Jens Axboe [Fri, 9 Mar 2012 16:45:43 +0000 (17:45 +0100)]
client: send probe from fio_client_connect()
We only do it from fio_clients_connect() now, better move it
into the lower per-client handler. This gets probes working
with gfio again.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 16:39:05 +0000 (17:39 +0100)]
gfio: better handle multiple threads/processes in a job
Make the name entry a combo, before we were just overwriting
so it would contain the last entry noted.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 16:24:07 +0000 (17:24 +0100)]
gfio: make results notebook scrollable too
And also adds menus to both results and job notebook.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 16:22:01 +0000 (17:22 +0100)]
gfio: make notebooks scrollable if there are too many
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 16:16:51 +0000 (17:16 +0100)]
gfio: hold a reference to the fio_client structure
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 16:11:04 +0000 (17:11 +0100)]
gfio: more error reporting
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 16:10:54 +0000 (17:10 +0100)]
server: return real error from fio_send_data()
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 16:09:55 +0000 (17:09 +0100)]
client: return more errors
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 16:00:01 +0000 (17:00 +0100)]
client: return errno on connection failure
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 14:14:06 +0000 (15:14 +0100)]
gfio: handle clients individually
Now we have per-client support, don't use any of the "do FOO for
all clients" functions.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 13:59:42 +0000 (14:59 +0100)]
gfio: only start one client net handler
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 13:37:29 +0000 (14:37 +0100)]
gfio: start of support for notebooked jobs
We split struct gui into a main window and per-client part,
having a separate style of layout for each.
This needs a lot of work still, but it gets the point across.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 9 Mar 2012 07:53:28 +0000 (08:53 +0100)]
gfio: ignore group stats for now
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Thu, 8 Mar 2012 20:34:18 +0000 (21:34 +0100)]
Add basic cconv test
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Thu, 8 Mar 2012 19:36:28 +0000 (20:36 +0100)]
Separate the act of adding and starting a job
CMD_JOB or CMD_JOB_LINE only adds the necessary jobs, which will
automatically return the job option info to the client. Let the
client do what it wants with the options, then have the client
send a CMD_RUN command to actually start the jobs.
Next we'll have a command for passing back edited options, so that
the client can update existing options.
This adds a new button to gfio, to note this split of add and run
as well.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Thu, 8 Mar 2012 19:35:56 +0000 (20:35 +0100)]
example: bump netio example to 100g transfer
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Thu, 8 Mar 2012 18:59:16 +0000 (19:59 +0100)]
server: rename CMD_RUN to CMD_SERVER_START
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Thu, 8 Mar 2012 18:45:37 +0000 (19:45 +0100)]
client/server: pass back packed thread_options in add_job()
Now the client can see the parsed options. As it turns out,
we need to parse locally since the client and server can be
vastly different operating systems and architectures. Pass
back the options, allowing the client to change them before
starting the job.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Thu, 8 Mar 2012 16:00:10 +0000 (17:00 +0100)]
gfio: Add y-axis labels
"Kbytes /sec" for bandwidth, and "IOs / sec" for IOPS.
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Thu, 8 Mar 2012 15:58:50 +0000 (16:58 +0100)]
gfio: move graph area a bit to the right
To make more room for y-axis title text.
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Thu, 8 Mar 2012 15:57:29 +0000 (16:57 +0100)]
gfio: fix a bug in the y-axis label shortening algorithm
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Thu, 8 Mar 2012 13:52:25 +0000 (14:52 +0100)]
Add thread_options_pack structure
First step towards having an on-disk and on-wire independent
options structure, with the goal being that we parse options
locally (hence allowing them to be edited sanely) and send them
to the server in one nice package.
Additionally, this is a step towards being able to have a new
fio/gfio "project" file format, which includes all the info a
given client needs. This includes job options, where to connect,
etc.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Thu, 8 Mar 2012 11:37:31 +0000 (12:37 +0100)]
Abstract out the thread_options structure
We need to prepare to have an on-disk/net version of the same.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Thu, 8 Mar 2012 11:30:48 +0000 (12:30 +0100)]
client/server: store/convert/send jobs_eta->nr_threads too
We still need to handle jobs_eta->run_str[] as well...
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Thu, 8 Mar 2012 09:51:36 +0000 (10:51 +0100)]
gfio: improve stop handling
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 19:31:28 +0000 (20:31 +0100)]
gfio: make the end results window a scrolled one
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 19:00:02 +0000 (20:00 +0100)]
gfio: don't use gtk_dialog_get_content_area()
It's not available in gtk 2.10, which we'd like to work with.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 18:51:36 +0000 (19:51 +0100)]
gfio: add missing disk utilization stat to end results
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Wed, 7 Mar 2012 18:40:58 +0000 (19:40 +0100)]
gfio: allow graph to stretch or shrink as window expands or contracts
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Wed, 7 Mar 2012 18:39:55 +0000 (19:39 +0100)]
gfio: right justify y-axis labels
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Wed, 7 Mar 2012 18:37:57 +0000 (19:37 +0100)]
gfio: fix problem with graph finding data range.
In finding minimum value, it used zero as an initializer
but if the graph had no data below say, 1000, it would still
find zero as the minimum, which was incorrect.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Wed, 7 Mar 2012 18:37:25 +0000 (19:37 +0100)]
gfio: get rid of invisible data
At first the invisible data seemed like an elegant way to get
reasonable graph ticks when no data was really present, but it
proved to be an ugly hack that caused other problems later,e.g.
when you want to cap other data so the xy graph scrolls, the
invisible data kept it from scrolling, etc.
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Wed, 7 Mar 2012 18:36:18 +0000 (19:36 +0100)]
gfio: shorten graph tick lables so x000000000 becomes
xG, x000000 becomes xM, and x000 becomes xK
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Wed, 7 Mar 2012 18:34:19 +0000 (19:34 +0100)]
gfio: cap graph data count
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 15:22:50 +0000 (16:22 +0100)]
gfio: change color of graphs
Green for reads, and red for writes. Seems to be what people
use mostly, so lets make it less confusing. Will also want
a title for that.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 15:16:50 +0000 (16:16 +0100)]
gfio: remove a few debug printfs
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 14:21:16 +0000 (15:21 +0100)]
gfio: remove boxes between graphs
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Wed, 7 Mar 2012 13:49:37 +0000 (14:49 +0100)]
gfio: add units to x-axis labels and remove redundant y-axis labels
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Wed, 7 Mar 2012 13:47:38 +0000 (14:47 +0100)]
gfio: make empty graph show grid lines, not "No good data"
This is done by adding a bit of "invisible data" to the graph
at the beginning.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Wed, 7 Mar 2012 13:47:03 +0000 (14:47 +0100)]
gfio: make graph background white
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 13:24:57 +0000 (14:24 +0100)]
gfio: add license
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 13:18:28 +0000 (14:18 +0100)]
gfio: update about dialog
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 12:14:32 +0000 (13:14 +0100)]
gfio: add font selection and resize original window
Font doesn't quite work yet, since we still setup the graphs
before anything is loaded. Will fix that later.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 11:31:40 +0000 (12:31 +0100)]
graph: use unsigned ints
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 10:45:07 +0000 (11:45 +0100)]
gfio: be a bit more clever on when to auto-start backend or not
If we are connecting remotely, don't show the option.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 09:53:29 +0000 (10:53 +0100)]
parser: relax restriction on postfix ranges
Currently we misparse this option:
bs=1k,4k
as read_bs=1, and write_bs=4096, since we compare the postfix
strictly against "k,4k". Use range compares instead.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 08:32:58 +0000 (09:32 +0100)]
Fix batch mismatch for verifies
If we set verify_backlog=32, we expect 32 writes followed by 32
reads that verify the written data. But if we fail getting a
verify io_u, then we may decrement the batch counter when we should
not. The result is that we write 32, read 32, write 1, read 1, repeat.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 08:26:13 +0000 (09:26 +0100)]
Add regression test for
f7078f7b
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 08:24:05 +0000 (09:24 +0100)]
Fix bug that causes early termination of fio with verify_backlog
Running this job:
[global]
bs=4k
ioengine=sync
size=100m
direct=1
filename=xxx
[write]
verify=md5
verify_backlog=32
rw=write
causes fio to exit when reads + writes have exceeded 100m, where
it should only exit when writes have exceeded 100m. Fix this by
checking the appropriate counters based on what type of job it is.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 07:19:30 +0000 (08:19 +0100)]
gfio: add *.ini as fio job file filter
Lots of people seem to name their job files that, so
lets add the filter.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Wed, 7 Mar 2012 07:07:39 +0000 (08:07 +0100)]
gfio: Add graphs for read/write bandwidth and read/write IOPS
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stephen M. Cameron [Wed, 7 Mar 2012 06:56:16 +0000 (07:56 +0100)]
gfio: Add mini library to draw bar graphs and line graphs
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 19:47:15 +0000 (20:47 +0100)]
gfio: clear progress bar on job end
And detach fio server thread.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 19:08:59 +0000 (20:08 +0100)]
gfio: first attempt at actually automatically starting backend server
It doesn't handle multiple connections, after the first job completes,
gfio must be restarted.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 18:34:22 +0000 (19:34 +0100)]
server: fix bad win32 indentation
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 18:26:42 +0000 (19:26 +0100)]
Add regression test for
8a99fdf6
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 18:24:49 +0000 (19:24 +0100)]
Fix another verify segfault
If you run this job:
[global]
ioengine=libaio
direct=1
filename=/tmp/foo
iodepth=128
size=10M
loops=1
group_reporting=1
readwrite=write
do_verify=0
verify=md5
numjobs=1
thread
verify_dump=1
[small_writes]
offset=0G
blocksize=512
verify_interval=1M
[large_writes]
stonewall
offset=1G
blocksize=1M
verify_interval=512
fio crashes because verify_interval is larger than the block size
in the small_writes job. Check for this in get_hdr_inc().
