Add the zonemode job option

Fio's zone support makes fio perform I/O inside a zone before it skips
to the next zone. That behavior is the opposite of the behavior needed
for zoned block devices, namely to consider all zones when performing
random I/O. Hence introduce a new job option that allows users to
choose between fio's traditional zone mode and the behavior needed for
zoned block devices. This patch makes fio behave identically with
--zonemode=none and --zonemode=zbd. A later patch will implement new
behavior for --zonemode=zbd.

Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

diff --git a/HOWTO b/HOWTO
index 383946145794961dfb419be326c70f041a0b8625..a0fadcbbd52359c24d65f3641d09c4fa5697b069 100644 (file)
--- a/HOWTO
+++ b/HOWTO
@@ -952,24 +952,50 @@ Target file/device
 
        Unlink job files after each iteration or loop.  Default: false.
 
 
        Unlink job files after each iteration or loop.  Default: false.
 

+.. option:: zonemode=str
+
+       Accepted values are:
+
+               **none**
+                               The :option:`zonerange`, :option:`zonesize` and
+                               :option:`zoneskip` parameters are ignored.
+               **strided**
+                               I/O happens in a single zone until
+                               :option:`zonesize` bytes have been transferred.
+                               After that number of bytes has been
+                               transferred processing of the next zone
+                               starts.
+               **zbd**
+                               Zoned block device mode. I/O happens
+                               sequentially in each zone, even if random I/O
+                               has been selected. Random I/O happens across
+                               all zones instead of being restricted to a
+                               single zone. The :option:`zoneskip` parameter
+                               is ignored. :option:`zonerange` and
+                               :option:`zonesize` must be identical.
+

 .. option:: zonerange=int
 
 .. option:: zonerange=int
 

-       Size of a single zone in which I/O occurs. See also :option:`zonesize`
-       and :option:`zoneskip`.
+       Size of a single zone. See also :option:`zonesize` and
+       :option:`zoneskip`.

 
 .. option:: zonesize=int
 
 
 .. option:: zonesize=int
 

-       Number of bytes to transfer before skipping :option:`zoneskip`
-       bytes. If this parameter is smaller than :option:`zonerange` then only
-       a fraction of each zone with :option:`zonerange` bytes will be
-       accessed.  If this parameter is larger than :option:`zonerange` then
-       each zone will be accessed multiple times before skipping
+       For :option:`zonemode` =strided, this is the number of bytes to
+       transfer before skipping :option:`zoneskip` bytes. If this parameter
+       is smaller than :option:`zonerange` then only a fraction of each zone
+       with :option:`zonerange` bytes will be accessed.  If this parameter is
+       larger than :option:`zonerange` then each zone will be accessed
+       multiple times before skipping to the next zone.
+
+       For :option:`zonemode` =zbd, this is the size of a single zone. The
+       :option:`zonerange` parameter is ignored in this mode.

 
 .. option:: zoneskip=int
 
 
 .. option:: zoneskip=int
 

-       Skip the specified number of bytes when :option:`zonesize` data have
-       been transferred. The three zone options can be used to do strided I/O
-       on a file.
+       For :option:`zonemode` =strided, the number of bytes to skip after
+       :option:`zonesize` bytes of data have been transferred. This parameter
+       must be zero for :option:`zonemode` =zbd.

 
 
 I/O type
 
 
 I/O type

diff --git a/cconv.c b/cconv.c
index 534bfb07214078f4a0d12f84691579edb8259a57..1d7f6f22350327e99ef070acb3a3764ace470711 100644 (file)
--- a/cconv.c
+++ b/cconv.c
@@ -223,6 +223,7 @@ void convert_thread_options_to_cpu(struct thread_options *o,
        o->zone_range = le64_to_cpu(top->zone_range);
        o->zone_size = le64_to_cpu(top->zone_size);
        o->zone_skip = le64_to_cpu(top->zone_skip);
        o->zone_range = le64_to_cpu(top->zone_range);
        o->zone_size = le64_to_cpu(top->zone_size);
        o->zone_skip = le64_to_cpu(top->zone_skip);

