[Intel-gfx] [ANNOUNCE] intel-gpu-tools 1.19
Petri Latvala
petri.latvala at intel.com
Fri Jun 9 13:51:13 UTC 2017
A new intel-gpu-tools quarterly release is available with the
following changes:
Library changes:
- Changed debugfs handlers to open files for the DRM device in use,
for setups with more than one DRM device. (Tomeu Vizoso)
- Added support for 4K and audio HDMI EDID injection. (Abdiel
Janulgue, Marius Vlad)
- Added support for AMDGPU devices. (Chris Wilson)
Tools changes:
- intel_error_decode now automatically opens a pager. (Chris Wilson)
- intel_error_decode now dumps the GuC firmware logs if available.
(Chris Wilson)
Benchmark changes:
- gem_wsim: New benchmark that simulates command submission
workloads. (Tvrtko Ursulin)
Scripts changes:
- trace.pl: New tool to parse i915 tracepoints for performance
analysis. (Tvrtko Ursulin)
- media-bench.pl: New tool to programmatically analyze simulated media
workloads using gem_wsim to find the optimal load balancing
strategy. (Tvrtko Ursulin)
Test changes:
- Imported amdgpu tests from libdrm. (Chris Wilson)
- Multiple other new tests.
And many other bug fixes and improvements.
And the full changelog follows:
Abdiel Janulgue (4):
intel-ci: Add initial generic testlist
lib/igt_kms: Add support for 4K and audio HDMI EDID injection.
tests/kms_hdmi_inject: Add test for HDMI injection capabilities.
lib/tests: Add kmstest_edid_add_* selftests
Ander Conselvan de Oliveira (3):
lib/dummyload: Handle timeout in a new thread instead of signal handler
lib/debugfs: Close dir before returning open debugs file
lib/igt_kms: Force outputs to use full range RGB
Arkadiusz Hiler (15):
tests/drm_import_export: Include {i915_,}drm.h properly
Make conditions on HAVE_UDEV consistent
lib/igt_aux: Include unistd.h for gettid() on Android
lib/igt_aux: Make procps optional
chamelium: Fix build issues on Android
tools/Android.mk: Add guc_logger and l3_parity skip list
tests/Android.mk: Add perf to skip list
Android.mk: Fix libkmod use
Android.mk: Filter out *.h from src files
Android.mk: Use drm stubs
tools/Android.mk: Fix zlib inclusion
tests/gem_exec_nop: Disable headless subtest on cairoless Android
tests/gem_exec_nop: Rename signal() to fence_signal()
benchmarks/gem_wsim: Add LOCAL fence defines
ffs: Include strings.h
Brian Starkey (3):
lib/igt_kms: Fix erroneous assert
lib/igt_debugfs: Remove igt_debugfs_t
lib/igt_debugfs: Only use valid values in igt_crc_to_str()
Chris Wilson (132):
Revert "Always expose IGT subtests for known kernel selftests"
Revert "lib: Add i915 and drm-mm selftest headers from the kernel"
igt/kms_pipe_crc_basic: Skip test before hang injection
igt/kms_pipe_crc_basic: Tighten timings for CRC readback
lib: Update i915_pciids.h
lib: Squelch a pair of ignore result warnings
lib: Remove requirements spam from gem_set_tiling()
igt/gem_userptr_blits: Move assert to caller
igt/gem_exec_schedule: Never use STORE_DWORD_IMM on SNB bsd engine
igt/drv_hangman: Setting wedged is serialised
igt/kms_vblank: Exercise contention of drmWaitVblank
igt/gem_spin_batch: Reduce impact from slow queueing
igt/gem_userptr_blits: Allow use of >4GiB
igt/gem_exec_nop: Sync before measuring total elapsed time
lib/debugfs: Only inspect errno after a confirmed error
Restore "lib: Open debugfs files for the given DRM device"
igt/gem_exec_reloc: Double memory requirements
igt/gem_exec_whisper: Do context requirement check in parent
lib: Clear unrelated errno for intel_detect_and_clear_missed_interrupts
lib: system() is bad, use libkmod instead
Improve utilisation of igt_debugfs_dir()
lib/debugfs: Phase out igt_debugfs_fopen()
igt/kms_sink_crc_basic: Fix debugfs error handling
igt/kms_pipe_crc_basic: Fix assertion for bad commands
lib/sysfs: stop before calling write with zero bytes
igt/kms_sink_crc_base: CRC is on 3 channels, not 4.
