fio.git
9 years agoFio 1.35 fio-1.35
Jens Axboe [Thu, 5 Nov 2009 07:58:55 +0000 (08:58 +0100)]
Fio 1.35

A few nasty bugs fixed since the last tag, so lets
cut a new one.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoDon't pass more than the queued IO depth in to getevents()
Radha Ramachandran [Thu, 5 Nov 2009 07:33:49 +0000 (08:33 +0100)]
Don't pass more than the queued IO depth in to getevents()

Otherwise we can hang fio, since it'll attempt to wait for more
IOs than we have in flight.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoFix race on io_u->flags
Jens Axboe [Wed, 4 Nov 2009 19:58:09 +0000 (20:58 +0100)]
Fix race on io_u->flags

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoMove BITS_PER_LONG define to arch
Jens Axboe [Wed, 4 Nov 2009 12:56:13 +0000 (13:56 +0100)]
Move BITS_PER_LONG define to arch

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoWait for async threads before freeing/killing IO buffers
Jens Axboe [Wed, 4 Nov 2009 11:51:17 +0000 (12:51 +0100)]
Wait for async threads before freeing/killing IO buffers

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoFixup iolog replay and add 'wait' command
Glen Ogilvie [Tue, 3 Nov 2009 20:52:36 +0000 (21:52 +0100)]
Fixup iolog replay and add 'wait' command

I have been working with Mike O'Sullivan, a senior lecturer at the University
of Auckland, New Zealand. He has been adapting fio 1.34.1 to work with iolog
files and also to be able to read a "wait" command to allow the iolog file to
specify a wait (in millisecs) before the next io. Mike's not completely
confident that his changes are correct in the overall fio picture, but they are
working with the iolog files we are using. What we did:

1) We incorporated wait into iolog file reading;
2) We made some modifications so that read_iolog worked.

We have attached a patch for fio-1.34.1.tar.bz2 along with some demonstration
fio config files and iolog files.

We are doing this so we can use an open source SPC-1 emulator to
generate SPC-1 like workloads and then run them in fio.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoFix race condition when using asynch verify
Radha Ramachandran [Tue, 3 Nov 2009 20:45:31 +0000 (21:45 +0100)]
Fix race condition when using asynch verify

I discovered the race condition when using asynch verify with libaio engine.
The code assumes that because the td->cur_depth value is not 0 that
there is still I/O pending and issues io_getevents when the I/O was
actually being verified by the asynchronous verify thread. This causes
the code to hang.

(Updated by Jens to use a new io_u->flag bitfield instead of adding a new
 integer to struct io_u).

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoSupport > 4 byte verify write patterns
Radha Ramachandran [Tue, 27 Oct 2009 19:14:27 +0000 (20:14 +0100)]
Support > 4 byte verify write patterns

This is the patch I have to write/read/verify hex patterns of size > 4
bytes.  I have tested these on hard drives and did not see any
performance change. I am assuming the performance drop if any might show
up in the read phase for probably solid state drives and maybe it can be
worked around using the asynchronous verify flag.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoFio 1.34.2 fio-1.34.2
Jens Axboe [Thu, 15 Oct 2009 05:11:30 +0000 (07:11 +0200)]
Fio 1.34.2

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoFix bug in bssplit parsing
Kenneth Waters [Thu, 15 Oct 2009 05:08:48 +0000 (07:08 +0200)]
Fix bug in bssplit parsing

I found a bug is bssplit parsing in bssplit_ddir() when trying to use
FIO on an ARM platform.  It looks like str_to_decimal() takes a "void
*" as its fourth argument, which it passes to get_multi_bytes(), which
is passes to fio_get_kb_base(), which casts it to a "struct
thread_data *".  bssplit_ddir() is passing in "struct thread_data **",
which I think is causing fio_get_kb_base() to go paw around in the
stack; I believe the different stack layouts between ARM and x86 is
causing me to see this problem.  The first change that introduces this
bug seems to be d6978a3242daad9cb7b0710b724f19225d1ed7e2.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoMerge branch 'master' of ssh://brick.kernel.dk/data/git/fio
Jens Axboe [Wed, 7 Oct 2009 20:02:15 +0000 (22:02 +0200)]
Merge branch 'master' of ssh://brick.kernel.dk/data/git/fio

9 years agoFix problem with mutexes and MONOTONIC clock
Roland Dreier [Wed, 7 Oct 2009 20:01:42 +0000 (22:01 +0200)]
Fix problem with mutexes and MONOTONIC clock

With the latest fio git tree, I get:

    fio: job startup hung? exiting.

on startup, and it looks like the latest commit, 69a852f5 ("Fix wrong
clock source in mutex") breaks things for me (Ubuntu 9.10 beta).  My
system seems to use CLOCK_REALTIME for pthread_cond_timedwait() by
default.

