Ensure that file descriptors will be closed on exit
Subsystems (2)
PATCH 2/76 i915/gem_basic: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_basic.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
tests/i915/gem_basic.c +5 -3
@@ -25,7 +25,6 @@* */ -#include "igt.h" #include <unistd.h> #include <stdlib.h> #include <stdio.h>
@@ -37,7 +36,10 @@#include <sys/ioctl.h> #include "drm.h" +#include "i915/gem_create.h" +#include "igt.h" +#include "igt_types.h" IGT_TEST_DESCRIPTION ("Tests basic gem_create and gem_close IOCTLs" );
@@ -78,10 +80,10 @@ test_create_fd_close(int fd)close (fd); } -int fd; - igt_main { +igt_fd_t (fd); +igt_fixture fd =drm_open_driver (DRIVER_INTEL );
Signature · 1 line
2.37.3
PATCH 5/76 i915/gem_blits: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_blits.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/gem_blits.c +1 -0
@@ -852,5 +852,6 @@ igt_main igt_fixture {put_ahnd (device.ahnd); +close (device.fd); } }
Signature · 1 line
2.37.3
cover Ensure that file descriptors will be closed on exit
From: Mauro Carvalho Chehab <[email protected]>
GCC is usually smart enough to close devices on exit, but that doesn't work with the logic inside igt_skip(), as it uses longjmp, causing the code to go out of scope. Such scoped variables must be tracked in an outer scope to the skipping subtest.
Change the logic on IGT tests in order to avoid keeping device opened at the end of the test.
Chris Wilson (76):
lib: Introduce typed cleanups
i915/gem_basic: Close device before exit i915/gem_flink_basic: Close device before exit i915/gem_linear_blits: Close device before exit i915/gem_blits: Close device before exit i915/gem_close: Close device before exit i915/gem_create: Close device before exit i915/gem_ctx_engines: Close device before exit i915/gem_ctx_shared: Close device before exit i915/gem_exec_alignment: Close device before exit i915/gem_render_copy: Close device before exit i915/gem_render_copy_redux: Close device before exit i915/gem_media_fill: Close device before exit i915/gem_userptr_blits: Close device before exit i915/gem_workarounds: Close device before exit i915/gem_unref_active_buffers: Close device before exit i915/gem_exec_endless: Close device before exit i915/gem_request_retire: Close device before exit i915/gem_unfence_active_buffers: Close device before exit i915/i915_pciid: Close device before exit
kms_cursor_legacy: Close device before exit
kms_flip: Close device before exit
kms_pipe_crc_basic: Close device before exit
kms_psr: Close device before exit
kms_flip: Close device before exit
kms_force_connector_basic: Close device before exit
dumb_buffer: Close device before exit
kms_atomic: Close device before exit
kms_atomic_interruptible: Close device before exit
kms_atomic_transition: Close device before exit i915/kms_big_fb: Close device before exit i915/kms_ccs: Close device before exit
kms_color: Close device before exit
kms_cursor_crc: Close device before exit i915/kms_big_joiner: Close device before exit i915/kms_cdclk: Close device before exit i915/kms_fence_pin_leak: Close device before exit i915/kms_flip_scaled_crc: Close device before exit i915/kms_flip_tiling: Close device before exit i915/kms_mmap_write_crc: Close device before exit i915/kms_pipe_b_c_ivb: Close device before exit i915/kms_psr2_sf: Close device before exit i915/kms_psr2_su: Close device before exit i915/kms_pwrite_crc: Close device before exit
kms_flip_event_leak: Close device before exit
kms_hdr: Close device before exit
kms_invalid_mode: Close device before exit
kms_panel_fitting: Close device before exit
kms_plane: Close device before exit
kms_plane_alpha_blend: Close device before exit
kms_plane_cursor: Close device before exit
kms_plane_multiple: Close device before exit
kms_plane_scaling: Close device before exit
kms_properties: Close device before exit
kms_rotation_crc: Close device before exit
kms_sequence: Close device before exit
kms_universal_plane: Close device before exit
kms_vblank: Close device before exit
syncobj_wait: Close device before exit
syncobj_basic: Close device before exit
syncobj_timeline: Close device before exit
drm_import_export: Close device before exit
kms_dp_aux_dev: Close device before exit
kms_hdmi_inject: Close device before exit
kms_3d: Close device before exit
kms_scaling_modes: Close device before exit
kms_dither: Close device before exit
drm_read: Close device before exit
kms_cursor_edge_walk: Close device before exit
kms_plane_lowres: Close device before exit
kms_content_protection: Close device before exit
feature_discovery: Close device before exit i915/i915_pm_dc: Close device before exit
core_auth: Close(master) before exit i915/gem_exec_balancer: Close device before exit i915/gem_ctx_isolation: Close device before exit
lib/igt_core.c | 6 ++ lib/igt_core.h | 2 + lib/igt_types.c | 17 +++ lib/igt_types.h | 47 ++++++++ lib/meson.build | 1 + lib/tests/bad_subtest_type.c | 19 ++++ lib/tests/igt_types.c | 136 ++++++++++++++++++++++++ lib/tests/meson.build | 2 + tests/core_auth.c | 8 +- tests/drm_import_export.c | 5 + tests/drm_read.c | 4 +- tests/dumb_buffer.c | 4 + tests/feature_discovery.c | 5 +- tests/i915/gem_basic.c | 8 +- tests/i915/gem_blits.c | 1 + tests/i915/gem_close.c | 3 +- tests/i915/gem_create.c | 3 +- tests/i915/gem_ctx_engines.c | 3 +- tests/i915/gem_ctx_isolation.c | 3 +- tests/i915/gem_ctx_shared.c | 3 +- tests/i915/gem_exec_alignment.c | 3 +- tests/i915/gem_exec_balancer.c | 3 +- tests/i915/gem_exec_endless.c | 3 +- tests/i915/gem_flink_basic.c | 11 +- tests/i915/gem_linear_blits.c | 3 +- tests/i915/gem_media_fill.c | 1 + tests/i915/gem_render_copy.c | 1 + tests/i915/gem_render_copy_redux.c | 4 + tests/i915/gem_request_retire.c | 5 +- tests/i915/gem_unfence_active_buffers.c | 1 + tests/i915/gem_unref_active_buffers.c | 1 + tests/i915/gem_userptr_blits.c | 4 +- tests/i915/gem_workarounds.c | 3 +- tests/i915/i915_pciid.c | 2 + tests/i915/i915_pm_dc.c | 1 + tests/i915/kms_big_fb.c | 1 + tests/i915/kms_big_joiner.c | 1 + tests/i915/kms_busy.c | 1 + tests/i915/kms_ccs.c | 4 +- tests/i915/kms_cdclk.c | 1 + tests/i915/kms_fence_pin_leak.c | 1 + tests/i915/kms_flip_scaled_crc.c | 1 + tests/i915/kms_flip_tiling.c | 1 + tests/i915/kms_mmap_write_crc.c | 1 + tests/i915/kms_pipe_b_c_ivb.c | 1 + tests/i915/kms_psr.c | 1 + tests/i915/kms_psr2_sf.c | 1 + tests/i915/kms_psr2_su.c | 1 + tests/i915/kms_pwrite_crc.c | 1 + tests/kms_3d.c | 1 + tests/kms_atomic.c | 1 + tests/kms_atomic_interruptible.c | 1 + tests/kms_atomic_transition.c | 1 + tests/kms_color.c | 1 + tests/kms_content_protection.c | 1 + tests/kms_cursor_crc.c | 1 + tests/kms_cursor_edge_walk.c | 4 +- tests/kms_cursor_legacy.c | 1 + tests/kms_dither.c | 1 + tests/kms_dp_aux_dev.c | 1 + tests/kms_flip.c | 3 + tests/kms_flip_event_leak.c | 1 + tests/kms_force_connector_basic.c | 7 +- tests/kms_hdmi_inject.c | 1 + tests/kms_hdr.c | 1 + tests/kms_invalid_mode.c | 1 + tests/kms_panel_fitting.c | 4 +- tests/kms_pipe_crc_basic.c | 1 + tests/kms_plane.c | 1 + tests/kms_plane_alpha_blend.c | 1 + tests/kms_plane_cursor.c | 1 + tests/kms_plane_lowres.c | 4 +- tests/kms_plane_multiple.c | 4 +- tests/kms_plane_scaling.c | 4 +- tests/kms_properties.c | 1 + tests/kms_rotation_crc.c | 1 + tests/kms_scaling_modes.c | 4 +- tests/kms_sequence.c | 4 + tests/kms_universal_plane.c | 1 + tests/kms_vblank.c | 4 + tests/syncobj_basic.c | 4 + tests/syncobj_timeline.c | 4 + tests/syncobj_wait.c | 4 + 83 files changed, 385 insertions(+), 33 deletions(-) create mode 100644 lib/igt_types.c create mode 100644 lib/igt_types.h create mode 100644 lib/tests/bad_subtest_type.c create mode 100644 lib/tests/igt_types.c