Replace more system("modprobe")
igt: Add gem_exec_capture to exercise capturing user requested buffers on hang
lib/sysfs: Fix using O_WRONLY in igt_sysfs_read()!
igt/kms_frontbuffer_tracking: Keep the debugfs dir around
igt/gem_tiled_pread_pwrite: Use streaming loads to speed up GTT readback
igt/pm_rc6_residency: Keep the sysfs directory open
benchmarks/gem_exec_trace: Enhanced multi-context capture
benchmarks/gem_exec_trace: Include client wait points
benchmarks/gem_exec_trace: Make the tracer threadsafe
benchmarks/gem_exec_trace: Use a constant delay
benchmarks/gem_exec_trace: Calibrate dummy workload
benchmarks/gem_exec_trace: Randomise batch duration
igt/kms_pipe_crc_basic: Use a fast hang
lib: Update kernel engine names
tests: More precautionary selfchecks to filter out broken GPUs
tests/gem_media_fill: Fixup typo
igt/gem_exec_whisper: Extend test patterns with a synchronous variant
tools/intel_error_decode: Do ascii85 decode first
tools/intel_error_decode: Add support for user specified bo
tools/intel_error_decode: Automatically open a pager
tools/intel_error_decode: Refactor matching known buffers
lib/pipe_crc: Cache debugfs directory
tests: More require GEM markup
igt/pm_rpm: Use directory fd to track and read entire directories
igt/prime_mmap_coherency: Reorder cache flushes to avoid excess pressure
igt/gem_exec_parallel: Can't run fds test on gen5
igt/gem_exec_whisper: Move the require(gen>5) out of the child
tools/intel_error_decode: Don't try instruction decoding on data buffers
tools/intel_error_decode: Dump the GuC log if available
igt/gem_seqno_wrap: Teach to skip if i915_next_seqno is no long writable
igt/gem_exec_fence: Restrict allowing hangs to hang tests
lib: Fix up internal engine names (again)
igt/gem_exec_reloc: Check interactions with WC domain
igt/gem_pwrite: Use DOMAIN_WC for access via gem_mmap__wc
igt/gem_exec_flush: Use DOMAIN_WC for access via gem_mmap__wc
igt/gem_concurrent_blit: Mark up with DOMAIN_WC
igt/gem_cs_tlb: Mark up with DOMAIN_WC
igt/gem_mmap_wc: Mark up with DOMAIN_WC
igt/gem_storedw_loop: Mark up with DOMAIN_WC
igt/gem_exec_fence: Enable hang detector
iget/gem_exec_reloc: Fix simulated relocations
igt/gem_exec_nop: Add some output to headless subtest
benchmarks/gem_latency: Provide LOCAL defines for old libdrm
igt/gem_exec_fence: Accumulate long history of fences
igt/gem_ctx_create: Exercise all engines + context creation
Fix build for intel_dp_compliance
igt: Import basic amdgpu tests from libdrm
igt: Add basic amd_cs_nop
igt: Add amd_prime for checking prime integration
igt/kms_frontbuffer_tracking: Declare dependency on functioning GEM
lib: spinning batches requires working GEM, so add the precheck
igt/gem_exec_schedule: Emit wide batches of requests
igt/amdgpu: Fix compiler errors
igt/gem_render_linear_blits: Add missing igt_require_gem()
igt/gem_workarounds: Mark up the requirement for an active GPU
igt/kms_cursor_legacy: Mark up busy tests as requiring a GPU
Add gem_exec_await
intel_reg: Use the drm device fd as a hint only
igt/gem_mmap_gtt: Check that userspace clflushing of the GTT mmap
wsim: Per-client prng pool for miscellaneous randoms
wsim: per-engine throttling
wsim: Send a periodic depth stamp down each queue
wsim: Add a small tolerance to rt balancing
wsim: Improve rt balancer to use history across sync points
wsim: Add rtavg balancer
wsim: Introduce verbosity
wsim: qd throttling now works independently of balancer->get_qd
wsim: Loop over the multiple u32 reads from the status page
wsim: Fixup breaking the read loop after the seqno advances.
