt_atomic_plane_properties {<br>
> > IGT_PLANE_ZPOS,<br>
> > IGT_PLANE_FB_DAMAGE_CLIPS,<br>
> > IGT_PLANE_SCALING_FILTER,<br>
> > - IGT_NUM_PLANE_PROPS<br>
> > + IGT_PLANE_HOTSPOT_X,<br>
> > + IGT_PLANE_HOTSPOT_Y,<br>
> > + IGT_NUM_PLANE_PROPS<br>
> > };<br>
> ><br>
> > /**<br>
> > @@ -448,6 +456,7 @@ struct igt_display {<br>
> > igt_pipe_t *pipes;<br>
> > bool has_cursor_plane;<br>
> > bool is_atomic;<br>
> > + bool set_hotspots;<br>
> > bool first_commit;<br>
> ><br>
> > uint64_t *modifiers;<br>
> <br>
<br>
</blockquote></div></div>
</blockquote></div></div>
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic