Jens Axboe [Fri, 22 Jun 2007 18:41:27 +0000 (20:41 +0200)]
netsplice: use zero-copy if possible
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Fri, 22 Jun 2007 18:40:10 +0000 (20:40 +0200)]
netsplice: fix a few typos
It actually works now.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Fri, 22 Jun 2007 13:42:21 +0000 (15:42 +0200)]
Support for 'netsplice' engine
This IO engine does network transfers, but uses splice() to move data
to/from a socket to a pipe and vmsplice() to move that data to/from
user space.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 19 Jun 2007 12:25:46 +0000 (14:25 +0200)]
guasi engine: remove STFU_GCC()
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 19 Jun 2007 10:49:52 +0000 (12:49 +0200)]
Fio 1.16.5
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 19 Jun 2007 10:48:41 +0000 (12:48 +0200)]
Adaptive io latency distrobution "histogram"
Log ranges all the way down to microseconds, and display the most
appropriate values for those ranges.
This makes the latency usable for sub-milisecond devices as well.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 19 Jun 2007 07:53:31 +0000 (09:53 +0200)]
HOWTO: note that clat/slat can be in micro or mili seconds
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 19 Jun 2007 07:50:28 +0000 (09:50 +0200)]
io_u_mark_depth: just set index, don't fall through
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 19 Jun 2007 07:48:44 +0000 (09:48 +0200)]
Cleanup msec <-> usec lat conversion
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 18 Jun 2007 07:48:57 +0000 (09:48 +0200)]
libaio engine: print warning for depth > 1 and buffered IO
Linux aio doesn't support queued buffered IO, it has to be
O_DIRECT. So print a warning to that effect, so that users
of fio don't get bad results by mistake.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 18 Jun 2007 07:41:23 +0000 (09:41 +0200)]
Make completion and submission latency use appropriate time base
It's hardcoded as msecs right now, switch the accounting to usecs
and print out the appropriate base determined by the value range.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Fri, 15 Jun 2007 08:33:49 +0000 (10:33 +0200)]
Add cpus_allowed option
cpumask can be cumbersome to use, since it requires you to calculate
a decimal mask to pass to fio. So add a cpus_allowed option that
takes a text list of allowed cpus, ala: cpus_allowed=1,3,8 will set
the affinity mask for CPUS 1, 3, and 8.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 14 Jun 2007 06:17:08 +0000 (08:17 +0200)]
Fio 1.16.4
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 13 Jun 2007 19:00:46 +0000 (21:00 +0200)]
HOWTO: expand the cpumask explanation a bit
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 13 Jun 2007 18:58:09 +0000 (20:58 +0200)]
Merge branch 'master' of ssh://brick.kernel.dk/data/git/fio
Gough, Corey D [Wed, 13 Jun 2007 18:58:01 +0000 (20:58 +0200)]
Fix cpu mask setting
Fio passed the mask by value, not by reference. So when
fill_cpu_mask() was done setting the valid CPUs, it was
only on the local value.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 13 Jun 2007 17:04:20 +0000 (19:04 +0200)]
Merge branch 'master' of ssh://router/data/git/fio
Jens Axboe [Wed, 13 Jun 2007 17:04:16 +0000 (19:04 +0200)]
splice engine: support for vmsplice to user space
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Fri, 8 Jun 2007 09:24:58 +0000 (11:24 +0200)]
splice flags are unsigned int, not unsigned long
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Fri, 25 May 2007 07:26:05 +0000 (09:26 +0200)]
Documentation update
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 24 May 2007 08:34:30 +0000 (10:34 +0200)]
Fio 1.16.3
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 24 May 2007 08:32:47 +0000 (10:32 +0200)]
Tweak meaning of size option
Consider it the full amount of IO to be done, even if the
files used are smaller. Then we'll just restart the workload
until we reach the desired number of bytes.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 23 May 2007 07:34:42 +0000 (09:34 +0200)]
Fix bug with time_based run
We need to clear the FIO_FILE_DONE flag and the number of
files done, to avoid do_io() exiting quickly for a re-run.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 22 May 2007 15:07:44 +0000 (17:07 +0200)]
Fio 1.16.2
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 22 May 2007 15:06:24 +0000 (17:06 +0200)]
Add comment in os/os-linux.h wrt glibc-2.3.2
It was a broken release wrt cpu affinity, they removed the
size argument from the set/get functions. If you want to
compile on such an ancient glibc, manually define GLIBC_2_3_2.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 22 May 2007 13:51:50 +0000 (15:51 +0200)]
Fix build on older systems
Avoid conflict with MINOR/MAJOR.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 22 May 2007 11:27:30 +0000 (13:27 +0200)]
Fix build/compile for non-blktrace platforms
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 22 May 2007 09:12:13 +0000 (11:12 +0200)]
IO verification needs to open and grab a reference to a file
Otherwise it can disappear in midair.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 22 May 2007 07:21:37 +0000 (09:21 +0200)]
Fix bug with random IO and network connections
It isn't a supported configuration.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 22 May 2007 07:13:31 +0000 (09:13 +0200)]
Update HOWTO on cpumask setting
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 22 May 2007 07:13:02 +0000 (09:13 +0200)]
Fix cpumask setting/getting on older kernels
Don't call into the affinity functions, unless the user has
specified a mask.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 21 May 2007 12:52:43 +0000 (14:52 +0200)]
mmap engine: fix open-for-size check
Don't call into mmap() yet, since we didn't setup ->io_size.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 21 May 2007 12:27:32 +0000 (14:27 +0200)]
Attempt to work-around possible time warp
A theory - we see gettimeofday() going backwards on some
systems, try and correct for that by just returning 0
for time passed.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 21 May 2007 11:23:30 +0000 (13:23 +0200)]
Improve random pattern without norandommap
Using the random map, random workloads have a natural tendency to
get less random at the end of the run, due to failures to find a
new unused random offset.