Signature · 1 line
2.37.3
PATCH 1/76 lib: Introduce typed cleanups
From: Chris Wilson <[email protected]>
Start introducing standard types with automatic cleanup courtesy of gcc's __attribute__((cleanup)). As an example, we start with an fd that will automatically call close() on going out of scope, and crucially before atexit where we will want to check for resource leaks.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
lib/igt_core.c | 6 ++ lib/igt_core.h | 2 + lib/igt_types.c | 17 +++++ lib/igt_types.h | 47 ++++++++++++ lib/meson.build | 1 + lib/tests/bad_subtest_type.c | 19 +++++ lib/tests/igt_types.c | 136 +++++++++++++++++++++++++++++++++++ lib/tests/meson.build | 2 + 8 files changed, 230 insertions(+) create mode 100644 lib/igt_types.c create mode 100644 lib/igt_types.h create mode 100644 lib/tests/bad_subtest_type.c create mode 100644 lib/tests/igt_types.c
lib/igt_core.c +6 -0
@@ -624,6 +624,12 @@ uint64_t igt_nsec_elapsed(struct timespec *start) (uint64_t)NSEC_PER_SEC*(now.tv_sec - start->tv_sec)); } +void __igt_assert_in_outer_scope (void ) +{ +internal_assert (!in_subtest, +"must only be called outside of a subtest" ); +} +bool __igt_fixture (void ) { internal_assert(!in_fixture,
lib/igt_core.h +2 -0
@@ -135,6 +135,8 @@ struct _GKeyFile *igt_load_igtrc(void);*/ #define IGT_EXIT_ABORT 112 +void __igt_assert_in_outer_scope (void ); +bool __igt_fixture (void );void __igt_fixture_complete (void );__noreturn void __igt_fixture_end (void );
lib/igt_types.c new file +17 -0
@@ -0,0 +1,17 @@ +// SPDX-License-Identifier: MIT +/* +* Copyright © 2022 Intel Corporation +*/ + +#include <unistd.h> + +#include "igt_types.h" + +void igt_cleanup_fd (volatile int *fd) +{ +if (!fd || *fd <0 ) +return ; + +close (*fd); + *fd =-1 ; +}
lib/igt_types.h new file +47 -0
47 lines not on this page · raw message
lib/meson.build +1 -0
@@ -38,6 +38,7 @@ lib_sources = [ 'igt_sysrq.c', 'igt_taints.c', 'igt_thread.c', + 'igt_types.c', 'igt_vec.c', 'igt_vgem.c', 'igt_x86.c',
lib/tests/bad_subtest_type.c new file +19 -0
@@ -0,0 +1,19 @@ +// SPDX-License-Identifier: MIT +/* +* Copyright © 2022 Intel Corporation +*/ + +#include "igt_core.h" +#include "igt_types.h" + +igt_main +{ + igt_subtest("bad-scoped-variable") { +/* +* Not allowed to nest a scoped variable inside a subtest as +* we expect to longjmp out of the subtest on failure/skip +* and automatic cleanup is not invoked for such jmps. +*/ +igt_fd_t (f); + } +}
lib/tests/igt_types.c new file +136 -0
136 lines not on this page · raw message
lib/tests/meson.build +2 -0
@@ -19,10 +19,12 @@ lib_tests = [ 'igt_stats', 'igt_subtest_group', 'igt_thread', + 'igt_types', 'i915_perf_data_alignment', ] lib_fail_tests = [ + 'bad_subtest_type', 'igt_no_subtest', 'igt_simple_test_subtests', 'igt_timeout',
Signature · 1 line
2.37.3
PATCH 6/76 i915/gem_close: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_close.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tests/i915/gem_close.c +2 -1
@@ -24,6 +24,7 @@#include "i915/gem.h" #include "i915/gem_create.h" #include "igt.h" +#include "igt_types.h" static int batch_create (int fd) {
@@ -125,7 +126,7 @@ static void test_many_handles(int fd)igt_main { -int fd =-1 ; +igt_fd_t (fd ); igt_fixture {/* Create an flink requires DRM_AUTH */
Signature · 1 line
2.37.3
PATCH 7/76 i915/gem_create: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_create.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tests/i915/gem_create.c +2 -1
@@ -54,6 +54,7 @@#include "intel_chipset.h" #include "igt_aux.h" #include "igt_dummyload.h" +#include "igt_types.h" #include "igt_x86.h" #include "i915/gem.h" #include "i915/gem_create.h"
@@ -828,7 +829,7 @@ static void create_ext_cpu_access_big(int fd)igt_main { -int fd =-1 ; +igt_fd_t (fd ); igt_fixture { fd =drm_open_driver (DRIVER_INTEL );
Signature · 1 line
2.37.3
PATCH 9/76 i915/gem_ctx_shared: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_ctx_shared.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tests/i915/gem_ctx_shared.c +2 -1
@@ -42,6 +42,7 @@#include "i915/gem_engine_topology.h" #include "i915/gem_vm.h" #include "igt.h" +#include "igt_types.h" #include "igt_rand.h" #include "igt_vgem.h" #include "sw_sync.h"
@@ -987,7 +988,7 @@ igt_main {const struct intel_execution_engine2 *e;intel_ctx_cfg_t cfg; -int i915 =-1 ; +igt_fd_t (i915); igt_fixture { i915 =drm_open_driver (DRIVER_INTEL );
Signature · 1 line
2.37.3
PATCH 11/76 i915/gem_render_copy: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_render_copy.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/gem_render_copy.c +1 -0
@@ -879,5 +879,6 @@ igt_main_args("dac", NULL, help_str, opt_handler, NULL)igt_stop_hang_detector ();buf_ops_destroy (data.bops);igt_collection_destroy (set); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 13/76 i915/gem_media_fill: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_media_fill.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/gem_media_fill.c +1 -0
@@ -173,5 +173,6 @@ igt_main igt_fixture {igt_collection_destroy (set);igt_stop_hang_detector (); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 16/76 i915/gem_unref_active_buffers: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_unref_active_buffers.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/gem_unref_active_buffers.c +1 -0
@@ -121,4 +121,5 @@ igt_simple_mainigt_spin_free (i915, spin);put_ahnd (ahnd); +close (i915); }
Signature · 1 line
2.37.3
PATCH 17/76 i915/gem_exec_endless: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_exec_endless.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tests/i915/gem_exec_endless.c +2 -1
@@ -28,6 +28,7 @@#include "igt.h" #include "igt_device.h" #include "igt_sysfs.h" +#include "igt_types.h" #include "sw_sync.h" #define MAX_ENGINES 64
@@ -341,7 +342,7 @@ static void unpin_rps(int sysfs) igt_main {const struct intel_execution_engine2 *e; -int i915 =-1 ; +igt_fd_t (i915);igt_skip_on_simulation ();
Signature · 1 line
2.37.3
PATCH 18/76 i915/gem_request_retire: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_request_retire.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
tests/i915/gem_request_retire.c +3 -2
@@ -48,6 +48,7 @@#include "i915/gem.h" #include "igt.h" +#include "igt_types.h" IGT_TEST_DESCRIPTION ("Collection of tests targeting request retirement code" " paths." );