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 18:23:54 +0000 (19:23 +0100)]
Add regression test for verify bug
It was a bug in the job files, not a zone bug after all.
The two block sizes were transposed between the pre and
post job.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 16:46:44 +0000 (17:46 +0100)]
Avoid segfault if verify header appears longer than io_u
Running this job:
[global]
ioengine=libaio
direct=1
filename=/tmp/foo
iodepth=128
size=10M
loops=1
group_reporting=1
readwrite=write
do_verify=0
verify=md5
numjobs=1
thread
verify_dump=1
[small_writes]
offset=0G
blocksize=512
[large_writes]
offset=1G
blocksize=1M
Followed by this job:
[global]
ioengine=libaio
direct=1
filename=/tmp/foo
iodepth=128
size=1M
loops=1
group_reporting=1
readwrite=read
do_verify=1
verify=md5
verify_fatal=1
numjobs=1
thread
verify_dump=1
bs=4k
[large_reads]
offset=0G
blocksize=1M
[small_reads]
offset=1G
blocksize=512
Will cause a segfault, since the bs=512 job appears to stumble
upon a 1mb header, making fio attempt to verify a much bigger
area than what we have read.
This looks like a zone bug. For now, check in a fix that makes
fio validate the verify header a bit harder.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 14:42:31 +0000 (15:42 +0100)]
Add client references
We have a use-after-free in the fio_handle_clients() loop.
If we receive a QUIT command, we remove the client in
fio_handle_client(). But fio_handle_clients() doesn't
have a way to detect this, so it checks client->error
after it has potentially been freed.
Add a simple reference to get rid of this problem.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 14:37:40 +0000 (15:37 +0100)]
server: fixup bug in sending jobs_eta command
We stored nr_running in elapsed_sec, causing ETA to only
be reliable on the client when the job neared completion.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 14:39:15 +0000 (15:39 +0100)]
Add my copyright to gfio.c
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 14:37:40 +0000 (15:37 +0100)]
server: fixup bug in sending jobs_eta command
We stored nr_running in elapsed_sec, causing ETA to only
be reliable on the client when the job neared completion.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 13:00:42 +0000 (14:00 +0100)]
gfio: connect buton fix
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 12:55:45 +0000 (13:55 +0100)]
gfio: enable tree view grid lines
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 12:49:12 +0000 (13:49 +0100)]
gfio: align log text to the left
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 12:46:21 +0000 (13:46 +0100)]
gfio: don't use gtk_combo_box_text
It's not available in gtk 2.18, which is fairly common.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 12:36:17 +0000 (13:36 +0100)]
gfio: make log window scrollable
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 11:55:29 +0000 (12:55 +0100)]
gfio: better handling of connect/disconnect button
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 11:14:42 +0000 (12:14 +0100)]
gfio: add disk utilization output to end results
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 10:02:52 +0000 (11:02 +0100)]
gfio: add log viewer
Replaces old text_op.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 08:52:49 +0000 (09:52 +0100)]
gfio: put results from same job file into a notebook
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 08:37:26 +0000 (09:37 +0100)]
gfio: use a window for results, not a dialog box
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 08:23:27 +0000 (09:23 +0100)]
gfio: ensure all command opcode handlers are properly gdk thread protected
Since we get this callback from our handling thread, we
need to use gdk_threads_enter() and gdk_threads_leave()
to ensure proper locking.
