fio.git
12 years agoFix potential crash in terminate_threads()
Jens Axboe [Sat, 1 Mar 2008 14:25:24 +0000 (15:25 +0100)]
Fix potential crash in terminate_threads()

td->io_ops can be NULL, if the thread is already gone. So copy the
pointer and check before dereferencing it.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoWhen retrieving a requeued IO, don't setup data pointers again
Jens Axboe [Sat, 1 Mar 2008 14:22:32 +0000 (15:22 +0100)]
When retrieving a requeued IO, don't setup data pointers again

This is a bug, if we had a residual data count we would be
retrying the full request again.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFix bad inlining in mutex.h
Jens Axboe [Sat, 1 Mar 2008 14:12:48 +0000 (15:12 +0100)]
Fix bad inlining in mutex.h

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFix recursive dump of options
Jens Axboe [Sat, 1 Mar 2008 14:12:14 +0000 (15:12 +0100)]
Fix recursive dump of options

If a child also had a child, we would miss it.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoConfusion between mutex and semaphore naming
Jens Axboe [Wed, 27 Feb 2008 17:58:00 +0000 (18:58 +0100)]
Confusion between mutex and semaphore naming

There's a semaphore implementation and option coming, so rename the
bits.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd --debug=parse for option parsing debug
Jens Axboe [Wed, 27 Feb 2008 17:32:33 +0000 (18:32 +0100)]
Add --debug=parse for option parsing debug

Adds log.h and debug.h to split the logging and debug bits out of
fio, so that the parser can use them.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agofio.1: add section option and remove spurious command text
Aaron Carroll [Tue, 26 Feb 2008 22:10:39 +0000 (23:10 +0100)]
fio.1: add section option and remove spurious command text

Signed-off-by: Aaron Carroll <aaronc@gelato.unsw.edu.au>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoextend_file need continue when the file exists
Zhang, Yanmin [Tue, 26 Feb 2008 14:35:52 +0000 (15:35 +0100)]
extend_file need continue when the file exists

If some testing files exist and some others don't exist, fio will
report failure on the no-exist file. Here is a patch to fix it.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoMove syslet/indirect bits into the OS specific directory
Jens Axboe [Sun, 24 Feb 2008 20:36:00 +0000 (21:36 +0100)]
Move syslet/indirect bits into the OS specific directory

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoRemove dead FIO_FILE_EXISTS flag
Jens Axboe [Thu, 21 Feb 2008 09:20:00 +0000 (10:20 +0100)]
Remove dead FIO_FILE_EXISTS flag

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoprune_io_piece_log() also needs to take list into account
Jens Axboe [Thu, 21 Feb 2008 08:52:35 +0000 (09:52 +0100)]
prune_io_piece_log() also needs to take list into account

If we didn't use the rbtree but have remaining entries to prune,
loop over the list as well and free those entries.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoRemove dead FIO_FILE_NOSORT flag
Jens Axboe [Thu, 21 Feb 2008 08:51:32 +0000 (09:51 +0100)]
Remove dead FIO_FILE_NOSORT flag

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agofix hugepage-size type mismatch
Aaron Carroll [Wed, 20 Feb 2008 08:14:13 +0000 (09:14 +0100)]
fix hugepage-size type mismatch

hugepage_size in thread_options is of type int, but listed in
options.c as type long.  This patch reconciles the two.

Signed-off-by: Aaron Carroll <aaronc@gelato.unsw.edu.au>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoaccount for delayed jobs when reaping threads
Aaron Carroll [Wed, 20 Feb 2008 08:14:12 +0000 (09:14 +0100)]
account for delayed jobs when reaping threads

reap_threads does not account for jobs that have been delayed
(e.g. via startdelay) which may lead to all threads being
terminated.  The following job file demonstrates the problem:

[job1]
filename=job1file
size=16k
startdelay=1

Signed-off-by: Aaron Carroll <aaronc@gelato.unsw.edu.au>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFio 1.19 fio-1.19
Jens Axboe [Tue, 19 Feb 2008 19:24:10 +0000 (20:24 +0100)]
Fio 1.19