Improve the logic to avoid this problem.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 21 May 2007 09:42:33 +0000 (11:42 +0200)]
Rename ffz() function to fio_ffz() to avoid conflicts
Newer gcc/glibc include an ffz function. Lets just do the easy
thing and rename ours to fio_ffz(), then we can look into reusing
a provided one later if necessary.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 16 May 2007 16:28:47 +0000 (18:28 +0200)]
blktrace support: major/minor fixups
We get passed the kernel variant, so we need a bit of manual
support to untangle it (major() and minor() wont cut it).
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 16 May 2007 12:40:29 +0000 (14:40 +0200)]
blktrace support: support for looking up a device
We can get the device number out of the trace, turn that into
a major/minor pair and look it up in /dev.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 16 May 2007 10:29:06 +0000 (12:29 +0200)]
blktrace support: discard notify messages
They are not interesting to us (they contain things like timestamps
and process information).
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 16 May 2007 10:16:44 +0000 (12:16 +0200)]
blktrace support: fixups/cleanups/improvements
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 16 May 2007 09:54:43 +0000 (11:54 +0200)]
fifo updates
Jens Axboe [Wed, 16 May 2007 06:59:27 +0000 (08:59 +0200)]
Fix timeout on 64-bit BE archs
We stored the timeout in an unsigned long long pointer, but
the value was really unsigned long.
Jens Axboe [Wed, 16 May 2007 07:26:23 +0000 (09:26 +0200)]
Fix fifo leak
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 16 May 2007 07:25:09 +0000 (09:25 +0200)]
blktrace support: speedup reading of data
We used to read in data in really small chunks (48 bytes at the time,
the size of the trace). This is really slow for large traces, so
add a fifo frontend to refill the cache in much larger sizes.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 16 May 2007 06:47:46 +0000 (08:47 +0200)]
blktrace support: adjust largest io_u buffer size to match trace
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 15 May 2007 12:29:58 +0000 (14:29 +0200)]
blktrace replay: delay support
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 15 May 2007 11:23:19 +0000 (13:23 +0200)]
blktrace fixup + documentation
Add a little documentation on using blktrace data files.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 15 May 2007 11:10:41 +0000 (13:10 +0200)]
blktrace support now works
One little buglet - the blktrace offset is the sector, not the
byte offset. For now it's hardcoded at 512, should really get
the value from the device.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 15 May 2007 09:54:21 +0000 (11:54 +0200)]
blktrace support update
This should essentially work, except that it will not delay between
io but just issue the entries as fast as it can. Missing bits:
- Utilize the ipo->delay to suitably delay issuing an io
- Extract device name from blktrace file
- Document it
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 15 May 2007 09:12:19 +0000 (11:12 +0200)]
iolog bug: TD_DDIR_READ should be TD_DDIR_WRITE
It's a typo, if we don't have reads stored but writes, then the
global direction should naturally be TD_DDIR_WRITE.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 15 May 2007 08:13:04 +0000 (10:13 +0200)]
Fix typo in blktrace code
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 15 May 2007 08:12:26 +0000 (10:12 +0200)]
Some more bits of blktrace support
Basic read/write is there. Still need to extract device name from
the trace, and check time delays to wait appropriately.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 15 May 2007 06:44:04 +0000 (08:44 +0200)]
Initial blktrace support
This doesn't work yet, just committing what little bits I did as not
to lose them.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 26 Apr 2007 13:24:20 +0000 (15:24 +0200)]
Log error on io buffer allocation
Currently it just exits silently.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 26 Apr 2007 11:47:07 +0000 (13:47 +0200)]
Allow termination during io_u building
If we are setting up gigabytes of io buffers, it can take many seconds
for that operation to complete. If the user wishes to abort, fio wont
honor that request. Fix that.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 26 Apr 2007 08:22:16 +0000 (10:22 +0200)]
Fio 1.16.1
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 26 Apr 2007 08:21:15 +0000 (10:21 +0200)]
Fix bug with file references
This could cause premature exit of fio with an error at the end of
the run.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 26 Apr 2007 07:02:34 +0000 (09:02 +0200)]
Disk stat improvements
Add an option to disable the disk stats. And only dump the disk
stat header, if we actually have stats to show.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 25 Apr 2007 12:24:12 +0000 (14:24 +0200)]
More iolog fixes
