qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Vivek Kasireddy <vivek.kasireddy@intel.com>
To: qemu-devel@nongnu.org
Cc: "Vivek Kasireddy" <vivek.kasireddy@intel.com>,
	"Gerd Hoffmann" <kraxel@redhat.com>,
	"Marc-André Lureau" <marcandre.lureau@redhat.com>,
	"Dmitry Osipenko" <dmitry.osipenko@collabora.com>,
	"Frediano Ziglio" <freddy77@gmail.com>,
	"Dongwon Kim" <dongwon.kim@intel.com>
Subject: [PATCH v4 6/7] ui/spice: Create a new texture with linear layout when gl=on is enabled
Date: Wed, 14 May 2025 19:45:44 -0700	[thread overview]
Message-ID: <20250515024734.758335-7-vivek.kasireddy@intel.com> (raw)
In-Reply-To: <20250515024734.758335-1-vivek.kasireddy@intel.com>

Since most encoders/decoders (invoked by Spice) may not work with
tiled memory associated with a texture, we need to create another
texture that has linear memory layout and use that instead.

Note that, there does not seem to be a direct way to indicate to the
GL implementation that a texture's backing memory needs to be linear.
Instead, we have to do it in a roundabout way where we need to first
create a tiled texture and import that as a memory object to create
a new texture that has a linear memory layout.

Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Marc-André Lureau <marcandre.lureau@redhat.com>
Cc: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Cc: Frediano Ziglio <freddy77@gmail.com>
Cc: Dongwon Kim <dongwon.kim@intel.com>
Signed-off-by: Vivek Kasireddy <vivek.kasireddy@intel.com>
---
 ui/spice-display.c | 76 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 76 insertions(+)

diff --git a/ui/spice-display.c b/ui/spice-display.c
index ed91521ac2..fb56da4ab0 100644
--- a/ui/spice-display.c
+++ b/ui/spice-display.c
@@ -893,6 +893,75 @@ static void spice_gl_update(DisplayChangeListener *dcl,
     ssd->gl_updates++;
 }
 
+static bool spice_gl_replace_fd_texture(SimpleSpiceDisplay *ssd,
+                                        int *fds, uint64_t *modifier,
+                                        int *num_planes)
+{
+    uint32_t offsets[DMABUF_MAX_PLANES], strides[DMABUF_MAX_PLANES];
+    GLuint texture = 0;
+    int fourcc;
+    bool ret;
+
+    if (!remote_client) {
+        return true;
+    }
+
+    if (*modifier == DRM_FORMAT_MOD_LINEAR) {
+        return true;
+    }
+
+    if (*num_planes > 1) {
+        error_report("spice: cannot replace texture with multiple planes");
+        return false;
+    }
+
+    /*
+     * We really want to ensure that the memory layout of the texture
+     * is linear; otherwise, the encoder's output may show corruption.
+     */
+    if (!surface_gl_create_texture_from_fd(ssd->ds, fds[0], &texture)) {
+        error_report("spice: cannot create new texture from fd");
+        return false;
+    }
+
+    /*
+     * A successful return after glImportMemoryFdEXT() means that
+     * the ownership of fd has been passed to GL. In other words,
+     * the fd we got above should not be used anymore.
+     */
+    ret = egl_dmabuf_export_texture(texture,
+                                    fds,
+                                    (EGLint *)offsets,
+                                    (EGLint *)strides,
+                                    &fourcc,
+                                    num_planes,
+                                    modifier);
+    if (!ret) {
+        glDeleteTextures(1, &texture);
+
+        /*
+         * Since we couldn't export our newly create texture (or create,
+         * an fd associated with it) we need to backtrack and try to
+         * recreate the fd associated with the original texture.
+         */
+        ret = egl_dmabuf_export_texture(ssd->ds->texture,
+                                        fds,
+                                        (EGLint *)offsets,
+                                        (EGLint *)strides,
+                                        &fourcc,
+                                        num_planes,
+                                        modifier);
+        if (!ret) {
+            surface_gl_destroy_texture(ssd->gls, ssd->ds);
+            warn_report("spice: no texture available to display");
+        }
+    } else {
+        surface_gl_destroy_texture(ssd->gls, ssd->ds);
+        ssd->ds->texture = texture;
+    }
+    return ret;
+}
+
 static void spice_server_gl_scanout(QXLInstance *qxl,
                                     const int *fd,
                                     uint32_t width, uint32_t height,
@@ -917,6 +986,7 @@ static void spice_gl_switch(DisplayChangeListener *dcl,
                             struct DisplaySurface *new_surface)
 {
     SimpleSpiceDisplay *ssd = container_of(dcl, SimpleSpiceDisplay, dcl);
+    bool ret;
 
     if (ssd->ds) {
         surface_gl_destroy_texture(ssd->gls, ssd->ds);
@@ -939,6 +1009,12 @@ static void spice_gl_switch(DisplayChangeListener *dcl,
             return;
         }
 
+        ret = spice_gl_replace_fd_texture(ssd, fd, &modifier, &num_planes);
+        if (!ret) {
+            surface_gl_destroy_texture(ssd->gls, ssd->ds);
+            return;
+        }
+
         trace_qemu_spice_gl_surface(ssd->qxl.id,
                                     surface_width(ssd->ds),
                                     surface_height(ssd->ds),
-- 
2.49.0



  parent reply	other threads:[~2025-05-15  2:50 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-15  2:45 [PATCH v4 0/7] ui/spice: Enable gl=on option for non-local or remote clients Vivek Kasireddy
2025-05-15  2:45 ` [PATCH v4 1/7] ui/egl-helpers: Error check the fds in egl_dmabuf_export_texture() Vivek Kasireddy
2025-05-15  2:45 ` [PATCH v4 2/7] ui/spice: Add an option for users to provide a preferred codec Vivek Kasireddy
2025-05-15  2:45 ` [PATCH v4 3/7] ui/spice: Enable gl=on option for non-local or remote clients Vivek Kasireddy
2025-05-24 14:44   ` Marc-André Lureau
2025-05-15  2:45 ` [PATCH v4 4/7] ui/spice: Add an option to submit gl_draw requests at fixed rate Vivek Kasireddy
2025-05-24 14:47   ` Marc-André Lureau
2025-05-24 14:57   ` Marc-André Lureau
2025-05-27  4:22     ` Kasireddy, Vivek
2025-05-15  2:45 ` [PATCH v4 5/7] ui/console-gl: Add a helper to create a texture with linear memory layout Vivek Kasireddy
2025-05-25  2:44   ` Dmitry Osipenko
2025-05-26 13:06   ` Michael Scherle
2025-05-26 15:16     ` Michael Scherle
2025-05-27  4:20       ` Kasireddy, Vivek
2025-05-27 13:45         ` Michael Scherle
2025-05-15  2:45 ` Vivek Kasireddy [this message]
2025-05-15  2:45 ` [PATCH v4 7/7] ui/spice: Blit the scanout texture if its memory layout is not linear Vivek Kasireddy
2025-05-24 14:43 ` [PATCH v4 0/7] ui/spice: Enable gl=on option for non-local or remote clients Marc-André Lureau
2025-05-27  4:18   ` Kasireddy, Vivek

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=20250515024734.758335-7-vivek.kasireddy@intel.com \
    --to=vivek.kasireddy@intel.com \
    --cc=dmitry.osipenko@collabora.com \
    --cc=dongwon.kim@intel.com \
    --cc=freddy77@gmail.com \
    --cc=kraxel@redhat.com \
    --cc=marcandre.lureau@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /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;
as well as URLs for NNTP newsgroup(s).