1.18x had some stupid bugs with random IO generation, so release
a new full version.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoDon't always print usage info
Jens Axboe [Tue, 19 Feb 2008 19:16:57 +0000 (20:16 +0100)]
Don't always print usage info

For instance, when we ask for help don't complain about missing job.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoMention new --debug=random option in README
Jens Axboe [Tue, 19 Feb 2008 19:11:41 +0000 (20:11 +0100)]
Mention new --debug=random option in README

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoMore random fixes
Jens Axboe [Tue, 19 Feb 2008 19:10:26 +0000 (20:10 +0100)]
More random fixes

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFix TO_MAP_BLOCK() taking offset into account twice
Jens Axboe [Tue, 19 Feb 2008 17:20:14 +0000 (18:20 +0100)]
Fix TO_MAP_BLOCK() taking offset into account twice

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd FIO_SIGQUIT engine option
Jens Axboe [Mon, 18 Feb 2008 20:11:24 +0000 (21:11 +0100)]
Add FIO_SIGQUIT engine option

For network engines, we need a signal to make it abort the connection
if it isn't already running. Otherwise fio would just stall on ctrl-c
before the connection was established.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoCleanup parse_cmd_line()
Jens Axboe [Mon, 18 Feb 2008 19:59:18 +0000 (20:59 +0100)]
Cleanup parse_cmd_line()

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd --section command line option
Jens Axboe [Mon, 18 Feb 2008 19:53:47 +0000 (20:53 +0100)]
Add --section command line option

It enables the user to ship bigger and more complex job files, while
still allowing fio to only run a part of it.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoExit on bad command line options
Jens Axboe [Mon, 18 Feb 2008 19:26:32 +0000 (20:26 +0100)]
Exit on bad command line options

It could be dangerous, eg if you expected --read-only to do its job
when you really needed to type --readonly.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd/correct missing command line options in help text
Jens Axboe [Mon, 18 Feb 2008 19:15:03 +0000 (20:15 +0100)]
Add/correct missing command line options in help text

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFio 1.18.1 fio-1.18.1
Jens Axboe [Mon, 18 Feb 2008 19:00:08 +0000 (20:00 +0100)]
Fio 1.18.1

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFix typo in mark_random_map()
Jens Axboe [Mon, 18 Feb 2008 13:14:53 +0000 (14:14 +0100)]
Fix typo in mark_random_map()

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFix random location generation with offset != 0
Jens Axboe [Mon, 18 Feb 2008 13:13:08 +0000 (14:13 +0100)]
Fix random location generation with offset != 0

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agofio: HOWTO update
Gurudas Pai [Fri, 8 Feb 2008 07:50:14 +0000 (08:50 +0100)]
fio: HOWTO update

In case of  cpuio , on smp machines we need increase numjobs
to <num of cpu's> to get the desired load across cpu's.
Just updated HOWTO with this info.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoVerify doesn't really work on mixed read/write workloads, so warn
Jens Axboe [Wed, 6 Feb 2008 13:40:14 +0000 (14:40 +0100)]
Verify doesn't really work on mixed read/write workloads, so warn

Unless the file is pre-populated, this cannot work. So warn the user
of such a bad construct, since the warning he/she receives from fio
only tells you about a verification error (which, technically, is what
is happening).

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFix misplaced exit(0) in cpus_allowed setting
Jens Axboe [Wed, 6 Feb 2008 13:17:30 +0000 (14:17 +0100)]
Fix misplaced exit(0) in cpus_allowed setting

Must have been a leftover from some debug session. It prevented any
job file with cpus_allowed set from running.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agofio: minor HOWTO fix
Gurudas Pai [Wed, 6 Feb 2008 10:16:15 +0000 (11:16 +0100)]
fio: minor HOWTO fix

writev(3) should be writev(2)

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoman page update
Jens Axboe [Tue, 5 Feb 2008 11:02:07 +0000 (12:02 +0100)]
man page update