@@ -103,10 +104,10 @@ test_retire_vma_not_inactive(int fd)put_ahnd (ahnd); } -int fd; -igt_main { +igt_fd_t (fd ); + igt_fixture { fd =drm_open_driver (DRIVER_INTEL );igt_require_gem (fd);
Signature · 1 line
2.37.3
PATCH 19/76 i915/gem_unfence_active_buffers: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_unfence_active_buffers.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/gem_unfence_active_buffers.c +1 -0
@@ -99,4 +99,5 @@ igt_simple_mainigt_spin_free (i915, spin);put_ahnd (ahnd); +close (i915); }
Signature · 1 line
2.37.3
PATCH 22/76 kms_flip: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_flip.c | 3 +++ 1 file changed, 3 insertions(+)
tests/kms_flip.c +3 -0
@@ -1877,4 +1877,7 @@ igt_main_args("e", NULL, help_str, opt_handler, NULL)run_pair (tests[i].duration, tests[i].flags); }igt_stop_signal_helper (); + +igt_fixture +close (drm_fd ); }
Signature · 1 line
2.37.3
PATCH 23/76 kms_pipe_crc_basic: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_pipe_crc_basic.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_pipe_crc_basic.c +1 -0
@@ -362,5 +362,6 @@ igt_main igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 27/76 dumb_buffer: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/dumb_buffer.c | 4 ++++ 1 file changed, 4 insertions(+)
tests/dumb_buffer.c +4 -0
@@ -392,4 +392,8 @@ igt_main igt_subtest("create-clear")always_clear (fd,30 ); + + igt_fixture { +close (fd); + } }
Signature · 1 line
2.37.3
PATCH 31/76 i915/kms_big_fb: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_big_fb.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_big_fb.c +1 -0
@@ -991,5 +991,6 @@ igt_main igt_fixture {igt_display_fini (&data.display);buf_ops_destroy (data.bops); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 35/76 i915/kms_big_joiner: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_big_joiner.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_big_joiner.c +1 -0
@@ -328,5 +328,6 @@ igt_main igt_fixture {igt_remove_fb (data.drm_fd, &data.fb);igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 40/76 i915/kms_mmap_write_crc: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_mmap_write_crc.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_mmap_write_crc.c +1 -0
@@ -299,5 +299,6 @@ igt_main_args("n", NULL, NULL, opt_handler, NULL)close (data.drm_fd);igt_stop_helper (&hog); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 42/76 i915/kms_psr2_sf: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_psr2_sf.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_psr2_sf.c +1 -0
@@ -1034,5 +1034,6 @@ igt_main igt_fixture {close (data.debugfs_fd);display_fini (&data); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 58/76 kms_vblank: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_vblank.c | 4 ++++ 1 file changed, 4 insertions(+)
tests/kms_vblank.c +4 -0
@@ -550,4 +550,8 @@ igt_mainfor_each_pipe_static (data.pipe) igt_subtest_grouprun_subtests_for_pipe (&data); + + igt_fixture { +close (fd); + } }
Signature · 1 line
2.37.3
PATCH 3/76 i915/gem_flink_basic: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_flink_basic.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)
tests/i915/gem_flink_basic.c +8 -3
@@ -25,7 +25,6 @@* */ -#include "igt.h" #include <stdlib.h> #include <stdio.h> #include <string.h>
@@ -34,8 +33,12 @@#include <errno.h> #include <sys/stat.h> #include <sys/ioctl.h> +#include "drm.h" +#include "igt.h" +#include "igt_types.h" +IGT_TEST_DESCRIPTION ("Tests for flink - a way to export a gem object by name" ); static void
@@ -152,12 +155,14 @@ test_flink_lifetime(int fd) ret =ioctl (fd2,DRM_IOCTL_GEM_OPEN , &open_struct);igt_assert_eq (ret,0 );igt_assert (open_struct.handle !=0 ); -} -int fd; +close (fd2); +} igt_main { +igt_fd_t (fd); +igt_fixture fd =drm_open_driver (DRIVER_INTEL );
Signature · 1 line
2.37.3
PATCH 10/76 i915/gem_exec_alignment: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_exec_alignment.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tests/i915/gem_exec_alignment.c +2 -1
@@ -44,6 +44,7 @@#include "i915/gem.h" #include "i915/gem_create.h" #include "igt.h" +#include "igt_types.h" IGT_TEST_DESCRIPTION ("Exercises the basic execbuffer using object alignments" );
@@ -525,7 +526,7 @@ static void single(int fd)igt_main { -int fd =-1 ; +igt_fd_t (fd ); igt_fixture { fd =drm_open_driver (DRIVER_INTEL );
Signature · 1 line
2.37.3
PATCH 12/76 i915/gem_render_copy_redux: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_render_copy_redux.c | 4 ++++ 1 file changed, 4 insertions(+)
tests/i915/gem_render_copy_redux.c +4 -0
@@ -233,4 +233,8 @@ igt_maincopy_flink (&data);igt_stop_signal_helper (); } + + igt_fixture { +data_fini (&data); + } }
Signature · 1 line
2.37.3
PATCH 15/76 i915/gem_workarounds: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_workarounds.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tests/i915/gem_workarounds.c +2 -1
@@ -31,6 +31,7 @@#include "i915/gem_create.h" #include "igt.h" #include "igt_device.h" +#include "igt_types.h" #define PAGE_SIZE 4096#define PAGE_ALIGN (x) ALIGN(x, PAGE_SIZE)
@@ -238,7 +239,6 @@ static void check_workarounds(int fd, enum operation op, unsigned int flags) igt_main {struct intel_mmio_data mmio_data; -int device =-1 ; const struct { const char *name; enum operation op;
@@ -258,6 +258,7 @@ igt_main { "-fd", FD }, { } }, *m; +igt_fd_t (device); igt_fixture {FILE *file;
Signature · 1 line
2.37.3
PATCH 20/76 i915/i915_pciid: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/i915_pciid.c | 2 ++ 1 file changed, 2 insertions(+)
tests/i915/i915_pciid.c +2 -0
@@ -64,4 +64,6 @@ igt_simple_mainint intel =drm_open_driver (DRIVER_INTEL );igt_assert (has_known_intel_chipset (intel)); + +close (intel); }
Signature · 1 line
2.37.3
PATCH 24/76 kms_psr: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_psr.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_psr.c +1 -0
@@ -592,5 +592,6 @@ igt_main_args("", long_options, help_str, opt_handler, &data)close (data.debugfs_fd);buf_ops_destroy (data.bops);display_fini (&data); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 25/76 kms_flip: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_busy.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_busy.c +1 -0
@@ -436,5 +436,6 @@ igt_main_args("e", NULL, help_str, opt_handler, NULL) igt_fixture {igt_display_fini (&display); +close (display.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 28/76 kms_atomic: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_atomic.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_atomic.c +1 -0
@@ -1436,5 +1436,6 @@ igt_mainigt_remove_fb (display.drm_fd, &fb);igt_display_fini (&display); +close (display.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 30/76 kms_atomic_transition: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_atomic_transition.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_atomic_transition.c +1 -0
