summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Chris Wilson [Mon, 30 Nov 2009 15:20:29 +0000 (15:20 +0000)]
decode: Split decoder and reuse outside of intel_gpu_dump
Introduce a second decoder that simpler operates on a file so that it
can be run on dumped batchbuffers independent of the gpu.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Wu Fengguang [Fri, 6 Nov 2009 03:06:22 +0000 (11:06 +0800)]
Add: tools/intel_audio_dump
Signed-off-by: Zhenyu Wang <zhenyu.z.wang@intel.com>
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Chris Wilson [Thu, 8 Oct 2009 05:48:02 +0000 (06:48 +0100)]
Add: tools/intel_gpu_time
A simple variant on time crossed with intel_gpu_top to estimate the
amount of GPU activity during a program's execution time.
The resource usage (sys+user, i.e. CPU %) is for the child, but the
GPU measure is global. Hmm.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Eric Anholt [Wed, 7 Oct 2009 00:25:10 +0000 (17:25 -0700)]
Add intel_gpu_dump from the 2D driver.
Eric Anholt [Mon, 21 Sep 2009 22:00:56 +0000 (15:00 -0700)]
Bump to 1.0.2 for release.
Carl Worth [Thu, 17 Sep 2009 22:05:24 +0000 (15:05 -0700)]
Print unmatched opcode value for all 3D UNKNOWN messages
This just simplifies one step for fixing up an incompletely-parsed
dump.
Xiang, Haihao [Thu, 3 Sep 2009 01:50:00 +0000 (09:50 +0800)]
Add support for new chips
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Thu, 3 Sep 2009 01:35:50 +0000 (09:35 +0800)]
CS FENCE in URB_FENCE is 11-bits wide
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Xiang, Haihao [Thu, 3 Sep 2009 01:33:36 +0000 (09:33 +0800)]
Fix the length of CONSTANT_BUFFER
Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>
Eric Anholt [Mon, 20 Jul 2009 18:46:11 +0000 (11:46 -0700)]
intel_gpu_dump: Don't pretend we understand vertex formats on 8xx.
Eric Anholt [Mon, 20 Jul 2009 18:44:56 +0000 (11:44 -0700)]
intel_gpu_dump: Parse LOAD_STATE_IMMEDIATE_2 for 830.
Eric Anholt [Mon, 20 Jul 2009 16:14:43 +0000 (09:14 -0700)]
intel_gpu_dump: Add IPEIR and PGTBL_ER to output.
Eric Anholt [Mon, 20 Jul 2009 16:07:16 +0000 (09:07 -0700)]
Add INSTDONE bits for 830-865.
Eric Anholt [Thu, 16 Jul 2009 21:15:47 +0000 (14:15 -0700)]
intel_gpu_dump: Fix decode for LOAD_STATE_IMMEDIATE_1 with S8 set.
Eric Anholt [Mon, 13 Jul 2009 17:03:17 +0000 (10:03 -0700)]
Add a manpage for intel_reg_write
Chris Wilson [Thu, 3 Sep 2009 09:58:59 +0000 (10:58 +0100)]
dump: Minor typo
Chris Wilson [Tue, 1 Sep 2009 09:09:55 +0000 (10:09 +0100)]
Add gem_fence_thrash test
This is a test case that overcommits fence registers between threads, which
are copying from one fenced bo to another. In earlier versions of the driver
this would cause excessive spinning as the first inactive (i.e. not in use
by the GPU) would be used to service the next page. After all the fence
registers had been allocated, in effect only the very first fence would then
be used for all subsequent faults.
Ben Gamari [Wed, 1 Jul 2009 20:52:20 +0000 (16:52 -0400)]
Add CS_USB_STATE decoding
Ben Gamari [Wed, 1 Jul 2009 20:52:19 +0000 (16:52 -0400)]
Add CONSTANT_BUFFER decoding
Ben Gamari [Wed, 1 Jul 2009 20:52:18 +0000 (16:52 -0400)]
Fix gs_fence output in URB_FENCE output
Ben Gamari [Wed, 1 Jul 2009 20:52:17 +0000 (16:52 -0400)]
add realloc bits to URB_FENCE decoding
Carl Worth [Fri, 8 May 2009 23:04:24 +0000 (16:04 -0700)]
Fix error message for case of kernel prior to 2.6.30
The error message was falsely claiming that a debugfs directory
didn't exist, when it was supposed to report that the ringbuffer
file did not exist within that directory.