--debug command line and bssplit job option missing. Add vsync io engine
and clarify bssrange.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFio 1.18 fio-1.18
Jens Axboe [Tue, 5 Feb 2008 09:23:54 +0000 (10:23 +0100)]
Fio 1.18

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFix offset retrieval problem for read and write having different bs ranges
Jens Axboe [Tue, 5 Feb 2008 09:05:50 +0000 (10:05 +0100)]
Fix offset retrieval problem for read and write having different bs ranges

"Zhang, Yanmin" <yanmin_zhang@linux.intel.com> reported that fio got
stuck with a job file that had different read and write block size
ranges. The problem is that get_next_offset() didn't take the data
direction into account, so when it generated an offset close to the
end of the file, it may not have been legal for that data direction.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoparser: always set off3/4 even if we have more
Jens Axboe [Tue, 5 Feb 2008 08:57:39 +0000 (09:57 +0100)]
parser: always set off3/4 even if we have more

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosync engine: missing fsync check in vsync
Jens Axboe [Mon, 4 Feb 2008 14:58:24 +0000 (15:58 +0100)]
sync engine: missing fsync check in vsync

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd vsync io engine
Jens Axboe [Mon, 4 Feb 2008 09:59:07 +0000 (10:59 +0100)]
Add vsync io engine

It uses readv/writev to transfer the data and coalesces adjacent
data into a single system call (emulating queueing).

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoImprove depth marking
Jens Axboe [Mon, 4 Feb 2008 09:56:26 +0000 (10:56 +0100)]
Improve depth marking

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoDecrement io_issue count when requeuing an io_u
Jens Axboe [Mon, 4 Feb 2008 09:56:07 +0000 (10:56 +0100)]
Decrement io_issue count when requeuing an io_u

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd debug trace for io_u_queue_complete()
Jens Axboe [Mon, 4 Feb 2008 09:49:41 +0000 (10:49 +0100)]
Add debug trace for io_u_queue_complete()

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoDon't allow FIO_SYNCIO io engine flag to override queue depth
Jens Axboe [Mon, 4 Feb 2008 09:48:13 +0000 (10:48 +0100)]
Don't allow FIO_SYNCIO io engine flag to override queue depth

Another bad check, just kill it.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosync engine: cleanup psync/sync seperation
Jens Axboe [Mon, 4 Feb 2008 08:35:30 +0000 (09:35 +0100)]
sync engine: cleanup psync/sync seperation

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoDon't force iodepth == 1 for SYNC io engine
Jens Axboe [Mon, 4 Feb 2008 08:35:03 +0000 (09:35 +0100)]
Don't force iodepth == 1 for SYNC io engine

Most will not do queuing, but allow it to be set in case some of
them do.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFix bug in fio mutex initialization
Zhang, Yanmin [Mon, 4 Feb 2008 08:17:52 +0000 (09:17 +0100)]
Fix bug in fio mutex initialization

Fio sem->cond isn't initialized properly.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoMerge branch 'master' of ssh://git.kernel.dk/data/git/fio
Jens Axboe [Fri, 1 Feb 2008 20:51:29 +0000 (21:51 +0100)]
Merge branch 'master' of ssh://git.kernel.dk/data/git/fio

12 years agoA few debug debug log fixes
Jens Axboe [Fri, 1 Feb 2008 20:51:02 +0000 (21:51 +0100)]
A few debug debug log fixes

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd option for not including the tracing stuff
Jens Axboe [Fri, 1 Feb 2008 19:37:09 +0000 (20:37 +0100)]
Add option for not including the tracing stuff

It eats some space.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd blktrace and verify debug tracing
Jens Axboe [Fri, 1 Feb 2008 19:27:52 +0000 (20:27 +0100)]
Add blktrace and verify debug tracing

Also add "all" as a keyword for setting all values, and pretty up
the output a bit by justifying and aligning the output.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoCPU burn engine fix
Jens Axboe [Fri, 1 Feb 2008 17:33:23 +0000 (18:33 +0100)]
CPU burn engine fix

It's ok to have zero buflen, if we are not doing real IO.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoA bunch of fixes
Jens Axboe [Fri, 1 Feb 2008 12:16:38 +0000 (13:16 +0100)]
A bunch of fixes

Really should have been split up, but...