+       o->zone_mode = le32_to_cpu(top->zone_mode);

        o->lockmem = le64_to_cpu(top->lockmem);
        o->offset_increment = le64_to_cpu(top->offset_increment);
        o->number_ios = le64_to_cpu(top->number_ios);
        o->lockmem = le64_to_cpu(top->lockmem);
        o->offset_increment = le64_to_cpu(top->offset_increment);
        o->number_ios = le64_to_cpu(top->number_ios);


@@ -548,6 +549,7 @@ void convert_thread_options_to_net(struct thread_options_pack *top,
        top->zone_range = __cpu_to_le64(o->zone_range);
        top->zone_size = __cpu_to_le64(o->zone_size);
        top->zone_skip = __cpu_to_le64(o->zone_skip);
        top->zone_range = __cpu_to_le64(o->zone_range);
        top->zone_size = __cpu_to_le64(o->zone_size);
        top->zone_skip = __cpu_to_le64(o->zone_skip);

+       top->zone_mode = __cpu_to_le32(o->zone_mode);

        top->lockmem = __cpu_to_le64(o->lockmem);
        top->ddir_seq_add = __cpu_to_le64(o->ddir_seq_add);
        top->file_size_low = __cpu_to_le64(o->file_size_low);
        top->lockmem = __cpu_to_le64(o->lockmem);
        top->ddir_seq_add = __cpu_to_le64(o->ddir_seq_add);
        top->file_size_low = __cpu_to_le64(o->file_size_low);

diff --git a/filesetup.c b/filesetup.c
index 94a025e6c1064e442dc18254d08e3f002dcae644..2ab251531ae8491fbcb0515547d4ea2aaf006559 100644 (file)
--- a/filesetup.c
+++ b/filesetup.c
@@ -1142,9 +1142,6 @@ int setup_files(struct thread_data *td)
        if (err)
                goto err_out;
 
        if (err)
                goto err_out;
 

-       if (!o->zone_size)
-               o->zone_size = o->size;
-

        /*
         * iolog already set the total io size, if we read back
         * stored entries.
        /*
         * iolog already set the total io size, if we read back
         * stored entries.

diff --git a/fio.1 b/fio.1
index 4071947f24f303edb8739f2b8b048ea9a6ee84d7..3b961193e9ed0bcc61686fe2f568bcd2e74db094 100644 (file)
--- a/fio.1
+++ b/fio.1
@@ -724,21 +724,43 @@ false.
 .BI unlink_each_loop \fR=\fPbool
 Unlink job files after each iteration or loop. Default: false.
 .TP
 .BI unlink_each_loop \fR=\fPbool
 Unlink job files after each iteration or loop. Default: false.
 .TP

-Fio supports strided data access. After having read \fBzonesize\fR bytes from an area that is \fBzonerange\fR bytes big, \fBzoneskip\fR bytes are skipped.
+.BI zonemode \fR=\fPstr
+Accepted values are:
+.RS
+.RS
+.TP
+.B none
+The \fBzonerange\fR, \fBzonesize\fR and \fBzoneskip\fR parameters are ignored.
+.TP
+.B strided
+I/O happens in a single zone until \fBzonesize\fR bytes have been transferred.
+After that number of bytes has been transferred processing of the next zone
+starts.
+.TP
+.B zbd
+Zoned block device mode. I/O happens sequentially in each zone, even if random
+I/O has been selected. Random I/O happens across all zones instead of being
+restricted to a single zone.
+.RE
+.RE

 .TP
 .BI zonerange \fR=\fPint
 .TP
 .BI zonerange \fR=\fPint

-Size of a single zone in which I/O occurs.
+Size of a single zone. See also \fBzonesize\fR and \fBzoneskip\fR.

 .TP
 .BI zonesize \fR=\fPint
 .TP
 .BI zonesize \fR=\fPint

-Number of bytes to transfer before skipping \fBzoneskip\fR bytes. If this
-parameter is smaller than \fBzonerange\fR then only a fraction of each zone
-with \fBzonerange\fR bytes will be accessed.  If this parameter is larger than
-\fBzonerange\fR then each zone will be accessed multiple times before skipping
-to the next zone.
+For \fBzonemode\fR=strided, this is the number of bytes to transfer before
+skipping \fBzoneskip\fR bytes. If this parameter is smaller than
+\fBzonerange\fR then only a fraction of each zone with \fBzonerange\fR bytes
+will be accessed.  If this parameter is larger than \fBzonerange\fR then each
+zone will be accessed multiple times before skipping to the next zone.
+
+For \fBzonemode\fR=zbd, this is the size of a single zone. The \fBzonerange\fR
+parameter is ignored in this mode.

