[Intel-gfx] [ANNOUNCE] intel-gpu-tools 1.17

Marius Vlad marius.c.vlad at intel.com
Fri Dec 2 17:03:03 UTC 2016


Awesome. Looks good to me and Thanks!

On Fri, Dec 02, 2016 at 06:19:40PM +0200, Petri Latvala wrote:
> A new intel-gpu-tools quarterly release is available with the
> following changes:
> 
> Library changes:
> 
> - Added an iterator that generate primes for creating input data that
>   should not fall into any patterns that may be optimised by the
>   drivers. (Chris Wilson)
> 
> - Crashes in fixture blocks now print a stacktrace. (Marius Vlad)
> 
> - Added support for various system suspend/resume options. (Imre Deak)
> 
> - Added linked list helpers from the Wayland project. (Lyude)
> 
> - Added a generic dummy workload helper for submitting GPU workloads
>   that consume exactly a specified amount of time. (Abdiel Janulgue)
> 
> - Added C functions for driver loading/unloading, pkill and lsof, for
>   converting shell script tests to C code. (Marius Vlad)
> 
> Tools changes:
> 
> - intel_reg: Add Kabylake support. (Jani Nikula)
> 
> - intel_bios_reader: Also dump PSR info. (Ville Syrjälä)
> 
> - intel_guc_logger: New tool for capturing logs from the GuC
>   firmware. (Akash Goel)
> 
> - intel_aubdump: Added commandline option to stream the dump to
>   another process. (Lionel Landwerlin)
> 
> - intel_aubdump: Annotate the dump with the application name and the
>   used PCI ID. (Jason Ekstrand)
> 
> Benchmark changes:
> 
> - gem_latency: Added support for measuring fence wakeup latencies. (Chris Wilson)
> 
> - prime_lookup: New microbenchmark for stressing prime_fd_to_handle
>   and prime_handle_to_fd. (Chris Wilson)
> 
> Test changes:
> 
> - Multiple new tests.
> 
> - Added an explicit list of tests used for Intel CI. (Petri Latvala)
> 
> - Converted multiple shell script tests to C. (Marius Vlad)
> 
> 
> And many other bug fixes and improvements.
> 
> 
> And the full changelog follows:
> 
> Abdiel Janulgue (8):
>       lib: Make signal helper definitions reusable
>       lib: add igt_dummyload
>       igt/gem_wait: Use new igt_spin_batch
>       igt/kms_flip: Use new igt_spin_batch
>       igt/kms_busy.c: Use new igt_spin_batch
>       igt_dummyload: clear signal handler on the desructor as well
>       igt/kms_flip: Fix set_dpms called with an idle bo
>       igt_dummyload: Don't clear handler for invalid signal
> 
> Akash Goel (1):
>       This patch provides a test utility which helps capture GuC firmware logs and
> 
> Chris Wilson (99):
>       benchmarks/gem_latency: Measure fence wakeup latencies
>       benchmarks/gem_busy: Merge all the sync_file fences together
>       igt/gem_exec_nop: Relax assertion for parallel execution
>       igt/gem_exec_nop: Check submitting nops to each engine in parallel
>       igt/gem_exec_nop: Refine the target calculation
>       igt/gem_ctx_switch: Increase execution weight
>       igt/gem_exec_latency: Initial sketch for measuring execbuf cost in GPU cycles
>       igt/gem_exec_nop: Show more timing details
>       igt/gem_exec_whisper: Add competing child processes
>       igt/gem_busy: Replace extended busy-ioctl testing in bat with simpler tests
>       igt/gem_busy: Actually flag the hang tests to cause a GPU hang
>       igt/gem_busy: Prevent banning when running multiple hang tests
>       igt/kms_cursor_legacy: Reduce flip/cursor ordering to basics
>       lib/kms: Probe connectors from igt_enable_connectors()
>       lib: Add an iterator to generate primes
>       lib: Export igt_debugfs_mount()
>       lib: Add support for DROP_FREED in igt_drop_caches_set()
>       lib: Add igt_primes.h to SOURCES
>       igt/prime_vgem: Initialise pfd for both parent/child
>       igt/gem_concurrent_all: Leave bufmgr destroyed before forking
>       igt/gem_concurrent_blit: Limit number of reads for dmabufs
>       benchmarks: Add some missing .gitignore binaries
>       igt/gem_concurrent_blit: Only set the pixels to be compared (GTT paths)
>       benchmarks: Add prime_lookup microbenchmark
>       benchmarks/gem_blt: Remove () causing integer overflow