- Check this_io_bytes at the bottom of do_io() so that async engines
  have a chance to queue pending IO before deeming this job done.

- dprint() should use log_info(), may not be stdout we want.

- last block bug in get_next_free_block(), if max blocks wasn't a
  multiple of the bitmap size.

- Add more debug points.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoTypo in 'init' debug option mask
Jens Axboe [Fri, 1 Feb 2008 12:12:28 +0000 (13:12 +0100)]
Typo in 'init' debug option mask

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd --debug for enabling internal dumps on various actions
Jens Axboe [Fri, 1 Feb 2008 09:30:20 +0000 (10:30 +0100)]
Add --debug for enabling internal dumps on various actions

A little weak currently, when it's fully integrated everywhere in
fio it'll help find fio problems that I can't trigger.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoStop job if it fails to do any IO for a new invocation
Jens Axboe [Fri, 1 Feb 2008 08:29:35 +0000 (09:29 +0100)]
Stop job if it fails to do any IO for a new invocation

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFio 1.17.3 fio-1.17.3
Jens Axboe [Thu, 31 Jan 2008 13:39:16 +0000 (14:39 +0100)]
Fio 1.17.3

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoMerge branch 'master' of ssh://router/data/git/fio
Jens Axboe [Thu, 31 Jan 2008 12:28:40 +0000 (13:28 +0100)]
Merge branch 'master' of ssh://router/data/git/fio

12 years agoFix wrong accouning of zone bytes
Jens Axboe [Thu, 31 Jan 2008 12:25:42 +0000 (13:25 +0100)]
Fix wrong accouning of zone bytes

Ryan Thomas <Ryan.Thomas@nuance.com> writes:

**********

With the following job description

[global]
bs=1k
direct=1
rw=read
ioengine=libaio
iodepth=2
zonesize=1k
zoneskip=1023k
write_bw_log

[/dev/cciss/c0d1]
write_iolog=foo2

The idea here is that I wanted to line up my zones to start at 1M
boundaries across the disk by writing 1k and skipping the next 1023k.
In practice I don't get the alignment because of an extra initial I/O.
I get an iolog that looks like

fio version 2 iolog
/dev/cciss/c0d1 add
/dev/cciss/c0d1 open
/dev/cciss/c0d1 read 0 1024
/dev/cciss/c0d1 read 1024 1024
/dev/cciss/c0d1 read 1049600 1024
/dev/cciss/c0d1 read 2098176 1024

There's a read that I don't expect in that log, namely the read starting
at byte 1024.  Because that read is there, the disk zones get offset by
one block.  I expected output like

fio version 2 iolog
/dev/cciss/c0d1 add
/dev/cciss/c0d1 open
/dev/cciss/c0d1 read 0 1024
/dev/cciss/c0d1 read 1048576 1024
/dev/cciss/c0d1 read 2097152 1024
/dev/cciss/c0d1 read 3145728 1024

***********

The problem is due to the fact that fio account zone_bytes when the IO
completes, which is clearly not correct for io depth > 1. So move the
accounting to when we fill the io_u instead.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoRandom IO fixes
Jens Axboe [Fri, 25 Jan 2008 13:02:15 +0000 (14:02 +0100)]
Random IO fixes

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoMore file extend fixes
Jens Axboe [Thu, 24 Jan 2008 12:13:12 +0000 (13:13 +0100)]
More file extend fixes

Simplify the logic a bit, hope this works...

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agofile extent fix
Gurudas Pai [Thu, 24 Jan 2008 12:04:44 +0000 (13:04 +0100)]
file extent fix

I ran a test with latest fio using jobfile,

[global]
size=10m
[job1]
ioengine=sync
rw=write
filename=testfile

And second time,
[global]
size=30m --> size increased.
[job1]
ioengine=sync
rw=write
filename=testfile

fio did not extend the file to 30m.