 .TP
 .BI zoneskip \fR=\fPint
 .TP
 .BI zoneskip \fR=\fPint

-Skip the specified number of bytes after \fBzonesize\fR bytes of data have been
-transferred.
+For \fBzonemode\fR=strided, the number of bytes to skip after \fBzonesize\fR
+bytes of data have been transferred. This parameter must be zero for
+\fBzonemode\fR=zbd.

 
 .SS "I/O type"
 .TP
 
 .SS "I/O type"
 .TP

diff --git a/init.c b/init.c
index 7514d1ab0af536b288eb5b35be04fe661a2c3c99..38f51288c638ba6ee613697076f958e5ef4b3e23 100644 (file)
--- a/init.c
+++ b/init.c
@@ -618,17 +618,29 @@ static int fixup_options(struct thread_data *td)
                ret |= warnings_fatal;
        }
 
                ret |= warnings_fatal;
        }
 

+       if (o->zone_mode == ZONE_MODE_NONE && o->zone_size) {
+               log_err("fio: --zonemode=none and --zonesize are not compatible.\n");
+               ret |= 1;
+       }
+
+       if (o->zone_mode == ZONE_MODE_NOT_SPECIFIED) {
+               if (o->zone_size)
+                       o->zone_mode = ZONE_MODE_STRIDED;
+               else
+                       o->zone_mode = ZONE_MODE_NONE;
+       }
+

        /*
        /*

-        * only really works with 1 file
+        * Strided zone mode only really works with 1 file.

         */
         */

-       if (o->zone_size && o->open_files > 1)
-               o->zone_size = 0;
+       if (o->zone_mode == ZONE_MODE_STRIDED && o->open_files > 1)
+               o->zone_mode = ZONE_MODE_NONE;

 
        /*
         * If zone_range isn't specified, backward compatibility dictates it
         * should be made equal to zone_size.
         */
 
        /*
         * If zone_range isn't specified, backward compatibility dictates it
         * should be made equal to zone_size.
         */

-       if (o->zone_size && !o->zone_range)
+       if (o->zone_mode == ZONE_MODE_STRIDED && !o->zone_range)

                o->zone_range = o->zone_size;
 
        /*
                o->zone_range = o->zone_size;
 
        /*

diff --git a/io_u.c b/io_u.c
index c58dcf0493bb50347367f8a1d6886ded60ec7748..e050d9f7c775e7f4fabe10b80f9add5d87f8e510 100644 (file)
--- a/io_u.c
+++ b/io_u.c
@@ -64,7 +64,7 @@ static uint64_t last_block(struct thread_data *td, struct fio_file *f,
        if (max_size > f->real_file_size)
                max_size = f->real_file_size;
 
        if (max_size > f->real_file_size)
                max_size = f->real_file_size;
 

-       if (td->o.zone_range)
+       if (td->o.zone_mode == ZONE_MODE_STRIDED && td->o.zone_range)

                max_size = td->o.zone_range;
 
        if (td->o.min_bs[ddir] > td->o.ba[ddir])
                max_size = td->o.zone_range;
 
        if (td->o.min_bs[ddir] > td->o.ba[ddir])


@@ -815,10 +815,14 @@ void requeue_io_u(struct thread_data *td, struct io_u **io_u)
        *io_u = NULL;
 }
 
        *io_u = NULL;
 }
 

-static void __fill_io_u_zone(struct thread_data *td, struct io_u *io_u)
+static void setup_strided_zone_mode(struct thread_data *td, struct io_u *io_u)

 {
        struct fio_file *f = io_u->file;
 
 {
        struct fio_file *f = io_u->file;
 

+       assert(td->o.zone_mode == ZONE_MODE_STRIDED);
+       assert(td->o.zone_size);
+       assert(td->o.zone_range);
+

        /*
         * See if it's time to switch to a new zone
         */
        /*
         * See if it's time to switch to a new zone
         */