The cleanest solution seems to be to explicitly set which clock to use
in pthread_cond_timedwait() via pthread_condattr_setclock() when
initializing the condition variable, although I'm not sure how portable
this is away from modern Linux (ie do all platforms that fio cares about
have pthread_condattr_setclock()?).

Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoFix wrong clock source in mutex
Jens Axboe [Tue, 6 Oct 2009 06:50:01 +0000 (08:50 +0200)]
Fix wrong clock source in mutex

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoFio 1.34.1 fio-1.34.1
Jens Axboe [Thu, 1 Oct 2009 17:45:31 +0000 (19:45 +0200)]
Fio 1.34.1

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoReally fix dm/md slave disk utilization
Jens Axboe [Thu, 1 Oct 2009 17:44:36 +0000 (19:44 +0200)]
Really fix dm/md slave disk utilization

We should just handle it internally in the disk util inc/dec. Make
it iterate each slave and modify their usage count, too.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoRevert "Fix disk utilization for md/dm slaves"
Jens Axboe [Thu, 1 Oct 2009 17:39:26 +0000 (19:39 +0200)]
Revert "Fix disk utilization for md/dm slaves"

This reverts commit f03f8e473b7c8a5c4ed2de0b99717cf75521bf69.

9 years agoMerge branch 'master' of ssh://brick.kernel.dk/data/git/fio
Jens Axboe [Thu, 1 Oct 2009 17:38:41 +0000 (19:38 +0200)]
Merge branch 'master' of ssh://brick.kernel.dk/data/git/fio

9 years agoFix disk utilization for md/dm slaves
Jens Axboe [Thu, 1 Oct 2009 09:20:03 +0000 (11:20 +0200)]
Fix disk utilization for md/dm slaves

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
9 years agoAdd generic bdev_size function
Jens Axboe [Tue, 22 Sep 2009 15:12:29 +0000 (17:12 +0200)]
Add generic bdev_size function

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFix compile error on IA64
Jens Axboe [Tue, 15 Sep 2009 17:54:43 +0000 (19:54 +0200)]
Fix compile error on IA64

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd a bit of huge page documentation
Jens Axboe [Mon, 14 Sep 2009 06:56:53 +0000 (08:56 +0200)]
Add a bit of huge page documentation

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd hint on being root for EPERM and shmhuge error
Jens Axboe [Mon, 14 Sep 2009 06:51:55 +0000 (08:51 +0200)]
Add hint on being root for EPERM and shmhuge error

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFio 1.34 fio-1.34
Jens Axboe [Sat, 12 Sep 2009 19:34:19 +0000 (21:34 +0200)]
Fio 1.34

Two critical fixes since 1.33 warrant a new release. This now works
properly with shm huge pages again, and a deadlock with lvm
component resolving is fixed.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMake sure shmhuge allocations are huge page size aligned
Jens Axboe [Sat, 12 Sep 2009 19:33:51 +0000 (21:33 +0200)]
Make sure shmhuge allocations are huge page size aligned

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoSilence the style/whitespace police
Jens Axboe [Fri, 11 Sep 2009 20:29:54 +0000 (22:29 +0200)]
Silence the style/whitespace police

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoCleanup diskutil.c
Jens Axboe [Fri, 11 Sep 2009 20:22:41 +0000 (22:22 +0200)]
Cleanup diskutil.c

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMerge branch 'master' of ssh://brick.kernel.dk/data/git/fio
Jens Axboe [Fri, 11 Sep 2009 19:34:49 +0000 (21:34 +0200)]
Merge branch 'master' of ssh://brick.kernel.dk/data/git/fio

10 years agoBug fix: handles disk device used multiple times
Alan D. Brunelle [Fri, 11 Sep 2009 18:57:10 +0000 (14:57 -0400)]
Bug fix: handles disk device used multiple times

There were issues in having the same underlying device being referenced
multiple times (via different paths) when reporting storage I/O
statistics. As an example: having two (or more) LVM2/DM volumes crafted
out of the same MD array.