In case of f->file_offset=0, f->io_size equals to f->real_file_size, hence no extend.
Following patch worked for me. But this patch will not handle the case where we have
to extend a file and with offset :(

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoUpdate iodepth_batch documentation
Jens Axboe [Fri, 18 Jan 2008 09:30:07 +0000 (10:30 +0100)]
Update iodepth_batch documentation

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoMake iodepth_batch=1 by default
Jens Axboe [Fri, 18 Jan 2008 09:28:11 +0000 (10:28 +0100)]
Make iodepth_batch=1 by default

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agolibio: fix bug in commit hook
Jens Axboe [Fri, 18 Jan 2008 09:26:58 +0000 (10:26 +0100)]
libio: fix bug in commit hook

Also simplifies it.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosurface-scan: make bs 64k again
Jens Axboe [Wed, 16 Jan 2008 10:10:58 +0000 (11:10 +0100)]
surface-scan: make bs 64k again

I reduced it for testing, put it back at 64k to speed it up.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd sample file for surface scan
Shawn Lewis [Wed, 16 Jan 2008 10:08:03 +0000 (11:08 +0100)]
Add sample file for surface scan

Does a write phase and then a verify phase and verifies the written
data.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoRevert "Remove verify_pattern option, replace with verify=pattern:x"
Shawn Lewis [Wed, 16 Jan 2008 10:01:33 +0000 (11:01 +0100)]
Revert "Remove verify_pattern option, replace with verify=pattern:x"

This reverts commit bfb41d98f63cb9fb72cb7c82618be10bc6d52c33.

Reverting lets us have verify=meta and specify a verify_pattern at the same
time. Sample surfacescan conf file coming up.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoDon't truncate files if we don't need to extend them
Shawn Lewis [Fri, 11 Jan 2008 08:45:11 +0000 (09:45 +0100)]
Don't truncate files if we don't need to extend them

Don't truncate files if we don't need to extend them.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd fill_device option
Shawn Lewis [Fri, 11 Jan 2008 08:45:11 +0000 (09:45 +0100)]
Add fill_device option

Add fill_device option. Causes fio to write until ENOSPC occurs (assuming
rw=write).

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agofio: Fix for new file creation
Gurudas Pai [Fri, 21 Dec 2007 09:50:33 +0000 (10:50 +0100)]
fio: Fix for new file creation

fio fails to create new file and just comes out without doing anything..
This happens only when you are creating file freshly.

./fio ../jobfile/jobfile.test
job1: (g=0): rw=randrw, bs=8K-8K/8K-8K, ioengine=sync, iodepth=1
Starting 1 process

Run status group 0 (all jobs):

Disk stats (read/write):
  sda: ios=0/0, merge=0/0, ticks=0/0, in_queue=0, util=0.00%

jobfile:

[global]
 bs=8k
 iodepth_batch=1
 randrepeat=1
 size=100m
 [job1]
 ioengine=sync
 rw=randrw
 filename=newfile

Manually bisected it to commit 509eab12448823a8eefbe925804b5308ee63bf5e
Here is the patch which fixed the issue for me. Please review and apply.
In case of new file "f->io_size > f->real_file_size" becomes true, since
f->real_file_size is 0.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoShow fio help page when no arguments are given
Jens Axboe [Wed, 19 Dec 2007 12:54:38 +0000 (13:54 +0100)]
Show fio help page when no arguments are given

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoOverflow bugs in random map
Jens Axboe [Fri, 14 Dec 2007 11:42:53 +0000 (12:42 +0100)]
Overflow bugs in random map

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd Intel IOMeter file access server example job file
Jens Axboe [Fri, 14 Dec 2007 11:41:04 +0000 (12:41 +0100)]
Add Intel IOMeter file access server example job file

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd support for 'bssplit' option, fine grained block size contrl
Jens Axboe [Fri, 14 Dec 2007 11:21:19 +0000 (12:21 +0100)]
Add support for 'bssplit' option, fine grained block size contrl