>       igt/vgem: Add a basic test for handling of module unload
>       overlay: Silence compiler warning for unused i81x_info
>       igt/gem_userptr_blits: Tidy up the debug spam
>       igt/gem_exec_whisper: Hang injection
>       igt/kms_cursor_legacy: Check that a busy flip doesn't block the cursor
>       igt/vgem_basic: Allow mmap() to persist without the vgem module loaded
>       igt/kms_cursor_legacy: Deduplicate make_busy()
>       igt/gem_exec_whisper: Tweaking the global seqno is not multiuser safe
>       igt/gem_ctx_threads: Repeat threads until timeout
>       igt/gem_ctx_thrash: Update context size estimates
>       igt/gem_ctx_thrash: Include with-execlists indicator
>       Remove igt/vgem_reload_basic
>       And remove vgem_reload_basic form Makefile.sources
>       igt/gem_wait: Use explicit timers
>       igt/gem_exec_big: Secure dispatch is run through the GGTT, limit it such
>       lib/igt_gt: Use /sys/.../error instead of /debug/.../i915_error_state
>       igt/gem_wait: munmap the batch along all paths
>       igt/gem_ctx_param: Tidy error messages
>       igt/gem_ctx_param: Update invalid parma number
>       lib/sysfs: Use a fallback for builtin modules
>       benchmarks/gem_blt: Directly compare the cmdparser
>       tools/l3_parity: Skip on ivb with only a single slice
>       igt: Trim timeouts for basic GEM tests
>       drmtest: Install exithandler for i915, even in multidriver tests
>       igt: Check the physical swizzle status
>       igt/gem_tiled_pread_basic: Only print the erroneous location
>       igt/gem_exec_flush: Seed a GGTT mappable address
>       igt/kms_frontbuffer_tracking: Don't fail if there is no hw overlay
>       igt/drv_hangman: Use manual error-state generation
>       igt/gem_exec_parse: Double check hw validation
>       igt/drv_hangman: Skip format expectations for compressed output
>       igt/kms_flip: Print more information to debug for flip timing checks
>       igt/gem_mmap_gtt: Document the "expected" behaviour
>       igt/kms_setmode: Clear all connectors before starting the test
>       igt/kms_flip: Calibrate timestamp errors
>       igt/kms_setmode: Test that the vblank interval matches the dotclock
>       igt: drop gem_storedw_loop from BAT
>       intel-ci: CI is missing any wait tests in BAT
>       igt/gem_ctx_create: Exercise allocating as many contexts as we can
>       igt/kms_flip: Mark frame_time() as coming from the mode
>       igt/kms_flip: Use the computed vblank interval for TS checking
>       igt/gem_exec_reloc: Check we write the full 64bit relocation
>       igt/gem_mmap_gtt: Use SSE4.1 streaming to speed up per-pixel checks
>       igt: Add test case for EXEC_OBJECT_ASYNC
>       igt: Add exerciser for execbuf fence-out <-> fence-in
>       igt: Add a test for reordering execbufs
>       igt/gem_exec_reloc: Reduce being interrupted to a warn
>       igt/gem_exec_parse: Exercise lots of different allocation sizes
>       igt/kms_frontbuffer_tracking: Drop unsynchronized pageflip test
>       igt/gem_mmap_gtt: Add forked contention to the mmap copies
>       igt/gem_mmap_gtt: Swap thrash copy
>       igt/gem_exec_schedule: Update param for final
>       igt/gem_exec_reloc: Trivial test for softpin ABI
>       igt/gem_eio: Add inflight test
>       igt/gem_exec_reloc: Don't execute an empty object for basic-softpin
>       igt/gem_eio: Apply the cork to the execbuf
>       igt/gem_mmap_gtt: Silence compile warning for !sse4.1
>       igt/gem_seqno_wrap: Tidy assertion failure message for reading i915_next_seqno
>       intel_ci: Add gem_exec_reloc/basic-softpin to BAT
>       igt/kms_setmode: Adjust timing assertion from percentage to scanline
>       igt: Add kselftest runner for drm_mm
>       igt/kms_setmode: Skip unconnected connectors
>       igt/drv_missed_irq: Initialise expect_rings before use
>       igt/gem_gtt_speed: Spare clang the embarrassment
>       igt/drv_hangman: Promote nsec calc to u64
>       Silence a bunch of "const discard" warnings
>       lib/dummyload: Clear the right signal afterwards
>       Revert "igt_dummyload: Don't clear handler for invalid signal"