This patch simply skips over any devices previously seen.

Signed-off-by: Alan D. Brunelle <alan.brunelle@hp.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFio 1.33.1 fio-1.33.1
Jens Axboe [Thu, 10 Sep 2009 07:54:55 +0000 (09:54 +0200)]
Fio 1.33.1

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFix typo in write_barrier() on Alpha
Martin Steigerwald [Thu, 10 Sep 2009 07:44:29 +0000 (09:44 +0200)]
Fix typo in write_barrier() on Alpha

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFio 1.33 fio-1.33
Jens Axboe [Sat, 5 Sep 2009 18:44:59 +0000 (20:44 +0200)]
Fio 1.33

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoDocument the verify behaviour wrt read and write workloads
Jens Axboe [Sat, 5 Sep 2009 18:37:35 +0000 (20:37 +0200)]
Document the verify behaviour wrt read and write workloads

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoDon't assume read IO are always verifies if verify is enabled
Jens Axboe [Fri, 4 Sep 2009 18:53:53 +0000 (20:53 +0200)]
Don't assume read IO are always verifies if verify is enabled

If it's a mixed read/write workload, we're likely just wanting to
verify the writes we do.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMIPS support
Jens Axboe [Tue, 1 Sep 2009 05:23:57 +0000 (07:23 +0200)]
MIPS support

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoDon't start new groups for numjobs=
Jens Axboe [Tue, 25 Aug 2009 20:14:29 +0000 (22:14 +0200)]
Don't start new groups for numjobs=

No point in doing that, and it makes two identical jobs behave
differently if typed manually or implicitly with numjobs=x

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agofio.1: fix hyphen-used-as-minus-sign
Michael Prokop [Wed, 12 Aug 2009 15:24:41 +0000 (17:24 +0200)]
fio.1: fix hyphen-used-as-minus-sign

Signed-off-by: Michael Prokop <mika@grml.org>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMerge branch 'master' of ssh://brick.kernel.dk/data/git/fio
Jens Axboe [Sun, 9 Aug 2009 20:40:50 +0000 (22:40 +0200)]
Merge branch 'master' of ssh://brick.kernel.dk/data/git/fio

10 years agoAdd Linus' optimized sha1 hash for checksum
Jens Axboe [Sun, 9 Aug 2009 20:40:33 +0000 (22:40 +0200)]
Add Linus' optimized sha1 hash for checksum

It does appear to be pretty fast.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd support for reserved keywords
Jens Axboe [Wed, 5 Aug 2009 09:42:37 +0000 (11:42 +0200)]
Add support for reserved keywords

Currently $pagesize, $mb_memory, and $ncpus are supported.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd strong madvise() hint for cache pruning
Jens Axboe [Tue, 4 Aug 2009 21:17:02 +0000 (23:17 +0200)]
Add strong madvise() hint for cache pruning

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoRemove old FIXME comment
Jens Axboe [Tue, 4 Aug 2009 21:04:18 +0000 (23:04 +0200)]
Remove old FIXME comment

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFreeBSD: use generic bdev size retrieval
Jens Axboe [Tue, 4 Aug 2009 21:03:39 +0000 (23:03 +0200)]
FreeBSD: use generic bdev size retrieval

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agosolaris: implement blockdev size getting
Jens Axboe [Tue, 4 Aug 2009 10:49:59 +0000 (12:49 +0200)]
solaris: implement blockdev size getting

Just use lseek(..., SEEK_END), we should actually be able to use that
on all platforms.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agosolarisaio: fix aio errno assignment
Jens Axboe [Tue, 4 Aug 2009 10:49:36 +0000 (12:49 +0200)]
solarisaio: fix aio errno assignment

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAllow ':' in filenames
Jens Axboe [Tue, 4 Aug 2009 07:51:48 +0000 (09:51 +0200)]
Allow ':' in filenames