From the HOWTO addition:

bssplit=str Sometimes you want even finer grained control of the
block sizes issued, not just an even split between them.
This option allows you to weight various block sizes,
so that you are able to define a specific amount of
block sizes issued. The format for this option is:

bssplit=blocksize/percentage:blocksize/percentage

for as many block sizes as needed. So if you want to define
a workload that has 50% 64k blocks, 10% 4k blocks, and
40% 32k blocks, you would write:

bssplit=4k/10:64k/50:32k/40

Ordering does not matter. If the percentage is left blank,
fio will fill in the remaining values evenly. So a bssplit
option like this one:

bssplit=4k/50:1k/:32k/

would have 50% 4k ios, and 25% 1k and 32k ios. The percentages
always add up to 100, if bssplit is given a range that adds
up to more, it will error out.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years ago'1' is minimum value for queue depth, don't allow 0
Jens Axboe [Wed, 12 Dec 2007 18:42:13 +0000 (19:42 +0100)]
'1' is minimum value for queue depth, don't allow 0

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosyslet: we have fls(), use it for rounding ring size up
Jens Axboe [Wed, 12 Dec 2007 18:39:59 +0000 (19:39 +0100)]
syslet: we have fls(), use it for rounding ring size up

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agounsigned vs signed warnings
Jens Axboe [Tue, 11 Dec 2007 12:23:27 +0000 (13:23 +0100)]
unsigned vs signed warnings

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosyslet: can't free the stack
Jens Axboe [Tue, 11 Dec 2007 11:36:41 +0000 (12:36 +0100)]
syslet: can't free the stack

Need to check up on this...

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoChange io engine version
Jens Axboe [Tue, 11 Dec 2007 09:54:26 +0000 (10:54 +0100)]
Change io engine version

Minor change from V7, just the getevents() prototype.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years ago->getevents() should take unsigned args
Jens Axboe [Tue, 11 Dec 2007 09:49:39 +0000 (10:49 +0100)]
->getevents() should take unsigned args

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosyslet: cleanup event reaping
Jens Axboe [Tue, 11 Dec 2007 09:45:36 +0000 (10:45 +0100)]
syslet: cleanup event reaping

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosyslet: add proper read barrier between user_tail and completion read
Jens Axboe [Tue, 11 Dec 2007 07:55:53 +0000 (08:55 +0100)]
syslet: add proper read barrier between user_tail and completion read

Also fixup a bug with ring indexing, it needs to use the real ring size
mask, not the io depth.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosyslet: error handling
Jens Axboe [Tue, 11 Dec 2007 07:46:16 +0000 (08:46 +0100)]
syslet: error handling

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosyslet: make sure ring buffer is a power-of-2
Jens Axboe [Tue, 11 Dec 2007 07:22:53 +0000 (08:22 +0100)]
syslet: make sure ring buffer is a power-of-2

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosyslet: memory allocation error handling
Jens Axboe [Mon, 10 Dec 2007 19:35:10 +0000 (20:35 +0100)]
syslet: memory allocation error handling

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd kcompat.h
Jens Axboe [Mon, 10 Dec 2007 19:25:24 +0000 (20:25 +0100)]
Add kcompat.h

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosyslet: update to -v7 as posted by Zach
Jens Axboe [Mon, 10 Dec 2007 19:24:44 +0000 (20:24 +0100)]
syslet: update to -v7 as posted by Zach

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFix failure in opening files with O_NOATIME
Aaron Carroll [Mon, 26 Nov 2007 08:08:53 +0000 (09:08 +0100)]
Fix failure in opening files with O_NOATIME

If we try to open a file with O_NOATIME and it fails, turn off
O_NOATIME and try again.  This occurs if the user does not own
the file and is unprivileged.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoUpdate HOWTO terse description
Shawn Lewis [Wed, 21 Nov 2007 08:38:34 +0000 (09:38 +0100)]
Update HOWTO terse description

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agofix utime_since overflow
Shawn Lewis [Wed, 21 Nov 2007 08:38:13 +0000 (09:38 +0100)]
fix utime_since overflow

utime_since was using a long for its return which overflows when the difference
between times is more than about 35 minutes.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agodon't log iopiece when do_verify is off
Shawn Lewis [Wed, 21 Nov 2007 08:35:41 +0000 (09:35 +0100)]
don't log iopiece when do_verify is off

Don't log an iopiece if do_verify is off. This allows us to do large write
phases (with verify headers) without using a lot of memory.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agofio.1: minor fixes
Aaron Carroll [Fri, 16 Nov 2007 11:12:45 +0000 (12:12 +0100)]
fio.1: minor fixes

 - Document reading job files from stdin with `-';
 - Fix a typo and remove a spurious line.