wsim: Convert the RT multiple reads to a latched read
wsim: Compact the per-engine heartbeat into a single buffer
autotools requires headers to be listed
Revert "lib/igt_kms: Force outputs to use full range RGB"
wsim: Stop treating wrk->status_page as just a single uint32_t
wsim: Avoid the workload_step being tracked simultaneously on multiple engines
wsim: Cache the heartbeat batch and location
wsim: Limit heartbeats to sync points
wsim: Limit the information updated during the heartbeat
wsim: Switch off heartbeat by default
wsim: Feed qd into ewma
wsim: Use a loop over engines to calculate RT
overlay: Fixup new layout of tracepoints
overlay: Convert to per-context seqno tracking
intel_error_decode: Fix off-by-one when dumping the binary objects
igt/gem_exec_reloc: Exercise relocations across the full GTT range
igt/gem_exec_reloc: Filter out unavailable addresses for !ppgtt
intel_error_decode: Tell zlib the correct amount of memory we allocated
wsim: Set the seqno/time stamp on each batch to every engine
igt/gem_exec_nop: Include the impact of signaling a fence
igt/gem_exec_nop: Restore check on available signal rings
igt/amd_cs_nop: Spread submissions across multiple processes
amdgpu/amd_cs_nop: Use a counter not seqno
igt/gem_exec_store: Add welcome screen
wsim: Only require execbuf wr ioctl for FENCE_OUT
wsim: Fix reporting of workload/s for slaves
wsim: Allow assigning priorities to each workload
igt/gem_spin_batch: Avoid interleave throttle within open spin batches
lib: Refactor testing for ability to use MI_STORE_DATA_IMM
igt/gem_exec_fence: Skip await test if store-dword is not support
igt/pm_rps: Allow CUR to be greater than MAX (overclocking)
igt/gem_exec_store: Exercise more patterns for MI_STORE_DWORD_IMM
igt/gem_exec_store: And free the new obj/reloc arrays
igt/gem_exec_latency: Check for WC mmaps before use
igt/gem_mmap_wc: Move the require mmap-wc to the fixture
igt/gem_shrink: Exercise concurrent calls to i915_gem_shrink()
Revert "kms_cursor_legacy: Add a burner thread to make basic-busy-* pass."
Daniel Stone (1):
tests/kms_*: Use correct DRM context version
Daniel Vetter (1):
tests/kms_properties: Add GET_PROPERTY ioctl sanity check
Eric Anholt (1):
igt/vc4_dmabuf_poll: Add a test for polling to wait for dmabuf fences.
Gabriel Krisman Bertazi (1):
kms_frontbuffer_tracking: Don't poke compressing status for old cpus
Harry Wentland (1):
tests/kms_setmode: Dynamic crtc/connector combinations
Jari Tahvanainen (1):
tests/feat_profile.json: legacy features list for piglit summary feature
Jason Ekstrand (1):
aubdump: Don't bail if a GEM handle of 0 is passed into execbuf
Lukasz Fiedorowicz (1):
lib: Moving gem_execbuf_wr to ioctl_wrappers
Maarten Lankhorst (9):
kms_properties: Blacklist legacy link-status connector property.
lib/igt_kms: Do not refresh all outputs in igt_display_commit
tests/kms_cursor_legacy: Increase tolerance for failing flip-before/after-cursor.
tests/kms_cursor_legacy: Fix failing tests harder
tests/kms_cursor_legacy: Prepare for non-square rotation tests
tests/kms_rotation_crc: Implement stricter rotation tests
tests/kms_rotation_crc: Keep primary plane enabled while testing.
tests/kms_properties: Allow setting all connector properties now.
kms_cursor_legacy: Add a burner thread to make basic-busy-* pass.
Marta Lofstedt (1):
igt/meta_test: Fix dmesg-warn test
Marta Löfstedt (1):
tests/meta_test: Add a meta test for sanity checks of CI systems
Mika Kahola (2):
tests/kms_atomic: test that TEST_ONLY does not clobber state
tests/kms_concurrent: Concurrent and interruptible subtests for atomic
Mika Kuoppala (1):
tests/gem_spin_batch: Add multiengine test
Oscar Mateo (1):
tests/pm_sseu: Re-enable the test
Petri Latvala (18):
Update MAINTAINERS file
igt_command_line.sh: Handle the special cases of drv_selftest and drm_mm
igt_command_line.sh: Actually check things during distcheck
lib/igt_kmod: Don't call igt_assert or igt_require without a fixture
intel-ci: Add extended.testlist for wider testing
benchmarks: Add gem_wsim to .gitignore
Revert "autotools requires headers to be listed"
benchmarks: Make sure ewma.h and ilog2.h are bundled in dist
Revert "Revert "lib/igt_kms: Force outputs to use full range RGB""
configure.ac: Define HAVE_CHAMELIUM in config.h
tests/Makefile.sources: Remove unused XFAIL_TESTS
tests: Merge single_kernel_tests and multi_kernel_tests in the build system
intel-ci: Add comments on test order to fast-feedback.testlist
tests/Makefile.am: Only ignore generated gitignore in gitignore
intel-ci: Have the directory included in the distributed tarballs
intel-ci: Document the public CI results web page
Fix a couple of typos CLFAGS -> CFLAGS
Update NEWS, bump version to 1.19.
Radoslaw Szwichtenberg (1):
tests/pm_rps: Replace custom load function with dummy load
Robert Foss (1):
tools: Remove accidentally included binary intel_bios_reader
Terrence Xu (1):
Fix the mac address definition typo in tools/intel_gvtg_test.
