Commit Graph

16 Commits

Author SHA1 Message Date
Jonas Ådahl
7ddbcd1fd2 cogl/journal: Don't sometimes hold a ref on the framebuffer
d42f1873fc introduced a semi circular
reference between the CoglFramebuffer, and CoglJournal, where
CoglJournal would keep a reference on the CoglFramebuffer when there
were any entries in the journal log.

To avoid risking leaking these objects indefinitely, when freeing
objects without doing anything that triggered a flush, CoglFramebuffer
had a "filter" on cogl_object_unref() calls, which knew
about under what conditions CoglJournal had a reference to it. When it
could detect that there were only the journal itself holding such a
reference, it'd flush the journal, effectively releasing the reference
the journal held, thus freeing itself, as well as the journal.

When CoglFramebuffer was ported to be implemented using GObject instead
of CoglObject, this "filter" was missed, causing not only awkward but
infrequent leaks, but also situations where we'd flush journals when
only the journal itself held the last reference to the framebuffer,
meaning the journal would free the framebuffer, thus itself, in the
middle of flushing, causing memory corruption and crashes.

A way to detect this, by asserting on CoglObject reference count during
flush, is by adding the `g_assert()` as described below, which will
assert instead cause memory corruption.

void
_cogl_journal_flush (CoglJournal *journal
{
   ...
   _cogl_journal_discard (journal);
+  g_assert (journal->_parent.ref_count > 0);
   ...
}

Fix this by making CoglFramebuffer the owner of the journal, which it
already was, and remove any circle referencing that was there before, as
it is not needed given that the CoglFramebuffer pointer is guaranteed to
be valid for the lifetime of CoglJournal as the framebuffer is the owner
of the journal.

However, to not miss flushing before tearing down, which is important as
this flushes painting calls to the driver that is important for e.g.
using the result of those journal entries, flush the journal the first
time cogl_framebuffer_dispose() is called, before doing anything else.

This also adds a test case. Without having broken the circular
reference, the test would fail on g_assert_null (offscreen), as it would
have been "leaked" at this point, but the actual memory corruption would
be a result of the `cogl_texture_get_data()` call, which flushes the
framebuffer, and causes the 'mid-flush' destruction of the journal
described above. Note that the texture keeps track of dependent
framebuffers, but it does not hold any references to them.

Closes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1474
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1735>
2021-02-18 16:59:00 +00:00
Georges Basile Stavracas Neto
1f0e8fb23c cogl/tests: Remove euler test
Next commits, and this patchset in general, will make this patchset
obsolete, since it'll only test graphene types against each other.
If at all useful, the Euler test should be moved to graphene.

https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1439
2020-10-06 15:34:46 +00:00
Simon McVittie
59a2bff8e2 cogl tests: Normally skip tests that are not expected to succeed
If a test is not expected to succeed, then running it could be considered
to be a waste of resources, particularly if the failure might manifest
as an indefinite hang (see cogl!11), or if the test is likely to dump core
and trigger "expensive" crash-reporting mechanisms like systemd-coredump,
corekeeper, abrt or apport.

Skip the tests that are expected to fail. They can still be requested via
an environment variable, which can be set after fixing a bug to check which
tests are now passing.

Originally cogl!15, adapted for mutter's fork of cogl to use gboolean
instead of CoglBool.

https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1272

Signed-off-by: Simon McVittie <smcv@debian.org>
2020-06-02 20:15:26 +00:00
Simon McVittie
c3bf10d19a cogl test-premult: Don't free texture data until CoglBitmap is freed
According to the cogl_bitmap_new_for_data documentation, the data is not
copied, so the application must keep the buffer alive for the lifetime
of the CoglBitmap. Freeing it too early led to a use-after-free in the
cogl unit tests. With that fixed, the test passes, so remove the known
failure annotation.

This AddressSanitizer trace is from the original cogl, but the bug and
fix apply equally to mutter's fork of cogl:

==6223==ERROR: AddressSanitizer: heap-use-after-free on address 0x62100001a500 at pc 0x7f3e2d4e7f4e bp 0x7ffcd9c41f30 sp 0x7ffcd9c416e0
READ of size 4096 at 0x62100001a500 thread T0
    #0 0x7f3e2d4e7f4d  (/usr/lib/x86_64-linux-gnu/libasan.so.5+0x96f4d)
    #1 0x7f3e260c7f6b in util_copy_box ../src/gallium/auxiliary/util/u_surface.c:131
    #2 0x7f3e268c6c10 in u_default_texture_subdata ../src/gallium/auxiliary/util/u_transfer.c:67
    #3 0x7f3e26486459 in st_TexSubImage ../src/mesa/state_tracker/st_cb_texture.c:1480
    #4 0x7f3e26487029 in st_TexImage ../src/mesa/state_tracker/st_cb_texture.c:1709
    #5 0x7f3e26487029 in st_TexImage ../src/mesa/state_tracker/st_cb_texture.c:1691
    #6 0x7f3e2644bdba in teximage ../src/mesa/main/teximage.c:3105
    #7 0x7f3e2644bdba in teximage_err ../src/mesa/main/teximage.c:3132
    #8 0x7f3e2644d84f in _mesa_TexImage2D ../src/mesa/main/teximage.c:3170
    #9 0x7f3e2cd1f7df in _cogl_texture_driver_upload_to_gl driver/gl/gl/cogl-texture-driver-gl.c:347
    #10 0x7f3e2ccd441b in allocate_from_bitmap driver/gl/cogl-texture-2d-gl.c:255
    #11 0x7f3e2ccd441b in _cogl_texture_2d_gl_allocate driver/gl/cogl-texture-2d-gl.c:462
    #12 0x7f3e2ce3a6c0 in cogl_texture_allocate cogl/cogl-texture.c:1398
    #13 0x7f3e2ce3e116 in _cogl_texture_pre_paint cogl/cogl-texture.c:359
    #14 0x7f3e2cdee177 in _cogl_pipeline_layer_pre_paint cogl/cogl-pipeline-layer.c:864
    #15 0x7f3e2cd574af in _cogl_rectangles_validate_layer_cb cogl/cogl-primitives.c:542
    #16 0x7f3e2cdd742f in cogl_pipeline_foreach_layer cogl/cogl-pipeline.c:735
    #17 0x7f3e2cd5c8b0 in _cogl_framebuffer_draw_multitextured_rectangles cogl/cogl-primitives.c:658
    #18 0x7f3e2cd60152 in cogl_rectangle cogl/cogl-primitives.c:858
    #19 0x5570a71ed6a0 in check_texture tests/conform/test-premult.c:103
    #20 0x5570a71ed946 in test_premult tests/conform/test-premult.c:159
    #21 0x5570a71df0d6 in main tests/conform/test-conform-main.c:58
    #22 0x7f3e2bcd809a in __libc_start_main ../csu/libc-start.c:308
    #23 0x5570a71e0869 in _start (/home/smcv/src/debian/cogl/tests/conform/.libs/test-conformance+0x33869)

0x62100001a500 is located 0 bytes inside of 4096-byte region [0x62100001a500,0x62100001b500)
freed by thread T0 here:
    #0 0x7f3e2d5581d7 in __interceptor_free (/usr/lib/x86_64-linux-gnu/libasan.so.5+0x1071d7)
    #1 0x5570a71ed58b in make_texture tests/conform/test-premult.c:69

previously allocated by thread T0 here:
    #0 0x7f3e2d558588 in malloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0x107588)
    #1 0x7f3e2d384500 in g_malloc ../../../glib/gmem.c:99

This was originally cogl!12.

https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1274

Signed-off-by: Simon McVittie <smcv@debian.org>
2020-05-27 15:50:36 +01:00
Carlos Garnacho
a5294ce55f cogl: Remove CoglPath and the tesselator
This was barely used, and doesn't represent the way we want to
do 2D rendering.

https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1126
2020-04-08 11:38:48 +02:00
Georges Basile Stavracas Neto
16875340cb Remove CoglQuaternion
This is unused API, and there are no plans to actually
use it. Even if we want to use it in the future, we'll
be fully Graphene and won't need this API anyway.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/458
2019-10-16 11:38:13 +00:00
Adam Jackson
7e8a864992 cogl: Remove unused cogl-gles2 API
This was introduced in:

    commit 010d16f647
    Author: Robert Bragg <robert@linux.intel.com>
    Date:   Tue Mar 6 03:21:30 2012 +0000

        Adds initial GLES2 integration support

        This makes it possible to integrate existing GLES2 code with
        applications using Cogl as the rendering api.

That's maybe a reasonable thing for a standalone cogl to want, but our
cogl has only one consumer. So if we want additional rendering out of
our cogl layer, it makes more sense to just add that to cogl rather than
support clutter or mutter or the javascript bindings creating their own
GLES contexts.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/500
2019-08-16 06:35:35 +00:00
Adam Jackson
ce6acf9dca cogl: Remove rectangle-texture tests
https://gitlab.gnome.org/GNOME/mutter/merge_requests/546
2019-04-18 12:53:12 -04:00
Adam Jackson
28954e8271 cogl: Remove unused 3D texture support
We're not using this, and it's difficult to imagine we ever would.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/487
2019-04-18 01:26:32 +00:00
Adam Jackson
a94841abf1 cogl: Remove color write masks from the framebuffer and pipeline API
The only thing using this is its own tests, and it's difficult to
imagine a real use for it.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/468
2019-04-17 19:57:14 +00:00
Jonas Ådahl
982d135ace cogl: Add missing function declarations
In plenty of places a non-static function was defined but didn't have
the corresponding declaration. Fix this by adding them, or alternatively
making them static.
2019-01-22 18:31:52 +01:00
Jonas Ådahl
24d8d40deb cogl/tests: Mark failing test as known failure
This allows us to enable running the tests again. Not fixing the
failure is at least better than not running the tests at all.
2018-11-06 17:17:36 +01:00
Jonas Ådahl
98af9bb251 cogl: Remove notion of a "cogl" version
Mutters copy of cogl isn't a separate library with its own version, so
remove the old version macros.
2018-11-06 17:17:36 +01:00
Jonas Ådahl
fba7d8c79c cogl: Always build CoglPath support 2018-11-06 17:17:36 +01:00
Jonas Ådahl
bf71cb2e3c Don't use config.h in clutter and cogl
In cogl use cogl-config.h and in clutter use clutter-build-config.h. We
can't use clutter-config.h in clutter because its already used and
installed.

https://bugzilla.gnome.org/show_bug.cgi?id=768976
2016-07-20 14:23:48 +08:00
Rui Matos
2d2835f02a move everything into a cogl/ directory 2016-04-22 16:44:31 +02:00