Eric Anholt [Wed, 1 Jul 2009 01:05:57 +0000 (18:05 -0700)]
.gitignore: catch up to intel_reg_write addition
Eric Anholt [Wed, 1 Jul 2009 00:52:43 +0000 (17:52 -0700)]
intel_gpu_dump: Add instdone register values and decoding.
Eric Anholt [Wed, 1 Jul 2009 00:13:27 +0000 (17:13 -0700)]
intel_gpu_top: Move instdone bit definitions to lib to share with _dump.
Eric Anholt [Wed, 24 Jun 2009 02:35:19 +0000 (19:35 -0700)]
intel_gpu_dump: Use the correct IPEHR on pre-965.
Eric Anholt [Wed, 24 Jun 2009 01:43:59 +0000 (18:43 -0700)]
intel_gpu_dump: Add details on URB_FENCE.
Eric Anholt [Wed, 24 Jun 2009 01:16:42 +0000 (18:16 -0700)]
Oops, fix up printed hw offset since the HEAD pointer fixup
Eric Anholt [Wed, 24 Jun 2009 00:10:44 +0000 (17:10 -0700)]
Add more 965 INSTDONE bits.
This shows off the units that are stuck busy in the ut2004 hang.
Eric Anholt [Tue, 23 Jun 2009 23:22:41 +0000 (16:22 -0700)]
Correctly flag the acthd when it points within a packet.
Eric Anholt [Tue, 23 Jun 2009 23:21:34 +0000 (16:21 -0700)]
Add the acthd to the GPU dump.
Ben Gamari [Thu, 11 Jun 2009 03:13:40 +0000 (23:13 -0400)]
Add register writing tool
Signed-off-by: Eric Anholt <eric@anholt.net>
Jesse Barnes [Fri, 19 Jun 2009 01:07:47 +0000 (18:07 -0700)]
Four new tests for error handling
Add four new tests for error the error handling cases:
- gem_bad_address - store to a bad address, should generate a protection or
page table error
- gem_bad_batch - try to execute a bad batch, should generate a protection,
invalid instruction or page table error
- gem_bad_blit - blit to an invalid location, should generated a protection
or page table error
- gem_hang - hang the GPU on an event that will never happen, test hang
detection & recovery code
Eric Anholt [Tue, 9 Jun 2009 18:30:25 +0000 (11:30 -0700)]
Make the precentage bars fit in 80 columns.
Signed-off-by: Eric Anholt <eric@anholt.net>
Robert Bragg [Fri, 17 Apr 2009 18:01:04 +0000 (19:01 +0100)]
A pedantic addition of a format string to a printf call
The printf used to clear the screen didn't have a format string, this
adds one to avoid a compiler warning.
Signed-off-by: Eric Anholt <eric@anholt.net>
Robert Bragg [Fri, 17 Apr 2009 17:52:23 +0000 (18:52 +0100)]
Adds unicode bars for quickly visualizing the busy percentages
It can be a bit easier to digest the percentages with bar graphs than by
scanning continually changing numbers.
Signed-off-by: Eric Anholt <eric@anholt.net>
Robert Bragg [Fri, 17 Apr 2009 15:25:52 +0000 (16:25 +0100)]
Increases the sample frequency from 100/sec to 10,000/sec
For a typical vsync locked application running at 60fps, sampling at just
under twice a frame doesn't seem to give very stable lists of relevent hardware
units because there are a number of units involved that may not be sampled one
second to the next.
This bumps the sample rate to 10,000 instead which is ~ 170 samples per
frame so we tend to hit all the units involved.
It also changes the report threshold to a sample count >= 1, so you don't
see as many units with a percentage of 0.
Signed-off-by: Eric Anholt <eric@anholt.net>
Robert Bragg [Sun, 7 Jun 2009 22:26:18 +0000 (23:26 +0100)]
Revert "intel_gpu_top:print a visual clue on how big a certain use is"
This reverts commit
5e13b98da1e7e31cffba84fd257002357d5f2682.