@@ -1132,5 +1132,6 @@ igt_main_args("", long_opts, help_str, opt_handler, &data) igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 32/76 i915/kms_ccs: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_ccs.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tests/i915/kms_ccs.c +3 -1
@@ -699,6 +699,8 @@ igt_main_args("cs:", NULL, help_str, opt_handler, &data) } } - igt_fixture + igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); + } }
Signature · 1 line
2.37.3
PATCH 33/76 kms_color: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_color.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_color.c +1 -0
@@ -1058,5 +1058,6 @@ igt_main igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 34/76 kms_cursor_crc: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_cursor_crc.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_cursor_crc.c +1 -0
@@ -910,5 +910,6 @@ igt_main }igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 37/76 i915/kms_fence_pin_leak: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_fence_pin_leak.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_fence_pin_leak.c +1 -0
@@ -229,4 +229,5 @@ igt_simple_mainbuf_ops_destroy (data.bops);igt_display_fini (&data.display); +close (data.drm_fd); }
Signature · 1 line
2.37.3
PATCH 39/76 i915/kms_flip_tiling: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_flip_tiling.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_flip_tiling.c +1 -0
@@ -229,5 +229,6 @@ igt_main igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 41/76 i915/kms_pipe_b_c_ivb: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_pipe_b_c_ivb.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_pipe_b_c_ivb.c +1 -0
@@ -291,5 +291,6 @@ igt_main igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 43/76 i915/kms_psr2_su: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_psr2_su.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_psr2_su.c +1 -0
@@ -335,5 +335,6 @@ igt_main igt_fixture {close (data.debugfs_fd);display_fini (&data); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 44/76 i915/kms_pwrite_crc: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_pwrite_crc.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_pwrite_crc.c +1 -0
@@ -191,4 +191,5 @@ igt_simple_mainrun_test (&data);igt_display_fini (&data.display); +close (data.drm_fd); }
Signature · 1 line
2.37.3
PATCH 45/76 kms_flip_event_leak: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_flip_event_leak.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_flip_event_leak.c +1 -0
@@ -115,5 +115,6 @@ igt_main igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 46/76 kms_hdr: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_hdr.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_hdr.c +1 -0
@@ -615,5 +615,6 @@ igt_main igt_fixture {igt_display_fini (&data.display); +close (data.fd); } }
Signature · 1 line
2.37.3
PATCH 48/76 kms_panel_fitting: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_panel_fitting.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tests/kms_panel_fitting.c +3 -1
@@ -280,6 +280,8 @@ igt_main igt_subtest_with_dynamic("atomic-fastset")test_panel_fitting (&data,TEST_ATOMIC ); - igt_fixture + igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); + } }
Signature · 1 line
2.37.3
PATCH 49/76 kms_plane: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_plane.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_plane.c +1 -0
@@ -1213,5 +1213,6 @@ igt_main_args("", long_opts, help_str, opt_handler, &data) igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 50/76 kms_plane_alpha_blend: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_plane_alpha_blend.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_plane_alpha_blend.c +1 -0
@@ -586,5 +586,6 @@ igt_main COMMIT_ATOMIC : COMMIT_LEGACY);igt_display_fini (&data.display); +close (data.gfx_fd); } }
Signature · 1 line
2.37.3
PATCH 52/76 kms_plane_multiple: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_plane_multiple.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tests/kms_plane_multiple.c +3 -1
@@ -468,6 +468,8 @@ igt_main_args("", long_options, help_str, opt_handler, NULL)run_test (&data, subtests[i].modifier); } - igt_fixture + igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); + } }
Signature · 1 line
2.37.3
PATCH 53/76 kms_plane_scaling: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_plane_scaling.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tests/kms_plane_scaling.c +3 -1
@@ -983,6 +983,8 @@ igt_main_args("", long_opts, help_str, opt_handler, &data) igt_subtest_f("2x-scaler-multi-pipe")test_scaler_with_multi_pipe_plane (&data); - igt_fixture + igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); + } }
Signature · 1 line
2.37.3
PATCH 54/76 kms_properties: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_properties.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_properties.c +1 -0
@@ -792,5 +792,6 @@ igt_main igt_fixture {igt_display_fini (&display); +close (display.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 64/76 kms_hdmi_inject: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_hdmi_inject.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_hdmi_inject.c +1 -0
@@ -219,5 +219,6 @@ igt_main igt_fixture {drmModeFreeConnector (connector); +close (drm_fd); } }
Signature · 1 line
2.37.3
PATCH 65/76 kms_3d: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_3d.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_3d.c +1 -0
@@ -111,4 +111,5 @@ igt_simple_mainkmstest_force_edid (drm_fd, connector,NULL );drmModeFreeConnector (connector); +close (drm_fd); }
Signature · 1 line
2.37.3
PATCH 72/76 feature_discovery: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/feature_discovery.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
tests/feature_discovery.c +3 -2
@@ -29,15 +29,16 @@#include "igt_kms.h" #include "igt_psr.h" #include "igt_sysfs.h" +#include "igt_types.h" -static int fd; -static int debugfs_fd;static igt_display_t display;IGT_TEST_DESCRIPTION ("A metatest that checks for \"features\" presence. " "The subtests here should only skip or pass, " "anything else means we have a serious problem." ); igt_main { +igt_fd_t (debugfs_fd ); +igt_fd_t (fd); igt_fixture { fd =drm_open_driver_master (DRIVER_ANY );
Signature · 1 line
2.37.3
PATCH 74/76 core_auth: Close(master) before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Add a missing close(master) to the subtest group fixture to prevent a leak during later tests.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/core_auth.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
tests/core_auth.c +5 -3
@@ -25,7 +25,6 @@* Testcase: drmGetMagic() and drmAuthMagic() */ -#include "igt.h" #include <unistd.h> #include <stdlib.h> #include <stdint.h>
@@ -48,6 +47,9 @@# include <pthread.h> #endif +#include "igt.h" +#include "igt_types.h" +IGT_TEST_DESCRIPTION ("Call drmGetMagic() and drmAuthMagic() and see if it behaves." ); static bool