You can now use '\' to escape a colon in a filename.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoSolaris cpu affinity fix
Jens Axboe [Mon, 3 Aug 2009 08:54:41 +0000 (10:54 +0200)]
Solaris cpu affinity fix

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFio 1.32 fio-1.32
Jens Axboe [Fri, 31 Jul 2009 09:23:28 +0000 (11:23 +0200)]
Fio 1.32

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agosmalloc: enlarge pool size
Jens Axboe [Fri, 31 Jul 2009 09:23:19 +0000 (11:23 +0200)]
smalloc: enlarge pool size

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFix problem with bw log not working for writes
Jens Axboe [Fri, 31 Jul 2009 09:20:52 +0000 (11:20 +0200)]
Fix problem with bw log not working for writes

We inadvertently only updated the read stat sample time cache,
so writes never logged any activity since negativ time spent is
rounded to zero.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMerge branch 'master' of ssh://router/data/git/fio
Jens Axboe [Mon, 27 Jul 2009 10:28:21 +0000 (12:28 +0200)]
Merge branch 'master' of ssh://router/data/git/fio

10 years agoVerify pattern & header meta data
Radha Ramachandran [Mon, 27 Jul 2009 10:27:55 +0000 (12:27 +0200)]
Verify pattern & header meta data

Also verify meta data when doing pattern verify, if available.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoA few more kb_base fixups
Jens Axboe [Sat, 18 Jul 2009 19:04:09 +0000 (21:04 +0200)]
A few more kb_base fixups

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMake kb_base apply to parser k/m/g etc values
Jens Axboe [Sat, 18 Jul 2009 06:25:35 +0000 (08:25 +0200)]
Make kb_base apply to parser k/m/g etc values

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd tera/peta suffixes
Jens Axboe [Fri, 17 Jul 2009 21:16:17 +0000 (23:16 +0200)]
Add tera/peta suffixes

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMerge branch 'master' of ssh://brick.kernel.dk/data/git/fio
Jens Axboe [Fri, 17 Jul 2009 20:33:45 +0000 (22:33 +0200)]
Merge branch 'master' of ssh://brick.kernel.dk/data/git/fio

10 years agoAdd kb_base option to specify the base unit of a kilobyte
Jens Axboe [Fri, 17 Jul 2009 20:33:32 +0000 (22:33 +0200)]
Add kb_base option to specify the base unit of a kilobyte

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoGet rid of KiB vs KB distinction
Jens Axboe [Fri, 17 Jul 2009 20:29:23 +0000 (22:29 +0200)]
Get rid of KiB vs KB distinction

Confuses more than it does good, drop it and default to just using KB,
MB, etc.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoparser: dump valid option format for ->verify() fail
Jens Axboe [Fri, 17 Jul 2009 20:26:15 +0000 (22:26 +0200)]
parser: dump valid option format for ->verify() fail

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFio 1.31 fio-1.31
Jens Axboe [Fri, 10 Jul 2009 06:48:17 +0000 (08:48 +0200)]
Fio 1.31

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFix broken completion latency numbers
Jens Axboe [Fri, 10 Jul 2009 06:47:57 +0000 (08:47 +0200)]
Fix broken completion latency numbers

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agosmalloc: just round alloc_size to int aligned for post redzone pointer
Jens Axboe [Tue, 7 Jul 2009 21:03:06 +0000 (23:03 +0200)]
smalloc: just round alloc_size to int aligned for post redzone pointer

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMake smalloc redzone pointer properly aligned
Jens Axboe [Tue, 7 Jul 2009 15:43:11 +0000 (17:43 +0200)]
Make smalloc redzone pointer properly aligned

It would SIGBUS on archs that require proper alignment, and work
suboptimally on others. Real fix is pending, making sure that
we always return naturally aligned pointers. But this at least
makes it work on sparc64.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoPrint error on thread init failure
Jens Axboe [Mon, 6 Jul 2009 12:28:21 +0000 (14:28 +0200)]
Print error on thread init failure

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAsync verify HOWTO/man update
Jens Axboe [Mon, 6 Jul 2009 12:12:57 +0000 (14:12 +0200)]
Async verify HOWTO/man update