- Warn if nrfiles > 1, so that the user knows it's not supported
- Need to clear ipo on malloc, or ->file could be some random value.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 25 Apr 2007 12:11:18 +0000 (14:11 +0200)]
Fix read_iolog
Caveat - it only works with 1 file, unfortunately. Will need a format
change to get that right.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 25 Apr 2007 06:41:04 +0000 (08:41 +0200)]
Fio 1.16
Mainly minor bug fixes, but also a few features.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 23 Apr 2007 16:54:55 +0000 (18:54 +0200)]
Fio is also available over http://
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 19 Apr 2007 14:31:16 +0000 (16:31 +0200)]
Fix file unlinking
We never set FIO_FILE_UNLINK. Just unlink if it's a regular
file and unlink=1 was given.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 19 Apr 2007 07:53:28 +0000 (09:53 +0200)]
Fix command line parsing bug
The first parameters are for the global job, if no --name section
has been seen yet. We don't require a --name=global section to appear
first.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 19 Apr 2007 07:41:45 +0000 (09:41 +0200)]
Fix --showcmd bug
Only append 'fio' if this is the first time we see a [] section.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 18 Apr 2007 10:46:56 +0000 (12:46 +0200)]
Mask in BE ioprio class as default
We can't just set the ioprio value, we always have to set a class
as well. If no class is given, just assume it's the default best
effort class we want.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 18 Apr 2007 10:25:41 +0000 (12:25 +0200)]
Revert "We do need to send a SIGQUIT to a process/thread, even if it is running"
This reverts commit
ef5fffa591e583de26c6ce14f9448104d6dfbe79.
The problem with the commit is that if we SIGQUIT a thread, it
gets exited without updating runtime/etc info. And then we
don't get a new status display on exit.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 18 Apr 2007 07:02:32 +0000 (09:02 +0200)]
Have the job set ->terminate when runtime is exceeded
Then we don't have to do a seperate check for time_based exit.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 17 Apr 2007 18:14:42 +0000 (20:14 +0200)]
Add time_based option
This allows fio to keep running, even if the workload has completed.
It will simply restart the workload over and over, for as long as the
runtime setting allows.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 17 Apr 2007 17:52:18 +0000 (19:52 +0200)]
Replace random constant with the symbolic name (GR prime)
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 17 Apr 2007 17:48:46 +0000 (19:48 +0200)]
IO buffer allocation cleanup
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 17 Apr 2007 13:46:32 +0000 (15:46 +0200)]
Add zero_buffers option
This option will cause all IO buffers to be initialized to all zeroes.
The default is to fill them with random data.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 17 Apr 2007 13:37:47 +0000 (15:37 +0200)]
Reduntant write bit checks
No point in checking for td_rw(td) || td_write(), since td_write()
will also be true for td_rw().
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 17 Apr 2007 11:42:47 +0000 (13:42 +0200)]
We do need to send a SIGQUIT to a process/thread, even if it is running
It could be stuck in a syscall, and not abort on its own.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 17 Apr 2007 11:31:34 +0000 (13:31 +0200)]
Log and dump short ios
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 17 Apr 2007 11:27:32 +0000 (13:27 +0200)]
If writing to stdout, move stat output to stderr
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 17 Apr 2007 07:06:43 +0000 (09:06 +0200)]
Better checks for disk util init
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 17 Apr 2007 07:05:10 +0000 (09:05 +0200)]
No need to check for type before doing fsync()
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 16 Apr 2007 19:54:24 +0000 (21:54 +0200)]
Add support for using '-' as filename for stdin/stdout
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 16 Apr 2007 19:39:33 +0000 (21:39 +0200)]
Fix problem with f->last_completed_pos not being correct on requeues
This causes an unnecessary seek. Usually this is just noise, but if
the device isn't seekable, it becomes a real problem.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 16 Apr 2007 18:42:33 +0000 (20:42 +0200)]
Add check for pipe/fifo files
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 16 Apr 2007 10:56:00 +0000 (12:56 +0200)]
Warn more gently if the kernel is too old for io scheduler switching
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 16 Apr 2007 07:48:53 +0000 (09:48 +0200)]
Fio 1.15.2
Fixes:
- Segfault on stonewall option issues
- Disk utilization on raw devices on Linux
- Parsing fixes (blank at front/end, comments)
- --showcmd option, turning a job file into a command line
- Update to work with newest GUASI
- File creation problem.