@@ -192,8 +194,6 @@ static void test_basic_auth(int master) igt_main { -int master; -/* root (which we run igt as) should always be authenticated */ igt_describe ("Check drm client is always authenticated." ); igt_subtest("getclient-simple") {
@@ -220,6 +220,8 @@ igt_main/* above tests require that no drm fd is open */ igt_subtest_group { +igt_fd_t (master); +igt_fixture master =drm_open_driver_master (DRIVER_ANY );
Signature · 1 line
2.37.3
PATCH 36/76 i915/kms_cdclk: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_cdclk.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_cdclk.c +1 -0
@@ -373,5 +373,6 @@ igt_main igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 38/76 i915/kms_flip_scaled_crc: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/kms_flip_scaled_crc.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/kms_flip_scaled_crc.c +1 -0
@@ -676,5 +676,6 @@ igt_main }kmstest_set_vt_text_mode ();igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 47/76 kms_invalid_mode: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_invalid_mode.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_invalid_mode.c +1 -0
@@ -306,5 +306,6 @@ igt_main igt_fixture {igt_display_fini (&data.display);igt_reset_connectors (); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 51/76 kms_plane_cursor: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_plane_cursor.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_plane_cursor.c +1 -0
@@ -310,5 +310,6 @@ igt_main igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 55/76 kms_rotation_crc: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_rotation_crc.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_rotation_crc.c +1 -0
@@ -1187,5 +1187,6 @@ igt_main_args("", long_opts, help_str, opt_handler, &data) igt_fixture {igt_display_fini (&data.display); +close (data.gfx_fd); } }
Signature · 1 line
2.37.3
PATCH 56/76 kms_sequence: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_sequence.c | 4 ++++ 1 file changed, 4 insertions(+)
tests/kms_sequence.c +4 -0
@@ -295,4 +295,8 @@ igt_main } } } + + igt_fixture { +close (fd); + } }
Signature · 1 line
2.37.3
PATCH 57/76 kms_universal_plane: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_universal_plane.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_universal_plane.c +1 -0
@@ -823,5 +823,6 @@ igt_main igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 59/76 syncobj_wait: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/syncobj_wait.c | 4 ++++ 1 file changed, 4 insertions(+)
tests/syncobj_wait.c +4 -0
@@ -912,4 +912,8 @@ igt_main igt_subtest("wait-all-interrupted")test_wait_interrupted (fd,WAIT_ALL ); + + igt_fixture { +close (fd); + } }
Signature · 1 line
2.37.3
PATCH 60/76 syncobj_basic: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/syncobj_basic.c | 4 ++++ 1 file changed, 4 insertions(+)
tests/syncobj_basic.c +4 -0
@@ -231,4 +231,8 @@ igt_main igt_subtest("test-valid-cycle")test_valid_cycle (fd); + igt_fixture { +close (fd); + } + }
Signature · 1 line
2.37.3
PATCH 61/76 syncobj_timeline: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/syncobj_timeline.c | 4 ++++ 1 file changed, 4 insertions(+)
tests/syncobj_timeline.c +4 -0
@@ -1536,4 +1536,8 @@ igt_mainigt_describe (test_32bits_limit_desc); igt_subtest("32bits-limit")test_32bits_limit (fd); + + igt_fixture { +close (fd); + } }
Signature · 1 line
2.37.3
PATCH 62/76 drm_import_export: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/drm_import_export.c | 5 +++++ 1 file changed, 5 insertions(+)
tests/drm_import_export.c +5 -0
@@ -296,4 +296,9 @@ igt_main {pthread_join (test_thread_id3,NULL );pthread_join (test_thread_id4,NULL ); } + + igt_fixture { +close (fd); +close (fd1); + } }
Signature · 1 line
2.37.3
PATCH 63/76 kms_dp_aux_dev: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_dp_aux_dev.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_dp_aux_dev.c +1 -0
@@ -134,4 +134,5 @@ igt_simple_mainigt_require (valid_connectors);drmModeFreeResources (res); +close (drm_fd); }
Signature · 1 line
2.37.3
PATCH 66/76 kms_scaling_modes: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_scaling_modes.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tests/kms_scaling_modes.c +3 -1
@@ -132,6 +132,8 @@ igt_main igt_subtest_with_dynamic("scaling-mode-none")test_scaling_mode (&data,DRM_MODE_SCALE_NONE ); - igt_fixture + igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); + } }
Signature · 1 line
2.37.3
PATCH 67/76 kms_dither: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_dither.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_dither.c +1 -0
@@ -247,5 +247,6 @@ igt_main igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 68/76 drm_read: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/drm_read.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tests/drm_read.c +3 -1
@@ -46,6 +46,8 @@#include <pthread.h> #include "drm.h" +#include "igt_types.h" +IGT_TEST_DESCRIPTION ("Call read(drm) and see if it behaves." );static void sighandler (int sig,siginfo_t * info,void *context)
@@ -253,10 +255,10 @@ static void test_short_buffer_wakeup(int in, enum pipe pipe) igt_main { -int fd;igt_display_t display;struct igt_fb fb;enum pipe pipe; +igt_fd_t (fd); igt_fixture {struct sigaction alarm_action = {};
Signature · 1 line
2.37.3
PATCH 69/76 kms_cursor_edge_walk: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_cursor_edge_walk.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tests/kms_cursor_edge_walk.c +3 -1
@@ -362,6 +362,8 @@ igt_main_args("", long_opts, help_str, opt_handler, &data) } } - igt_fixture + igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); + } }
Signature · 1 line
2.37.3
PATCH 70/76 kms_plane_lowres: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_plane_lowres.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tests/kms_plane_lowres.c +3 -1
@@ -334,6 +334,8 @@ igt_mainrun_test (&data, subtests[i].modifier); } - igt_fixture + igt_fixture {igt_display_fini (&data.display); +close (data.drm_fd); + } }
Signature · 1 line
2.37.3
PATCH 71/76 kms_content_protection: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_content_protection.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_content_protection.c +1 -0
@@ -839,5 +839,6 @@ igt_main igt_fixture {test_content_protection_cleanup ();igt_display_fini (&data.display); +close (data.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 73/76 i915/i915_pm_dc: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/i915_pm_dc.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/i915_pm_dc.c +1 -0
@@ -573,6 +573,7 @@ igt_mainclose (data.debugfs_fd);close (data.msr_fd);display_fini (&data); +close (data.drm_fd); }igt_exit ();
Signature · 1 line
2.37.3
PATCH 75/76 i915/gem_exec_balancer: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_exec_balancer.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tests/i915/gem_exec_balancer.c +2 -1