@@ -869,11 +873,8 @@ static int fill_io_u(struct thread_data *td, struct io_u *io_u)
        if (!ddir_rw(io_u->ddir))
                goto out;
 
        if (!ddir_rw(io_u->ddir))
                goto out;
 

-       /*
-        * When file is zoned zone_range is always positive
-        */
-       if (td->o.zone_range)
-               __fill_io_u_zone(td, io_u);
+       if (td->o.zone_mode == ZONE_MODE_STRIDED)
+               setup_strided_zone_mode(td, io_u);

 
        /*
         * No log, let the seq/rand engine retrieve the next buflen and
 
        /*
         * No log, let the seq/rand engine retrieve the next buflen and

diff --git a/options.c b/options.c
index 86ab5d6d230cf0df8d218a66aabe90a01adf6cbf..2bac64a343b1030966bcea2658455fc14d8021bc 100644 (file)
--- a/options.c
+++ b/options.c
@@ -3239,6 +3239,30 @@ struct fio_option fio_options[FIO_MAX_OPTS] = {
                .help   = "Your platform does not support IO scheduler switching",
        },
 #endif
                .help   = "Your platform does not support IO scheduler switching",
        },
 #endif

+       {
+               .name   = "zonemode",
+               .lname  = "Zone mode",
+               .help   = "Mode for the zonesize, zonerange and zoneskip parameters",
+               .type   = FIO_OPT_STR,
+               .off1   = offsetof(struct thread_options, zone_mode),
+               .def    = "none",
+               .category = FIO_OPT_C_IO,
+               .group  = FIO_OPT_G_ZONE,
+               .posval = {
+                          { .ival = "none",
+                            .oval = ZONE_MODE_NONE,
+                            .help = "no zoning",
+                          },
+                          { .ival = "strided",
+                            .oval = ZONE_MODE_STRIDED,
+                            .help = "strided mode - random I/O is restricted to a single zone",
+                          },
+                          { .ival = "zbd",
+                            .oval = ZONE_MODE_ZBD,
+                            .help = "zoned block device mode - random I/O selects one of multiple zones randomly",
+                          },
+               },
+       },

        {
                .name   = "zonesize",
                .lname  = "Zone size",
        {
                .name   = "zonesize",
                .lname  = "Zone size",

diff --git a/thread_options.h b/thread_options.h
index 8bbf54bfbafad3fc41e7dac1e7d5a37d31d6afe9..9ef3d6b155ed95b15877c8152adfd4f0b79864db 100644 (file)
--- a/thread_options.h
+++ b/thread_options.h
@@ -10,6 +10,14 @@
 #include "lib/pattern.h"
 #include "td_error.h"
 
 #include "lib/pattern.h"
 #include "td_error.h"
 

+enum fio_zone_mode {
+       ZONE_MODE_NOT_SPECIFIED = 0,
+       ZONE_MODE_NONE          = 1,
+       ZONE_MODE_STRIDED       = 2, /* perform I/O in one zone at a time */
+       /* perform I/O across multiple zones simultaneously */
+       ZONE_MODE_ZBD           = 3,
+};
+

 /*
  * What type of allocation to use for io buffers
  */
 /*
  * What type of allocation to use for io buffers
  */


@@ -188,6 +196,7 @@ struct thread_options {
        unsigned long long zone_range;
        unsigned long long zone_size;
        unsigned long long zone_skip;
        unsigned long long zone_range;
        unsigned long long zone_size;
        unsigned long long zone_skip;

+       enum fio_zone_mode zone_mode;

        unsigned long long lockmem;
        enum fio_memtype mem_type;
        unsigned int mem_align;
        unsigned long long lockmem;
        enum fio_memtype mem_type;
        unsigned int mem_align;


@@ -601,6 +610,8 @@ struct thread_options_pack {
 
        uint32_t allow_create;
        uint32_t allow_mounted_write;
 
        uint32_t allow_create;
        uint32_t allow_mounted_write;

+
+       uint32_t zone_mode;

 } __attribute__((packed));
 
 extern void convert_thread_options_to_cpu(struct thread_options *o, struct thread_options_pack *top);
 } __attribute__((packed));
 
 extern void convert_thread_options_to_cpu(struct thread_options *o, struct thread_options_pack *top);