- disk-less IO engine init clean
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 12 Apr 2007 11:02:38 +0000 (13:02 +0200)]
Put the ->real_file_size handling into fio
Then we can remove it from the io engines, where the disk-less IO
engines provided a ->setup() hook just to set that.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Thu, 12 Apr 2007 10:57:27 +0000 (12:57 +0200)]
A ->close_file() operation isn't strictly needed
So be sure to check the pointer before calling it.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 11 Apr 2007 19:20:03 +0000 (21:20 +0200)]
For non-file engines, set ->real_file_size if total size is known
Fixes a segfault with rand* directions.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
gurudas pai [Wed, 11 Apr 2007 10:44:27 +0000 (12:44 +0200)]
Fix bug with stonewall option causing segfault
fio core dumped , with stonewall,
[global]
ioengine=libaio
direct=1
bs=8k
iodepth=1024
iodepth_batch=60
randrepeat=1
size=1024m
directory=/home/oracle
[job1]
rw=randread
filename=file1
[job2]
rw=randwrite
filename=file1
stonewall
[job3]
rw=randread
filename=file1
[job4]
rw=randwrite
filename=file1
>fio myjob
job1: (g=0): rw=randread, bs=8K-8K/8K-8K, ioengine=libaio, iodepth=1024
job2: (g=1): rw=randwrite, bs=8K-8K/8K-8K, ioengine=libaio, iodepth=1024
job3: (g=1): rw=randread, bs=8K-8K/8K-8K, ioengine=libaio, iodepth=1024
job4: (g=1): rw=randwrite, bs=8K-8K/8K-8K, ioengine=libaio, iodepth=1024
Starting 4 processes
Quit (core dumped)PP] [99.3% done] [ 1298/ 0 kb/s] [eta 00m:03s]
(gdb) bt
t_rate=0xfeeee790, m_rate=0xfeeee794) at fio.c:1011
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 11 Apr 2007 10:43:35 +0000 (12:43 +0200)]
Fix problem with treating ENOENT as an error
It's just a zero-sized file, as far as get_file_sizes() is concerned.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Davide Libenzi [Tue, 10 Apr 2007 20:29:29 +0000 (22:29 +0200)]
GUASI API change
There's been an API change in the latest GUASI library, and this fixes it
inside FIO.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 4 Apr 2007 09:09:19 +0000 (11:09 +0200)]
Add a '--showcmd' option for dumping a job file
It effectively shows what the equivelant command line string would be.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 4 Apr 2007 09:04:34 +0000 (11:04 +0200)]
Let parse_options() decide on when parsing failed
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Wed, 4 Apr 2007 09:04:06 +0000 (11:04 +0200)]
Have strip_blank_end() also strip off any comments
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Tue, 3 Apr 2007 23:33:04 +0000 (16:33 -0700)]
Fix sysfs_root assignment with multiple identical devices
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 2 Apr 2007 19:36:44 +0000 (21:36 +0200)]
Move os/arch/compiler headers into directories
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 2 Apr 2007 13:08:01 +0000 (15:08 +0200)]
Seperate disk util code out of fio.c
It's only used if FIO_HAVE_DISK_UTIL is set.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 2 Apr 2007 12:46:07 +0000 (14:46 +0200)]
Fix disk utilization and ioscheduler switch on raw devices
Add a method for mapping a raw device to the real block device,
so our sysfs lookup stuff works as expected.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 2 Apr 2007 09:11:34 +0000 (11:11 +0200)]
Fio 1.15.1
Minor bug fixes, do a maintenance release.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 2 Apr 2007 09:04:47 +0000 (11:04 +0200)]
sg engine: fix old comment
We don't even need to set f to files[0], the loop sets up
the poll appropriately.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 2 Apr 2007 08:51:10 +0000 (10:51 +0200)]
sg engine: do type check only when sd has been setup
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 2 Apr 2007 08:42:33 +0000 (10:42 +0200)]
sg engine: errno return value fixes
It was a little all-over-the-place, unify it - return -1 for error,
let the caller look at errno for the specific error.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Jens Axboe [Mon, 2 Apr 2007 08:36:59 +0000 (10:36 +0200)]
splice engine: return negative for error
Otherwise fio_splice_queue() will think it has handle errno number
of bytes.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>