This fixes various crashes.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 08:17:59 +0000 (09:17 +0100)]
gfio: correct depths view
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Tue, 6 Mar 2012 07:09:17 +0000 (08:09 +0100)]
gfio: move bandwidth stats
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 21:07:54 +0000 (22:07 +0100)]
gfio: add os resource usage and io depth display
IO depths is a bit of a hack, needs work.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 20:38:12 +0000 (21:38 +0100)]
gfio: add latency bucket display
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 19:49:10 +0000 (20:49 +0100)]
Move stat arrays into show/calc func
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 19:44:11 +0000 (20:44 +0100)]
gfio: fixup bw display
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 19:04:43 +0000 (20:04 +0100)]
gfio: Change some labels to entries and add bw stats to final output
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 15:43:49 +0000 (16:43 +0100)]
Add completion percentiles to results output
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 13:37:38 +0000 (14:37 +0100)]
End results fixes
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 13:09:03 +0000 (14:09 +0100)]
gfio: start of proper end results data display
Incomplete, it's just a start.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 13:08:34 +0000 (14:08 +0100)]
stat: move calc of latencies into show function
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 12:32:47 +0000 (13:32 +0100)]
Add client references
We have a use-after-free in the fio_handle_clients() loop.
If we receive a QUIT command, we remove the client in
fio_handle_client(). But fio_handle_clients() doesn't
have a way to detect this, so it checks client->error
after it has potentially been freed.
Add a simple reference to get rid of this problem.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 12:08:51 +0000 (13:08 +0100)]
Unify cli latency display
And export usec_to_msec().
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Mon, 5 Mar 2012 08:21:12 +0000 (09:21 +0100)]
Pass in client for thread/group status op
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Sun, 4 Mar 2012 14:15:08 +0000 (15:15 +0100)]
Store specific client ops in the fio_client structure
We could be handling a mix of gui and cli clients from the server,
so passing. Plus this is a lot cleaner.
Also unify ETA opcode handling between clients.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Sat, 3 Mar 2012 19:35:50 +0000 (20:35 +0100)]
Convert net commands in the generic command handler
Removes this from the cli/gui client, it can be handled
generically.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Sat, 3 Mar 2012 19:28:07 +0000 (20:28 +0100)]
Add new net command for text logging
Encapsulate the text in a payload, that also includes the
logging level and the time the event took place. The text
client will do as it always has done, but the gui client
can now do proper logging and knows if this is debug/info
or error messages.
No functional change in the GUI.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 2 Mar 2012 21:23:36 +0000 (22:23 +0100)]
Compression fixes
- memset() remainder of buffer instead of copying random segment.
Provides more reliable compression (yes yes, you told me so).
- Allow 100 as well, just means full memset. Make that just set
zero_buffers=1 instead, since we only need to do it once
instead of for every write.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 2 Mar 2012 20:02:12 +0000 (21:02 +0100)]
Add buffer_compress_percentage
The option is pending testing, so not documented yet.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 2 Mar 2012 18:44:15 +0000 (19:44 +0100)]
Add preferences help tooltips
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 2 Mar 2012 18:34:13 +0000 (19:34 +0100)]
Add Preferences dialog
Allows you to set debug options for the client.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 2 Mar 2012 18:09:31 +0000 (19:09 +0100)]
GUI fixes
- Don't use clear_ui_info() on normal quit
- Remove setting of all debug flags
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 2 Mar 2012 13:27:17 +0000 (14:27 +0100)]
Start of support for auto-starting the backend
When running fio as a client only, you need the backend running
on the system. Start adding support for letting gfio auto-fork
the backend, if it isn't already running.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 2 Mar 2012 13:04:30 +0000 (14:04 +0100)]
Clear UI/progress bar on disconnect/quit
And ensure that server idle loop sends run command once per
idle loop entry.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 2 Mar 2012 12:32:04 +0000 (13:32 +0100)]
Terminate clients when 'Disconnect' is clicked
And reports errors in error bar.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Jens Axboe [Fri, 2 Mar 2012 12:12:04 +0000 (13:12 +0100)]
Move connection details to separate dialog box
No need to view that in the main window, we need the space.
Signed-off-by: Jens Axboe <axboe@kernel.dk>