@@ -35,6 +35,7 @@#include "igt_gt.h" #include "igt_perf.h" #include "igt_sysfs.h" +#include "igt_types.h" #include "sw_sync.h" IGT_TEST_DESCRIPTION ("Exercise in-kernel load-balancing" );
@@ -3302,7 +3303,7 @@ static bool has_parallel_execbuf(int i915)igt_main { -int i915 =-1 ; +igt_fd_t (i915 ); igt_fixture { i915 =drm_open_driver (DRIVER_INTEL );
Signature · 1 line
2.37.3
PATCH 76/76 i915/gem_ctx_isolation: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_ctx_isolation.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tests/i915/gem_ctx_isolation.c +2 -1
@@ -25,6 +25,7 @@#include "i915/gem_create.h" #include "igt.h" #include "igt_dummyload.h" +#include "igt_types.h" #define MAX_REG 0x200000#define NUM_REGS (MAX_REG / sizeof(uint32_t))
@@ -979,7 +980,7 @@ igt_mainunsigned int has_context_isolation =0 ;const struct intel_execution_engine2 *e;intel_ctx_cfg_t cfg; -int i915 =-1 ; +igt_fd_t (i915); igt_fixture {int gen;
Signature · 1 line
2.37.3
PATCH 4/76 i915/gem_linear_blits: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_linear_blits.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tests/i915/gem_linear_blits.c +2 -1
@@ -47,6 +47,7 @@#include "i915/gem.h" #include "i915/gem_create.h" #include "igt.h" +#include "igt_types.h" IGT_TEST_DESCRIPTION ("Test doing many blits with a working set larger than the" " aperture size." );
@@ -236,7 +237,7 @@ igt_mainconst int ncpus =sysconf (_SC_NPROCESSORS_ONLN);uint64_t count =0 ;bool do_relocs; -int fd =-1 ; +igt_fd_t (fd); igt_fixture { fd =drm_open_driver (DRIVER_INTEL );
Signature · 1 line
2.37.3
PATCH 8/76 i915/gem_ctx_engines: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_ctx_engines.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
tests/i915/gem_ctx_engines.c +2 -1
@@ -41,6 +41,7 @@#include "i915/gem_context.h" #include "i915/gem_create.h" #include "igt.h" +#include "igt_types.h" #include "sw_sync.h" #define engine_class (e, n) ((e)->engines[(n)].engine_class)
@@ -593,7 +594,7 @@ static void independent_all(int i915, const intel_ctx_t *ctx) igt_main {const struct intel_execution_engine2 *e; -int i915 =-1 ; +igt_fd_t (i915); igt_fixture { i915 =drm_open_driver (DRIVER_INTEL );
Signature · 1 line
2.37.3
PATCH 14/76 i915/gem_userptr_blits: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_userptr_blits.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
tests/i915/gem_userptr_blits.c +3 -1
@@ -62,6 +62,7 @@#include "i915/gem_create.h" #include "igt.h" #include "igt_sysfs.h" +#include "igt_types.h" #include "sw_sync.h" #include "eviction_common.c"
@@ -2299,7 +2300,7 @@ static void test_userfault(int i915)uint64_t total_ram;uint64_t aperture_size; -int fd, count; +int count;static int opt_handler (int opt,int opt_index,void *data) {
@@ -2319,6 +2320,7 @@ const char *help_str = " -c\tBuffer count\n"; igt_main_args("c:", NULL, help_str, opt_handler, NULL) {int size =sizeof (linear); +igt_fd_t (fd); igt_fixture {unsigned int mmo_max =0 ;
Signature · 1 line
2.37.3
PATCH 21/76 kms_cursor_legacy: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_cursor_legacy.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_cursor_legacy.c +1 -0
@@ -1753,5 +1753,6 @@ igt_mainif (intel_psr2_restore)i915_psr2_sel_fetch_restore (display.drm_fd);igt_display_fini (&display); +close (display.drm_fd); } }
Signature · 1 line
2.37.3
PATCH 26/76 kms_force_connector_basic: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_force_connector_basic.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
tests/kms_force_connector_basic.c +6 -1
@@ -33,9 +33,9 @@ IGT_TEST_DESCRIPTION("Check the debugfs force connector/edid features work"static void reset_connectors (void ) { -int drm_fd =0 ;drmModeRes *res;drmModeConnector *connector =NULL ; +int drm_fd; drm_fd =drm_open_driver_master (DRIVER_ANY ); res =drmModeGetResources (drm_fd);
@@ -54,6 +54,8 @@ static void reset_connectors(void) }igt_set_module_param_int (drm_fd,"load_detect_test" ,0 ); + +close (drm_fd); }static int opt_handler (int opt,int opt_index,void *data)
@@ -224,6 +226,9 @@ igt_main_args("", long_opts, help_str, opt_handler, NULL)drmModeFreePlane (drm_plane); } + +igt_remove_fb (drm_fd, &xrgb_fb); +igt_remove_fb (drm_fd, &argb_fb); }igt_describe ("Test to check the forced connector state." );
Signature · 1 line
2.37.3
PATCH 29/76 kms_atomic_interruptible: Close device before exit
In reply to Mauro Carvalho ChehabFrom: Chris Wilson <[email protected]>
Close the device fd before we check for leaks during the atexit handlers.
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/kms_atomic_interruptible.c | 1 + 1 file changed, 1 insertion(+)
tests/kms_atomic_interruptible.c +1 -0
@@ -349,5 +349,6 @@ igt_main/* TODO: legacy gamma_set/get, object set/getprop, getcrtc, getconnector */ igt_fixture {igt_display_fini (&display); +close (display.drm_fd); } }
Signature · 1 line
2.37.3
Quoted 5 lines · Nirmoy
Hi Mauro,
Is there any reason we are avoid using |igt_fd_t for patches 27,26,22 ?|
|Ah never mind, just realized about igt_skip()|
|Nirmoy|
Quoted 186 lines
|For i915 patches Reviewed-by:| Nirmoy Das<[email protected]>
Regards,
Nirmoy
||Quoted 177 lines · Mauro Carvalho Chehab
From: Mauro Carvalho Chehab<[email protected]>
GCC is usually smart enough to close devices on exit, but that doesn't work with the logic inside
igt_skip(), as it uses longjmp, causing the code to go out of scope. Such scoped variables must be tracked in an outer scope to the skipping subtest.Change the logic on IGT tests in order to avoid keeping device opened at the end of the test.
Chris Wilson (76):
lib: Introduce typed cleanups
i915/gem_basic: Close device before exit i915/gem_flink_basic: Close device before exit i915/gem_linear_blits: Close device before exit i915/gem_blits: Close device before exit i915/gem_close: Close device before exit i915/gem_create: Close device before exit i915/gem_ctx_engines: Close device before exit i915/gem_ctx_shared: Close device before exit i915/gem_exec_alignment: Close device before exit i915/gem_render_copy: Close device before exit i915/gem_render_copy_redux: Close device before exit i915/gem_media_fill: Close device before exit i915/gem_userptr_blits: Close device before exit i915/gem_workarounds: Close device before exit i915/gem_unref_active_buffers: Close device before exit i915/gem_exec_endless: Close device before exit i915/gem_request_retire: Close device before exit i915/gem_unfence_active_buffers: Close device before exit i915/i915_pciid: Close device before exit
kms_cursor_legacy: Close device before exit
kms_flip: Close device before exit
kms_pipe_crc_basic: Close device before exit
kms_psr: Close device before exit
kms_flip: Close device before exit
kms_force_connector_basic: Close device before exit
dumb_buffer: Close device before exit
kms_atomic: Close device before exit
kms_atomic_interruptible: Close device before exit
kms_atomic_transition: Close device before exit i915/kms_big_fb: Close device before exit i915/kms_ccs: Close device before exit
kms_color: Close device before exit