>       igt/kms_busy: Add more information about the state of the bo
>       lib/dummyload: Don't clear the same signal twice
>       lib/igt_dummyload: Don't clear a signal if we haven't installed the timeout
>       lib/igt_kmod: Adopt igt_kselftests()
>       igt: Add kselftest runner for i915
>       lib/igt_kmod: Squelch the igt assert for a module with no subtests
> 
> Daniel Vetter (1):
>       lib/chipset: Properly skip on non-Intel
> 
> Derek Morton (1):
>       benchmarks/gem_busy: Fix compile error
> 
> Emil Velikov (3):
>       lib: use the local intel_aub.h file
>       lib: use the local intel_aub.h file
>       configure.ac: correctly manage DRM_INTEL_{CFLAGS, LIBS}
> 
> Francisco Jerez (1):
>       aubdump: Fix GTT setup for Gen8+
> 
> Imre Deak (4):
>       lib/igt_aux: Add support for various system suspend/resume options
>       tests/gem_exec_suspend: Add basic S3/S4-devices subtests
>       lib/igt_aux: Improve documentation for igt_system_suspend_autoresume()
>       intel-ci: Add gem_exec_suspend/S4-devices to BAT
> 
> Jani Nikula (11):
>       tools/intel_reg: add kabylake register spec file
>       igt/drv_module_reload_basic: add helper for checking module reloaded
>       igt/drv_module_reload_basic: let snd_hda_intel removal errors through
>       igt/drv_module_reload_basic: let intel_ips removal errors through
>       tests: prefix IGT kernel log messages with [IGT]
>       tests: add more checks for finding the debugfs in script based tests
>       tests: add facility to enable/disable hda dynamic debug in script based tests
>       tests: enable hda dynamic debug for module reload test
>       tests: debug log lsmod on module unload failure
>       tests: do snd_hda_intel unbind before unload in module reload test
>       tests: add more debugs on failure to unload snd_hda_intel
> 
> Jason Ekstrand (5):
>       aubdump: Handle 48-bit relocations properly
>       aubdump: Bump the GTT size to 256MB
>       Revert "aubdump: Bump the GTT size to 256MB"
>       aubdump: Set the application name
>       aubdump: Add a PCI-ID= entry to the AUB file comment
> 
> Kenneth Graunke (1):
>       aubdump: Fix intel_aubdump -o <filename>
> 
> Kristian H. Kristensen (3):
>       assembler: Fix assert when setting 3src sources
>       assembler: Set 3src options before sources
>       assembler: RSQ is a math function
> 
> Lionel Landwerlin (4):
>       tools: intel_aubdump: pass configuration through file descriptor
>       aubdump: add --command option to stream aubdump to another program
>       tests: kms_pipe_color: fix ctm tests
>       tests: kms_pipe_color: remove unused variable
> 
> Liu Ying (2):
>       kms_atomic: Expose atomic commit flags to crtc_commit_atomic()
>       kms_atomic: plane_primary: Allow modeset if atomic check returns -EINVAL
> 
> Lyude (1):
>       igt_aux: Add some list helpers from wayland
> 
> Maarten Lankhorst (13):
>       kms_properties: Add invalid property test.
>       kms_cursor_legacy: Add cursor crc tests.
>       kms_properties: Invalid atomic properties should return -ENOENT
>       igt_kms: Refresh output when setting pipe.
>       Revert "tests/kms_psr_sink_crc: commit before querying mode"
>       kms_cursor_legacy: Add cursor update vs busy page flip test.
>       kms_cursor_legacy: Split out the busy crc test.
>       kms_cursor_legacy: Add nonblocking modeset vs cursor update test.
>       kms_atomic_transition: Respect limits for sprite planes on earlier platforms.
>       kms_panel_fitting: Add some more tests for fastset testing.
>       kms_atomic_transition: Add test timeout to run_modeset_tests
>       kms_atomic_transition: Fix compile
>       kms_atomic_transition: Only run tests on valid pipes.
> 
> Marius Vlad (5):
>       lib/igt_core: Print stacktrace in initialization/fixture blocks.
>       lib/igt_kmod: New library to support driver loading/unloading and additional helpers.
>       lib/igt_gvt: Make use of libkmod helpers and fix reading gvt parameter.
>       tests/drv_module_reload: Convert sh script to C version.