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoCode indentation fix
Jens Axboe [Mon, 6 Jul 2009 12:12:52 +0000 (14:12 +0200)]
Code indentation fix

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMake async IO verify offload handle errors properly
Jens Axboe [Mon, 6 Jul 2009 11:51:05 +0000 (13:51 +0200)]
Make async IO verify offload handle errors properly

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoImprove async verify offload efficiency
Jens Axboe [Mon, 6 Jul 2009 11:43:46 +0000 (13:43 +0200)]
Improve async verify offload efficiency

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd support for async IO verification offload
Jens Axboe [Mon, 6 Jul 2009 10:59:04 +0000 (12:59 +0200)]
Add support for async IO verification offload

This adds support for setting up a number of IO verification offload
threads, instead of doing the offload inline. An option for controlling
the CPU affinity of those threads are always added.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agommap engine: remove code duplication
Jens Axboe [Sat, 4 Jul 2009 18:29:04 +0000 (20:29 +0200)]
mmap engine: remove code duplication

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agommap engine: allow full mmaps for 64-bit machines
Jens Axboe [Fri, 3 Jul 2009 20:52:38 +0000 (22:52 +0200)]
mmap engine: allow full mmaps for 64-bit machines

No need to do partial mmaps there.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFio 1.30 fio-1.30
Jens Axboe [Fri, 3 Jul 2009 06:40:55 +0000 (08:40 +0200)]
Fio 1.30

O_DIRECT accidentically got broken by the iomem_align option, so
lets just cut a new major release.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd warning about potentially missing O_DIRECT support or bad alignment
Jens Axboe [Fri, 3 Jul 2009 06:38:50 +0000 (08:38 +0200)]
Add warning about potentially missing O_DIRECT support or bad alignment

If the first O_DIRECT fails, then it's likely because:

1) The file system does not support O_DIRECT, or
2) The user set iomem_align to an unsupported value.

Let the user know, otherwise they only see EINVAL and may not realize
why things aren't working.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoDon't limit direct IO to 512b aligned
Jens Axboe [Fri, 3 Jul 2009 06:33:08 +0000 (08:33 +0200)]
Don't limit direct IO to 512b aligned

This is a platform restriction at best, eg on Linux block devices must
have 512b memory buffer alignment (or hw sector size). With NFS, this
is not required. And other operating systems may have other limitations
as well, so don't hard core anything.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFix O_DIRECT memory alignment
Jens Axboe [Fri, 3 Jul 2009 06:30:15 +0000 (08:30 +0200)]
Fix O_DIRECT memory alignment

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFio 1.29 fio-1.29
Jens Axboe [Thu, 2 Jul 2009 08:20:26 +0000 (10:20 +0200)]
Fio 1.29

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMerge branch 'master' of ssh://router/data/git/fio
Jens Axboe [Thu, 2 Jul 2009 06:31:30 +0000 (08:31 +0200)]
Merge branch 'master' of ssh://router/data/git/fio

10 years agoFix issue with fill_device=1 and continue_on_error
Radha Ramachandran [Thu, 2 Jul 2009 06:30:49 +0000 (08:30 +0200)]
Fix issue with fill_device=1 and continue_on_error

I added a change so when both the continue_on_error and fill_device
option is on, when it hits ENOSPC issue the code ignores the error.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd code to detect a task that exited prior to up'ing the startup mutex
Jens Axboe [Wed, 1 Jul 2009 21:02:10 +0000 (23:02 +0200)]
Add code to detect a task that exited prior to up'ing the startup mutex

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoDon't allow pre_read on IO engines that cannot seek
Jens Axboe [Wed, 1 Jul 2009 10:26:28 +0000 (12:26 +0200)]
Don't allow pre_read on IO engines that cannot seek

We cannot pre-read files, if we cannot seek back and read the same
data again. So just disable it with a warning to the user.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFio 1.29-rc1 fio-1.29-rc1
Jens Axboe [Wed, 1 Jul 2009 08:45:31 +0000 (10:45 +0200)]
Fio 1.29-rc1

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFix compile error
Jens Axboe [Wed, 1 Jul 2009 08:43:05 +0000 (10:43 +0200)]
Fix compile error