Tomeu Vizoso (3):
lib: Open debugfs files for the given DRM device
igt/kms_pipe_crc_basic: Don't assert on errno after success
Revert "lib: Open debugfs files for the given DRM device"
Tvrtko Ursulin (59):
igt/gem_exec_nop/headless: Verify GT performance in headless mode
benchmarks/gem_wsim: Command submission workload simulator
igt/scripts: trace.pl to parse the i915 tracepoints
benchmarks/gem_wsim: Fix no reloc handling
gem_wsim: Add some generic media workloads
gem_wsim: Simplify batch creation
gem_wsim: Support multiple dependencies
gem_wsim: Support VCS2 remapping
gem_wsim: Fix implicit sync on last workload step
gem_wsim: More simulated transcoding workloads
gem_wsim: Add RTR balancer
gem_wsim: Enable initial per-workload round-robin VCS engine assignment
gem_wsim: Seed random numbers per client
gem_wsim: Allow symbolic balancer selection
gem_wsim: Slightly more robust workload parsing
gem_wsim: Fix master workload handling and stats
gem_wsim: Refactor balancer selection and help text
igt/trace.pl: Collect perf data in quiet mode
gem_wsim: Minimize startup gap
gem_wsim: Two small tidies
gem_wsim: Add append workload
igt/media-bench.pl: Media workload analyzer
gem_wsim: Simplify batch offset block a bit
gem_wsim: Fence support
gem_wsync: Clearer step metadata handling
gem_wsim: Implement sw sync point support
media-bench: Fix scoreboard range stat
media-bench: Fix evaluating with no balancing
media-bench: Flag workloads which failed to balance
media-bench: Store the trace file for later
media-bench: Simplify combined scoring
media-bench: Print score range when evaluation workloads
gem_wsim: Fix client exit with more than one background workload
gem_wsim: Add missing help text for -p
gem_wsim: Convert clients to threads
gem_wsim: Move seqno allocation and query into helpers
gem_wsim: Move seqno status page access to helpers
gem_wsim: Add global balancing mode
media-bench.pl: Add global balancing mode
trace.pl: Add aggregate GPU idle/busy stat
trace.pl: Option to draw the aggreate GPU busy timeline
media-bench.pl: Add option to specify workloads directly
media-bench.pl: Support GPU aggregate idle stats
media-bench.pl: More tweak to saturation point finding
igt: Fix detection of missing flex
gem_wsim: Added missing -H help text
gem_wsim: Add static context balancer
media-bench.pl: Add context balancer to the list
gem_wsim: Late balancing decision mode
media-bench.pl: Include gem_wsim -d in the evaluation
gem_wsim: Detect invalid duration range
media-bench.pl: Verbose option
media-bench.pl: Add GT2 mode
gem_wsim: Fix heartbeats in GT2 mode
gem_wsim: Asymmetrical 1-to-n workloads
gem_wsim: Add some randomness to the 17i7 workload
gem_wsim: More interesting workloads
trace.pl: Handle multi-day uptime
media-bench.pl: Add automatic per-client target mode
Ville Syrjälä (2):
tests/kms_cursor_legacy: make_busy() before get_vblank()
tests/kms_cursor_crc: Test non-square cursors on IVB+
git tag: intel-gpu-tools-1.19
https://xorg.freedesktop.org/archive/individual/app/intel-gpu-tools-1.19.tar.bz2
MD5: 4fdfa56acca3b046fc61fb12686656f3 intel-gpu-tools-1.19.tar.bz2
SHA1: c7ad03b1637db6749690b22a366d675471dc353e intel-gpu-tools-1.19.tar.bz2
SHA256: a49c8ae7c0431c1f33cd9492f83ec4e06e98a48429178e4bd22d9a1a87e3b0f1 intel-gpu-tools-1.19.tar.bz2
PGP: https://xorg.freedesktop.org/archive/individual/app/intel-gpu-tools-1.19.tar.bz2.sig
https://xorg.freedesktop.org/archive/individual/app/intel-gpu-tools-1.19.tar.gz
MD5: d5e6a5f6e29b419bab60d7de55f79282 intel-gpu-tools-1.19.tar.gz
SHA1: 2184babf6cc60e6f0c84a4a44fe42eedbb817d6d intel-gpu-tools-1.19.tar.gz
SHA256: 70f6ecb45c5689a384402b9f103cf748c263d424a1cbc9bf11ce01b8082388da intel-gpu-tools-1.19.tar.gz
PGP: https://xorg.freedesktop.org/archive/individual/app/intel-gpu-tools-1.19.tar.gz.sig
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/intel-gfx/attachments/20170609/8f54103f/attachment-0001.sig>
More information about the Intel-gfx
mailing list