# patchcheck.conf
#
# This contains a test that takes two git commits and will test each
# commit between the two. The build test will look at what files the
# commit has touched, and if any of those files produce a warning, then
# the build will fail.
# PATCH_START is the commit to begin with and PATCH_END is the commit
# to end with (inclusive). This is similar to doing a git rebase -i PATCH_START~1
# and then testing each commit and doing a git rebase --continue.
# You can use a SHA1, a git tag, or anything that git will accept for a checkout
PATCH_START := HEAD~3
PATCH_END := HEAD
# Use the oldconfig if build_type wasn't defined
DEFAULTS IF NOT DEFINED BUILD_TYPE
DO_BUILD_TYPE := oldconfig
DEFAULTS ELSE
DO_BUILD_TYPE := ${BUILD_TYPE}
DEFAULTS
# Change PATCH_CHECKOUT to be the branch you want to test. The test will
# do a git checkout of this branch before starting. Obviously both
# PATCH_START and PATCH_END must be in this branch (and PATCH_START must
# be contained by PATCH_END).
PATCH_CHECKOUT := test/branch
# Usually it's a good idea to have a set config to use for testing individual
# patches.
PATCH_CONFIG := ${CONFIG_DIR}/config-patchcheck
# Change PATCH_TEST to run some test for each patch. Each commit that is
# tested, after it is built and installed on the test machine, this command
# will be executed. Usually what is done is to ssh to the target box and
# run some test scripts. If you just want to boot test your patches
# comment PATCH_TEST out.
PATCH_TEST := ${SSH} "/usr/local/bin/ktest-test-script"
DEFAULTS IF DEFINED PATCH_TEST
PATCH_TEST_TYPE := test
DEFAULTS ELSE
PATCH_TEST_TYPE := boot
# If for some reason a file has a warning that one of your patches touch
# but you do not care about it, set IGNORE_WARNINGS to that commit(s)
# (space delimited)
#IGNORE_WARNINGS = 39eaf7ef884dcc44f7ff1bac803ca2a1dcf43544 6edb2a8a385f0cdef51dae37ff23e74d76d8a6ce
# Instead of just checking for warnings to files that are changed
# it can be advantageous to check for any new warnings. If a
# header file is changed, it could cause a warning in a file not
# touched by the commit. To detect these kinds of warnings, you
# can use the WARNINGS_FILE option.
#
# If the variable CREATE_WARNINGS_FILE is set, this config will
# enable the WARNINGS_FILE during the patchcheck test. Also,
# before running the patchcheck test, it will create the
# warnings file.
#
DEFAULTS IF DEFINED CREATE_WARNINGS_FILE
WARNINGS_FILE = ${OUTPUT_DIR}/warnings_file
TEST_START IF DEFINED CREATE_WARNINGS_FILE
# WARNINGS_FILE is already set by the DEFAULTS above
TEST_TYPE = make_warnings_file
# Checkout the commit before the patches to test,
# and record all the warnings that exist before the patches
# to test are added
CHECKOUT = ${PATCHCHECK_START}~1
# Force a full build
BUILD_NOCLEAN = 0
BUILD_TYPE = ${DO_BUILD_TYPE}
# If you are running a multi test, and the test failed on the first
# test but on, say the 5th patch. If you want to restart on the
# fifth patch, set PATCH_START1. This will make the first test start
# from this commit instead of the PATCH_START commit.
# Note, do not change this option. Just define PATCH_START1 in the
# top config (the one you pass to ktest.pl), and this will use it,
# otherwise it will just use PATCH_START if PATCH_START1 is not defined.
DEFAULTS IF NOT DEFINED PATCH_START1
PATCH_START1 := ${PATCH_START}
TEST_START IF ${TEST} == patchcheck
TEST_TYPE = patchcheck
MIN_CONFIG = ${PATCH_CONFIG}
TEST = ${PATCH_TEST}
PATCHCHECK_TYPE = ${PATCH_TEST_TYPE}
PATCHCHECK_START = ${PATCH_START1}
PATCHCHECK_END = ${PATCH_END}
CHECKOUT = ${PATCH_CHECKOUT}
BUILD_TYPE = ${DO_BUILD_TYPE}
TEST_START IF ${TEST} == patchcheck && ${MULTI}
TEST_TYPE = patchcheck
MIN_CONFIG = ${PATCH_CONFIG}
TEST = ${PATCH_TEST}
PATCHCHECK_TYPE = ${PATCH_TEST_TYPE}
PATCHCHECK_START = ${PATCH_START}
PATCHCHECK_END = ${PATCH_END}
CHECKOUT = ${PATCH_CHECKOUT}
# Use multi to test different compilers?
MAKE_CMD = CC=gcc-4.5.1 make
BUILD_TYPE = ${DO_BUILD_TYPE}
ue='1'>1
drm/i915: Check for NULL i915_vma in intel_unpin_fb_obj()
I've seen this trigger twice now, where the i915_gem_object_to_ggtt()
call in intel_unpin_fb_obj() returns NULL, resulting in an oops
immediately afterwards as the (inlined) call to i915_vma_unpin_fence()
tries to dereference it.
It seems to be some race condition where the object is going away at
shutdown time, since both times happened when shutting down the X
server. The call chains were different:
- VT ioctl(KDSETMODE, KD_TEXT):
intel_cleanup_plane_fb+0x5b/0xa0 [i915]
drm_atomic_helper_cleanup_planes+0x6f/0x90 [drm_kms_helper]
intel_atomic_commit_tail+0x749/0xfe0 [i915]
intel_atomic_commit+0x3cb/0x4f0 [i915]
drm_atomic_commit+0x4b/0x50 [drm]
restore_fbdev_mode+0x14c/0x2a0 [drm_kms_helper]
drm_fb_helper_restore_fbdev_mode_unlocked+0x34/0x80 [drm_kms_helper]
drm_fb_helper_set_par+0x2d/0x60 [drm_kms_helper]
intel_fbdev_set_par+0x18/0x70 [i915]
fb_set_var+0x236/0x460
fbcon_blank+0x30f/0x350
do_unblank_screen+0xd2/0x1a0
vt_ioctl+0x507/0x12a0
tty_ioctl+0x355/0xc30
do_vfs_ioctl+0xa3/0x5e0
SyS_ioctl+0x79/0x90
entry_SYSCALL_64_fastpath+0x13/0x94
- i915 unpin_work workqueue:
intel_unpin_work_fn+0x58/0x140 [i915]
process_one_work+0x1f1/0x480
worker_thread+0x48/0x4d0
kthread+0x101/0x140
and this patch purely papers over the issue by adding a NULL pointer
check and a WARN_ON_ONCE() to avoid the oops that would then generally
make the machine unresponsive.
Other callers of i915_gem_object_to_ggtt() seem to also check for the
returned pointer being NULL and warn about it, so this clearly has
happened before in other places.
[ Reported it originally to the i915 developers on Jan 8, applying the
ugly workaround on my own now after triggering the problem for the
second time with no feedback.
This is likely to be the same bug reported as
https://bugs.freedesktop.org/show_bug.cgi?id=98829
https://bugs.freedesktop.org/show_bug.cgi?id=99134
which has a patch for the underlying problem, but it hasn't gotten to
me, so I'm applying the workaround. ]
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'tools/testing/selftests/powerpc/dscr/Makefile')