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMove gtod_cpu affinity check to verify callback
Jens Axboe [Wed, 1 Jul 2009 08:42:18 +0000 (10:42 +0200)]
Move gtod_cpu affinity check to verify callback

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMove rw= write && read_only check to ->verify callback
Jens Axboe [Wed, 1 Jul 2009 08:38:35 +0000 (10:38 +0200)]
Move rw= write && read_only check to ->verify callback

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd support for the iomem_align option
Jens Axboe [Wed, 1 Jul 2009 08:33:03 +0000 (10:33 +0200)]
Add support for the iomem_align option

This allows detailed control of the alignment of the IO buffers
that fio uses.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd verify hook to options
Jens Axboe [Wed, 1 Jul 2009 07:24:05 +0000 (09:24 +0200)]
Add verify hook to options

This allows us to individually verify options, instead of having
a big function that attempts to fix things up.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoOnly show maxval in --cmdhelp if it is given
Jens Axboe [Wed, 1 Jul 2009 07:06:12 +0000 (09:06 +0200)]
Only show maxval in --cmdhelp if it is given

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoCleanup IO buffer memory allocation
Jens Axboe [Wed, 1 Jul 2009 07:00:52 +0000 (09:00 +0200)]
Cleanup IO buffer memory allocation

This is in preparation for adding an option that controls the
alignment of the buffers.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agommap engine: fix for large files on 32-bit archs
Jens Axboe [Mon, 29 Jun 2009 09:54:56 +0000 (11:54 +0200)]
mmap engine: fix for large files on 32-bit archs

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agosg engine: IO should be done sync of O_SYNC and O_DIRECT
Jens Axboe [Tue, 16 Jun 2009 21:23:01 +0000 (23:23 +0200)]
sg engine: IO should be done sync of O_SYNC and O_DIRECT

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd support for fdatasync()
Jens Axboe [Tue, 16 Jun 2009 20:40:26 +0000 (22:40 +0200)]
Add support for fdatasync()

Adds a new option, fdatasync=. It's identical to the fsync= option,
but uses fdatasync() instead.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoMerge branch 'master' of ssh://router/data/git/fio
Jens Axboe [Tue, 16 Jun 2009 06:31:58 +0000 (08:31 +0200)]
Merge branch 'master' of ssh://router/data/git/fio

10 years agoAlter the order of checking time exceeded vs getting an io_u
Radha Ramachandran [Tue, 16 Jun 2009 06:31:29 +0000 (08:31 +0200)]
Alter the order of checking time exceeded vs getting an io_u

I was hitting a case where fio would loop for ever even though the
runtime was exceeded if it had difficulty getting the io_u struct, so
I suggest this change to the code so it wld not loop forever.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoDefault print pretty up
Jens Axboe [Mon, 15 Jun 2009 21:29:54 +0000 (23:29 +0200)]
Default print pretty up

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoCmd help type fix
Jens Axboe [Mon, 15 Jun 2009 12:45:43 +0000 (14:45 +0200)]
Cmd help type fix

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFix typo on continue_on_error patch
Jens Axboe [Mon, 15 Jun 2009 10:37:30 +0000 (12:37 +0200)]
Fix typo on continue_on_error patch

Also expand the error print a bit.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoAdd a 'continue_on_error' option to fio
Radha Ramachandran [Mon, 15 Jun 2009 06:40:16 +0000 (08:40 +0200)]
Add a 'continue_on_error' option to fio

Add option to make fio continue on non-fatal errors.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoFio 1.28 fio-1.28
Jens Axboe [Thu, 11 Jun 2009 20:06:40 +0000 (22:06 +0200)]
Fio 1.28

Should be good enough for another major release.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoYet another rate fixup
Jens Axboe [Thu, 11 Jun 2009 20:06:23 +0000 (22:06 +0200)]
Yet another rate fixup

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoImprove rate limiting
Jens Axboe [Thu, 11 Jun 2009 08:55:39 +0000 (10:55 +0200)]
Improve rate limiting

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
10 years agoGet rid of (now) unused rate_throttle()
Jens Axboe [Thu, 11 Jun 2009 08:41:28 +0000 (10:41 +0200)]
Get rid of (now) unused rate_throttle()

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