>       tests/kms_sysfs_edid_timing: Convert sh to C version.
> 
> Mika Kahola (1):
>       tests/kms_plane_multiple: CRC based atomic correctness test
> 
> Mika Kuoppala (2):
>       tests/kms_psr_sink_crc: commit before querying mode
>       lib: substitute cxt BAN_PERIOD with BANNABLE
> 
> Petri Latvala (10):
>       lib: Do two writes to /proc/sys/vm/drop_caches again
>       Update MAINTAINERS file
>       tests: Add explicit test lists for CI use.
>       intel-ci: Remove kms_psr_sink_crc at psr_basic from BAT
>       intel-ci: Remove vgem_reload_basic
>       tools: Add intel_guc_logger to .gitignore
>       benchmarks: Add prime_lookup to .gitignore
>       lib: Add missing include for sync()
>       tools: Add intel_gem_info to .gitignore
>       Update NEWS, bump version to 1.17.
> 
> Robert Bragg (12):
>       igt/gem_exec_parse: some minor cleanups
>       igt/gem_exec_parse: move hsw_load_register_reg down
>       igt/gem_exec_parse: update hsw_load_register_reg
>       igt/gem_exec_parse: make global vars local to main()
>       igt/gem_exec_parse: init global parser_version in fixture
>       igt/gem_exec_parse: req. v < 9 for oacontrol tracking test
>       igt/gem_exec_parse: make basic-rejected version agnostic
>       igt/gem_exec_parse: update bitmasks test for v >=8
>       igt/gem_exec_parse: update cmd-crossing-page for >= v8
>       igt/gem_exec_parse: update hsw_load_register_reg for v >= 8
>       igt/gem_exec_parse: update registers test for v >= 8
>       igt/gem_exec_parse: check oacontrol lri bad for >= v9
> 
> Rodrigo Vivi (1):
>       lib/igt_aux.h: Cast is confusing old gcc
> 
> Tomeu Vizoso (6):
>       lib/drmtest: Take DRIVER_ANY into account when opening the DRM device
>       lib/drmtest: Test for equivalence with DRIVER_ANY
>       Revert "lib: use the local intel_aub.h file"
>       lib: Pass tiling constant where that's expected
>       lib: Pass I915_TILING_Y to the kernel if Yf or Ys
>       lib: Pass I915_TILING_NONE if Yf or Ys
> 
> Ville Syrjälä (7):
>       tools/intel_bios_reader: Dump PSR block from VBT
>       tests: Use bash for debugfs_wedged and drm_lib.sh
>       tests: Leave basic breadcrumbs in dmesg for shell script based tests
>       tests/vgem_reload_basic: Leave breadcrumbs in dmesg
>       tests/tools_test: Fix it up for intel_reg
>       tools/intel_reg: Return SUCCESS after a succesful dump
>       build: Fix assmebler/etc. tools build for 32bit x86
> 
> marius vlad (1):
>       demo/Makefile.source: Compile intel_sprite_on when HAVE_LIBDRM_INTEL.
> 
> git tag: intel-gpu-tools-1.17
> 
> https://xorg.freedesktop.org/archive/individual/app/intel-gpu-tools-1.17.tar.bz2
> MD5:  f107b8d909243769f8bb8114ccda8531  intel-gpu-tools-1.17.tar.bz2
> SHA1: 2837571d395a765d03feb3e85c6f1f1cf27af661  intel-gpu-tools-1.17.tar.bz2
> SHA256: a6cbc4e0e52ea87c082987d1bb90a354f8ffac94a31b7b4cd48b38b7f4aafb1a  intel-gpu-tools-1.17.tar.bz2
> PGP:  https://xorg.freedesktop.org/archive/individual/app/intel-gpu-tools-1.17.tar.bz2.sig
> 
> https://xorg.freedesktop.org/archive/individual/app/intel-gpu-tools-1.17.tar.gz
> MD5:  e68c7484952766e521009a803a907dad  intel-gpu-tools-1.17.tar.gz
> SHA1: e24cc5c401afa33d8322134626b42c4d653cb680  intel-gpu-tools-1.17.tar.gz
> SHA256: 0d8407777afcd5252b950f21e8cfe0a3167ed7cc36232ebcd12fb58f8d49f21a  intel-gpu-tools-1.17.tar.gz
> PGP:  https://xorg.freedesktop.org/archive/individual/app/intel-gpu-tools-1.17.tar.gz.sig
> 



> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx



More information about the xorg mailing list