kms_cursor_crc: Close device before exit i915/kms_big_joiner: Close device before exit i915/kms_cdclk: Close device before exit i915/kms_fence_pin_leak: Close device before exit i915/kms_flip_scaled_crc: Close device before exit i915/kms_flip_tiling: Close device before exit i915/kms_mmap_write_crc: Close device before exit i915/kms_pipe_b_c_ivb: Close device before exit i915/kms_psr2_sf: Close device before exit i915/kms_psr2_su: Close device before exit i915/kms_pwrite_crc: Close device before exit
kms_flip_event_leak: Close device before exit
kms_hdr: Close device before exit
kms_invalid_mode: Close device before exit
kms_panel_fitting: Close device before exit
kms_plane: Close device before exit
kms_plane_alpha_blend: Close device before exit
kms_plane_cursor: Close device before exit
kms_plane_multiple: Close device before exit
kms_plane_scaling: Close device before exit
kms_properties: Close device before exit
kms_rotation_crc: Close device before exit
kms_sequence: Close device before exit
kms_universal_plane: Close device before exit
kms_vblank: Close device before exit
syncobj_wait: Close device before exit
syncobj_basic: Close device before exit
syncobj_timeline: Close device before exit
drm_import_export: Close device before exit
kms_dp_aux_dev: Close device before exit
kms_hdmi_inject: Close device before exit
kms_3d: Close device before exit
kms_scaling_modes: Close device before exit
kms_dither: Close device before exit
drm_read: Close device before exit
kms_cursor_edge_walk: Close device before exit
kms_plane_lowres: Close device before exit
kms_content_protection: Close device before exit
feature_discovery: Close device before exit i915/i915_pm_dc: Close device before exit
core_auth: Close(master) before exit i915/gem_exec_balancer: Close device before exit i915/gem_ctx_isolation: Close device before exitlib/igt_core.c | 6 ++ lib/igt_core.h | 2 + lib/igt_types.c | 17 +++ lib/igt_types.h | 47 ++++++++ lib/meson.build | 1 + lib/tests/bad_subtest_type.c | 19 ++++ lib/tests/igt_types.c | 136 ++++++++++++++++++++++++ lib/tests/meson.build | 2 + tests/core_auth.c | 8 +- tests/drm_import_export.c | 5 + tests/drm_read.c | 4 +- tests/dumb_buffer.c | 4 + tests/feature_discovery.c | 5 +- tests/i915/gem_basic.c | 8 +- tests/i915/gem_blits.c | 1 + tests/i915/gem_close.c | 3 +- tests/i915/gem_create.c | 3 +- tests/i915/gem_ctx_engines.c | 3 +- tests/i915/gem_ctx_isolation.c | 3 +- tests/i915/gem_ctx_shared.c | 3 +- tests/i915/gem_exec_alignment.c | 3 +- tests/i915/gem_exec_balancer.c | 3 +- tests/i915/gem_exec_endless.c | 3 +- tests/i915/gem_flink_basic.c | 11 +- tests/i915/gem_linear_blits.c | 3 +- tests/i915/gem_media_fill.c | 1 + tests/i915/gem_render_copy.c | 1 + tests/i915/gem_render_copy_redux.c | 4 + tests/i915/gem_request_retire.c | 5 +- tests/i915/gem_unfence_active_buffers.c | 1 + tests/i915/gem_unref_active_buffers.c | 1 + tests/i915/gem_userptr_blits.c | 4 +- tests/i915/gem_workarounds.c | 3 +- tests/i915/i915_pciid.c | 2 + tests/i915/i915_pm_dc.c | 1 + tests/i915/kms_big_fb.c | 1 + tests/i915/kms_big_joiner.c | 1 + tests/i915/kms_busy.c | 1 + tests/i915/kms_ccs.c | 4 +- tests/i915/kms_cdclk.c | 1 + tests/i915/kms_fence_pin_leak.c | 1 + tests/i915/kms_flip_scaled_crc.c | 1 + tests/i915/kms_flip_tiling.c | 1 + tests/i915/kms_mmap_write_crc.c | 1 + tests/i915/kms_pipe_b_c_ivb.c | 1 + tests/i915/kms_psr.c | 1 + tests/i915/kms_psr2_sf.c | 1 + tests/i915/kms_psr2_su.c | 1 + tests/i915/kms_pwrite_crc.c | 1 + tests/kms_3d.c | 1 + tests/kms_atomic.c | 1 + tests/kms_atomic_interruptible.c | 1 + tests/kms_atomic_transition.c | 1 + tests/kms_color.c | 1 + tests/kms_content_protection.c | 1 + tests/kms_cursor_crc.c | 1 + tests/kms_cursor_edge_walk.c | 4 +- tests/kms_cursor_legacy.c | 1 + tests/kms_dither.c | 1 + tests/kms_dp_aux_dev.c | 1 + tests/kms_flip.c | 3 + tests/kms_flip_event_leak.c | 1 + tests/kms_force_connector_basic.c | 7 +- tests/kms_hdmi_inject.c | 1 + tests/kms_hdr.c | 1 + tests/kms_invalid_mode.c | 1 + tests/kms_panel_fitting.c | 4 +- tests/kms_pipe_crc_basic.c | 1 + tests/kms_plane.c | 1 + tests/kms_plane_alpha_blend.c | 1 + tests/kms_plane_cursor.c | 1 + tests/kms_plane_lowres.c | 4 +- tests/kms_plane_multiple.c | 4 +- tests/kms_plane_scaling.c | 4 +- tests/kms_properties.c | 1 + tests/kms_rotation_crc.c | 1 + tests/kms_scaling_modes.c | 4 +- tests/kms_sequence.c | 4 + tests/kms_universal_plane.c | 1 + tests/kms_vblank.c | 4 + tests/syncobj_basic.c | 4 + tests/syncobj_timeline.c | 4 + tests/syncobj_wait.c | 4 + 83 files changed, 385 insertions(+), 33 deletions(-)create mode 100644
lib/igt_types.c
create mode 100644lib/igt_types.h
create mode 100644lib/tests/bad_subtest_type.c
create mode 100644lib/tests/igt_types.c
Re: [igt-dev] [PATCH 17/76] i915/gem_exec_endless: Close device before exit
In reply to a message not in the archiveQuoted 5 lines · On Fri, 30 Sep 2022 14:25:03 +0200 Kamil Konieczny…
Hi Mauro,
Quoted 2 lines · Mauro Carvalho Chehab
From: Chris Wilson <[email protected]>
Quoted 7 lines
Quoted 2 lines
Close the device fd before we check for leaks during the atexit handlers.
This is somewhat misleading, it only changes type but missed actually closing i915 var.
Maybe some place for future improvement,
When igt_fd_t() is used, there's no need to explicitly close it, as gcc/clang will use a typed cleanup to close it when the variable lifetime expires.
Quoted 38 lines
Reviewed-by Kamil Konieczny <[email protected]>Quoted 35 lines
Signed-off-by Chris Wilson <[email protected]>Signed-off-by Mauro Carvalho Chehab <[email protected]>---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
tests/i915/gem_exec_endless.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)tests/i915/gem_exec_endless.c +2
-1@@ -28,6 +28,7 @@#include "igt.h" #include "igt_device.h" #include "igt_sysfs.h" +#include "igt_types.h" #include "sw_sync.h" #define MAX_ENGINES 64@@ -341,7 +342,7 @@ static void unpin_rps(int sysfs) igt_main {const struct intel_execution_engine2 *e; -int i915 =-1 ; +igt_fd_t (i915);igt_skip_on_simulation ();Signature · 1 line
2.37.3
Re: [igt-dev] [PATCH 01/76] lib: Introduce typed cleanups
In reply to a message not in the archiveQuoted 84 lines · On Fri, 30 Sep 2022 17:16:27 +0200 Kamil Konieczny…
Hi Mauro,
Quoted 6 lines · Mauro Carvalho Chehab
From: Chris Wilson <[email protected]>
Start introducing standard types with automatic cleanup courtesy of gcc's __attribute__((cleanup)). As an example, we start with an fd that will automatically call
close()on going out of scope, and crucially before atexit where we will want to check for resource leaks.This is "what is changing" part, please also put here "why part" from cover letter, so it will stay in git history.
Quoted 7 lines
Suggested-by Andrzej Hajda <[email protected]>Signed-off-by Chris Wilson <[email protected]>Cc Andrzej Hajda <[email protected]>Cc Petri Latvala <[email protected]>Acked-by Nirmoy Das <[email protected]>Signed-off-by Mauro Carvalho Chehab <[email protected]>I will send cc also to Zbigniew.