Signed-off-by: Aaron Carroll <aaronc@gelato.unsw.edu.au>
Signed-off-by: Jens Axboe <axboe@carl.home.kernel.dk>
12 years agoSeperate ioprio value from setting
Jens Axboe [Thu, 8 Nov 2007 07:29:07 +0000 (08:29 +0100)]
Seperate ioprio value from setting

Newer kernels allow passing a value of '0' for resetting ioprio,
and our check for if (td->ioprio) doesn't allow passing a value of
0 into sys_ioprio_set(). So seperate the setting from the value by
adding a ->ioprio_set bool.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFio 1.17.2 fio-1.17.2
Jens Axboe [Tue, 30 Oct 2007 14:45:12 +0000 (15:45 +0100)]
Fio 1.17.2

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoFix spurious eta output
Aaron Carroll [Fri, 26 Oct 2007 09:10:28 +0000 (11:10 +0200)]
Fix spurious eta output

 - Pad eta output with enough spaces to overwrite remaining characters
   from the end of previous lines;
 - Wrap an overly-long line.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoMore cpu engine fixes
Jens Axboe [Thu, 25 Oct 2007 16:44:45 +0000 (18:44 +0200)]
More cpu engine fixes

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoMerge branch 'master' of ssh://git.kernel.dk/data/git/fio
Jens Axboe [Thu, 25 Oct 2007 16:34:14 +0000 (18:34 +0200)]
Merge branch 'master' of ssh://git.kernel.dk/data/git/fio

12 years agoFixes for CPU burn engine
Jens Axboe [Thu, 25 Oct 2007 16:34:02 +0000 (18:34 +0200)]
Fixes for CPU burn engine

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agoAdd pread/pwrite support to sync engine
gurudas pai [Tue, 23 Oct 2007 13:12:30 +0000 (15:12 +0200)]
Add pread/pwrite support to sync engine

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agofio: build warning fix on gcc-3.4.6-3
gurudas pai [Mon, 22 Oct 2007 20:10:39 +0000 (22:10 +0200)]
fio: build warning fix on  gcc-3.4.6-3

On gcc-3.4.6-3

 >make log.o
gcc -W -Wwrite-strings -Wall -D_GNU_SOURCE -D_LARGEFILE_SOURCE
-D_FILE_OFFSET_BITS=64 -O2 -g  -D_FORTIFY_SOURCE=2 -rdynamic  -c -o
log.o log.c
log.c: In function `read_iolog':
log.c:308: warning: int format, fio_ddir arg (arg 3)

Which corresponds to

enum fio_ddir rw;
..
..
if (sscanf(p, "%d,%llu,%u", &rw, &offset, &bytes) != 3) {

Following patch will fix the warning. This was added by commit
21bd2987dc5d82a18af485cd1e7841e94137fc0c

Thanks,
-Guru

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosplice: fix fallback from copy vmsplice to nothing
Jens Axboe [Tue, 16 Oct 2007 07:44:47 +0000 (09:44 +0200)]
splice: fix fallback from copy vmsplice to nothing

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
12 years agosplice: fix problem with current mainline kernels
Jens Axboe [Tue, 16 Oct 2007 07:42:40 +0000 (09:42 +0200)]
splice: fix problem with current mainline kernels

splice was updated to work well with the newer vmsplice patches,
where we don't have to copy data at all. But that broke the
regular vmsplice that copies data, which is all the mainline kernel
still support.

Fix that up, so we do appropriate fallbacks. Fixes the problem
reported by gurudas pai <gurudas.pai@oracle.com>

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>