We can print prettier graphs using unicode
Signed-off-by: Eric Anholt <eric@anholt.net>
Eric Anholt [Thu, 4 Jun 2009 14:38:17 +0000 (14:38 +0000)]
intel_gpu_top: Print GPU clock information, which seems interesting.
Arjan van de Ven [Tue, 28 Apr 2009 11:05:43 +0000 (12:05 +0100)]
intel_gpu_top:print a visual clue on how big a certain use is
humans are pretty bad at reading percentages quicky; this patch adds a
histogram capability to make it more visually clear as to which lines are
big ticket items
Eric Anholt [Tue, 19 May 2009 23:48:40 +0000 (16:48 -0700)]
Bump version to 1.0.1 for release.
Eric Anholt [Tue, 19 May 2009 23:32:39 +0000 (16:32 -0700)]
intel_gpu_dump: Stop decoding batchbuffers after MI_BATCH_BUFFER_END.
Eric Anholt [Tue, 19 May 2009 23:26:35 +0000 (16:26 -0700)]
Add decode of gen4 PIPE_CONTROL
Eric Anholt [Tue, 19 May 2009 23:12:43 +0000 (16:12 -0700)]
Add decode of dest buffer variables (destination format)
(cherry-pick from Mesa commit
0af7e9170fd7c0d906652378b9f78fe2ba9725ad)
Eric Anholt [Tue, 19 May 2009 23:09:08 +0000 (16:09 -0700)]
Add manpages for all installed binaries.
Eric Anholt [Tue, 19 May 2009 22:53:55 +0000 (15:53 -0700)]
Add gem_largeobject to ignore.
Jesse Barnes [Thu, 14 May 2009 21:39:48 +0000 (14:39 -0700)]
tests: add large object test
The large object test simply tries to allocate a 128M object, pin it, then
pwrite the whole thing. This should make obvious any leaks on close or
page pointer allocation failures.
Eric Anholt [Mon, 27 Apr 2009 16:45:22 +0000 (09:45 -0700)]
Fix distcheck
Eric Anholt [Mon, 27 Apr 2009 16:41:41 +0000 (09:41 -0700)]
.gitignore libtool junk
Eric Anholt [Mon, 27 Apr 2009 16:39:05 +0000 (09:39 -0700)]
intel_gpu_top: Add ring information.
Eric Anholt [Mon, 27 Apr 2009 16:35:40 +0000 (09:35 -0700)]
Add a test to try to fill the ring.
Did this while trying to debug the Ubuntu DRI1+compiz issue. Didn't succeed
in filling the ring.
Carl Worth [Mon, 13 Apr 2009 21:35:11 +0000 (14:35 -0700)]
intel_gpu_dump: Use asprintf instead of a static-sized array
Not that we're really concerned about buffer-overrun attacks
against this utility, but still, good habits and clean code
are always preferred.
Carl Worth [Mon, 13 Apr 2009 21:33:09 +0000 (14:33 -0700)]
intel_gpu_dump: Give a nicer message for missing batchbuffer-dumping code
This is the case where debugfs is mounted, the dri/0 directory
exists but there's no i915_ringbuffer_info file. Point the user
toward upgrading the driver.
Carl Worth [Mon, 13 Apr 2009 21:06:00 +0000 (14:06 -0700)]
intel_gpu_dump: Mention filename in error message
This helps for when the user makes a typo trying to pass a file
or directory name to intel_gpu_dump.
Carl Worth [Mon, 13 Apr 2009 21:02:04 +0000 (14:02 -0700)]
Add example command for mounting debugfs
This is a nice opportunity to teach the user something, (and help our
users to actually succeed in creating useful bug reports for us).
Eric Anholt [Mon, 13 Apr 2009 01:03:25 +0000 (18:03 -0700)]
intel_gpu_dump: Add no-arguments automatic mode, and clean up comments.
Eric Anholt [Mon, 13 Apr 2009 00:49:31 +0000 (17:49 -0700)]
intel_gpu_dump: Dump ring, batch, and head/tail info when given the directory.