Quoted 61 lines
---
To avoid mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 00/76] at: https://lore.kernel.org/all/[email protected]/
lib/igt_core.c | 6 ++ lib/igt_core.h | 2 + lib/igt_types.c | 17 +++++ lib/igt_types.h | 47 ++++++++++++ lib/meson.build | 1 + lib/tests/bad_subtest_type.c | 19 +++++ lib/tests/igt_types.c | 136 +++++++++++++++++++++++++++++++++++ lib/tests/meson.build | 2 + 8 files changed, 230 insertions(+) create mode 100644 lib/igt_types.c create mode 100644 lib/igt_types.h create mode 100644 lib/tests/bad_subtest_type.c create mode 100644 lib/tests/igt_types.clib/igt_core.c +6
-0@@ -624,6 +624,12 @@ uint64_t igt_nsec_elapsed(struct timespec *start) (uint64_t)NSEC_PER_SEC*(now.tv_sec - start->tv_sec)); } +void __igt_assert_in_outer_scope (void ) +{ +internal_assert (!in_subtest, +"must only be called outside of a subtest" ); +} +bool __igt_fixture (void ) { internal_assert(!in_fixture,lib/igt_core.h +2
-0@@ -135,6 +135,8 @@ struct _GKeyFile *igt_load_igtrc(void);*/ #define IGT_EXIT_ABORT 112 +void __igt_assert_in_outer_scope (void ); +bool __igt_fixture (void );void __igt_fixture_complete (void );__noreturn void __igt_fixture_end (void );lib/igt_types.c new file +6
-0@@ -0,0 +1,17 @@ +// SPDX-License-Identifier: MIT +/* +* Copyright © 2022 Intel Corporation +*/ + +#include <unistd.h>
Quoted 8 lines
lib/igt_types.c line 12
Quoted 6 lines
lib/igt_types.c new file +6
-0+ +#include "igt_types.h" + +void igt_cleanup_fd (volatile int *fd) +{ +if (!fd || *fd <0 )---------------------- ^
imho here should be != -1
In practice, it is identical, as open() returns only -1. On the other hand, close(-2) - or any other negative value - will produce an error, so, I prefer checking against < 0 here.
Quoted 2 lines
Consider also (*fd >= -1 && *fd <= 2) as these are stdout/in/error file descriptors.
In thesis, you're right, but the only case where 0, 1 or 2 would be used would be if someone would force using the typed cleanups with something like:
foo()
{
igt_fd_t(close_stdout_fd);
close_stdout_fd = STDOUT;
}this would cause STDOUT to be closed at the end of foo() function.
I can't see any cases where something like that would be used. On the other hand, if someone is doing some tricks with stdout, I can't see why the library should care enough to not let things like the above work ;-)
Quoted 46 lines
Quoted 33 lines
lib/igt_types.c new file +5
-0+return ; + +close (*fd); + *fd =-1 ; +}lib/igt_types.h new file +21
-0@@ -0,0 +1,47 @@ +/* SPDX-License-Identifier: MIT */ +/* +* Copyright © 2022 Intel Corporation +*/ + +#ifndef IGT_TYPES_H +#define IGT_TYPES_H + +/* +* GCC can automatically cleanup variables that go out of scope, but only +* through normal means. Breaking out of scope using longjmp (i.e. igt_skip) +* is not handled automatically by GCC. Such scoped variables must be tracked +* in an outer scope to the skipping subtest. +* +* BAD: +* igt_subtest("bad") { +* igt_fd_t(fd); +* +* fd = drm_open_driver(); +* } +* lib/igt_types.h line 31
Quoted 10 lines
lib/igt_types.h new file +10
-0+* GOOD: +* igt_subtest_group() { +* igt_fd_t(fd); +* +* igt_fixture { +* fd = drm_open_driver(); +* } +* +* igt_subtest("good") +* ; Should we close here fd ? This is example of good practice so imho here:
Quoted 5 lines
igt_fixture close(fd);even if this is redundant in case of assert or fail, or just close(fd) without fixture ?
Just close won't work when longjmp is used, as this is an undefined STD-C behavior. See this comment at lib/igt_core.c:
*
* # Magic Control Blocks
*
* i-g-t makes heavy use of C macros which serve as magic control blocks. They
* work fairly well and transparently but since C doesn't have full-blown
* closures there are caveats:
...
*
* - Code blocks with magic control flow are implemented with setjmp()
* and longjmp(). This applies to #igt_fixture, #igt_subtest, * #igt_subtest_with_dynamic and #igt_dynamic * blocks and all the three variants to finish test: igt_success(), * igt_skip() and igt_fail(). Mostly this is of no concern, except * when such a control block changes stack variables defined in the * same function as the control block resides. Any store/load * behaviour after a longjmp() is ill-defined for these * variables. Avoid such code. * * Quoting the man page for longjmp(): * * "The values of automatic variables are unspecified after a call to * longjmp() if they meet all the following criteria:" * - "they are local to the function that made the corresponding setjmp() call; * - "their values are changed between the calls to setjmp() and longjmp(); and * - "they are not declared as volatile."
The hole idea of this series is to teach gcc and clang how to deal when igt_skip() & similar functions are used. Basically, when a fd variable is declared with:
igt_fd_t(fd);
gcc/clang will then free the variable even when longjmp is used.
Regards,
Mauro
Re: [igt-dev] [PATCH 01/76] lib: Introduce typed cleanups
In reply to a message not in the archiveQuoted 11 lines · On Tue, 04 Oct 2022 12:17:20 +0200 Janusz Krzysztofik…
Quoted 10 lines
Review
Quoted 2 lines
(quoted diff) +2
-0+#include "igt_core.h" +#include "igt_types.h"Please put here description with
IGT_TEST_DESCRIPTION()macro.lib/tests/bad_subtest_type.c line 10
Quoted 3 lines
lib/tests/bad_subtest_type.c new file +3
-0+ +igt_main +{Please put description before test with
igt_describe().
Quoted 4 lines
AFAICS none of
lib/tests/*, exceptigt_describethat needs those clauses because it exercises them, follows that convention. If you found them helpful then OK, that could be a field for improvement, but that should not block adding new self IGT tests as needed, I believe.
True, but on the other hand, it doesn't hurt adding descriptions. As such change is trivial enough, I'll add them.
Quoted 7 lines
Quoted 4 lines
Also copy comment from main test that after running this SIGABORT is intentional.
Since this test is added to
lib_fail_testscategory (see below changes tolib/tests/meson.build) then I don't think we need further clarifications.
Same here.
Quoted 3 lines
Thanks,
Janusz
Re: [igt-dev] [PATCH 01/76] lib: Introduce typed cleanups
In reply to a message not in the archiveQuoted 6 lines · On Wed, 5 Oct 2022 09:45:32 +0200 Zbigniew Kempczyński…
lib/igt_types.h line 45
Quoted 3 lines
lib/igt_types.h new file +3
-0+void igt_cleanup_fd (volatile int *fd); +#define igt_fd_t (x__) \ + volatile int x__ cleanup_with(igt_cleanup_fd) = IGT_OUTER_SCOPE_INIT(-1)Why do we need volatile here? Variable after initalization (open) likely won't be touched (until exiting its scope) so what's the risk?
With current gcc versions, volatile is not needed, but this fails with clang. The real problem is that C spec doesn't define what happens with an automatic stack variable after doing a longjmp.
From Petri's feedback, the only solution that works is to declare such variables as volatile.
Regards,
Mauro