Eric Anholt [Mon, 13 Apr 2009 00:04:31 +0000 (17:04 -0700)]
intel_gpu_dump: Decode the GTT offset of batchbuffers.
Eric Anholt [Sun, 12 Apr 2009 22:57:07 +0000 (15:57 -0700)]
intel_gpu_dump: Get the devid so we can decode correctly on gen4.
Eric Anholt [Wed, 8 Apr 2009 02:07:18 +0000 (19:07 -0700)]
Fix warning about assert in intel_batchbuffer.h
Eric Anholt [Tue, 7 Apr 2009 23:21:16 +0000 (16:21 -0700)]
Add a test for reading tiled objects with pread.
Eric Anholt [Tue, 7 Apr 2009 23:18:11 +0000 (16:18 -0700)]
Add a test for pread after blitting to an object.
This caught a bug with get_user_pages usage in the kernel, which would
result in zeroes being read out of the object when faulting in a new page.
Carl Worth [Tue, 7 Apr 2009 23:22:46 +0000 (16:22 -0700)]
Add very preliminary intel_gpu_dump program
Right now it requires a filename on the command line, (such as
/debug/dri/0/i915_batchbuffers). We plan to make it smarter in
the future by going out and collecting all the data it needs.
Carl Worth [Thu, 2 Apr 2009 18:53:38 +0000 (11:53 -0700)]
Add Eric's batchbuffer-decode code
This comes from the 3D driver within mesa, with identically named files.
In its new life here it will become a standalone program.
Eric Anholt [Fri, 3 Apr 2009 21:23:06 +0000 (14:23 -0700)]
Add a regression test for tiled object blitting.
Eric Anholt [Tue, 31 Mar 2009 02:17:12 +0000 (19:17 -0700)]
Add a couple of other intel_upload_blit_large variants for comparison.
Eric Anholt [Tue, 31 Mar 2009 02:09:00 +0000 (19:09 -0700)]
intel_upload_blit_*: Fix reporting and generate junk instead of using random.
Turns out that random completely overwhelmed the cost of the driver stuff
in _large.c. I definitely want to generate data, but still be able to see
the driver's responsibility.
Eric Anholt [Tue, 31 Mar 2009 02:08:43 +0000 (19:08 -0700)]
Add intel_chipset.h to fix dist.
Eric Anholt [Fri, 27 Mar 2009 23:11:50 +0000 (16:11 -0700)]
Move i810_reg.h to lib/intel_reg.h and get it disted.
Eric Anholt [Fri, 27 Mar 2009 23:04:25 +0000 (16:04 -0700)]
tests/: warning fixes
Eric Anholt [Fri, 27 Mar 2009 23:03:12 +0000 (16:03 -0700)]
rename project
Eric Anholt [Fri, 27 Mar 2009 23:03:02 +0000 (16:03 -0700)]
readme fixups
Eric Anholt [Fri, 27 Mar 2009 20:13:47 +0000 (13:13 -0700)]
intel_gpu_top: Add more 965 bits.
Eric Anholt [Fri, 27 Mar 2009 19:25:09 +0000 (12:25 -0700)]
Port intel_idle from 2D driver as intel_gpu_top with a better interface.
Eric Anholt [Fri, 27 Mar 2009 18:07:33 +0000 (11:07 -0700)]
intel_stepping: Add G45 and GM45.
Eric Anholt [Fri, 27 Mar 2009 17:59:09 +0000 (10:59 -0700)]
Add intel_stepping from the 2D driver.
Eric Anholt [Fri, 27 Mar 2009 01:47:42 +0000 (18:47 -0700)]
Add the tests we want from libdrm.
Eric Anholt [Fri, 27 Mar 2009 01:45:01 +0000 (18:45 -0700)]
Fix project name/version for distfiles.
Eric Anholt [Fri, 27 Mar 2009 01:44:13 +0000 (18:44 -0700)]
Add notes about running things as root.
Eric Anholt [Fri, 27 Mar 2009 00:15:11 +0000 (17:15 -0700)]
Initial import of intel-graphics-tools with some microbenchmarks.