6b1e4e732d
This patch also drops support for automatically unbinding devices from the kernel - run scripts/setup.sh first. Our generic pci interface is now hidden behind include/spdk/pci.h and implemented in lib/util/pci.c. We no longer wrap the calls in nvme_impl.h or ioat_impl.h. The implementation now only uses DPDK and the libpciaccess dependency has been removed. If using a version of DPDK earlier than 16.07, enumerating devices by class code isn't available and only Intel SSDs will be discovered. DPDK 16.07 adds enumeration by class code and all NVMe devices will be correctly discovered. Change-Id: I0e8bac36b5ca57df604a2b310c47342c67dc9f3c Signed-off-by: Ben Walker <benjamin.walker@intel.com> |
||
|---|---|---|
| .. | ||
| .gitignore | ||
| example_config.fio | ||
| fio_plugin.c | ||
| Makefile | ||
| README.md | ||
Compiling
First, clone the fio source repository from http://github.com/axboe/fio
git clone http://github.com/axboe/fio
Then check out the fio 2.8 tag
cd fio && git checkout fio-2.8
Finally, compile the code with
./configure && make
Next, edit the CONFIG file located in the root of the SPDK repository and set CONFIG_FIO_PLUGIN to y and FIO_SOURCE_DIR to the location of the fio repository that was just created.
Further, you'll need to build DPDK with -fPIC set. You can do this by modifying your DPDK config file (i.e. config/defconfig_x86_64-native-linuxapp-gcc) to include the line
EXTRA_CFLAGS=-fPIC
At this point, build SPDK as per normal. The fio plugin will be placed in the same directory as this README.
Usage
To use the SPDK fio plugin with fio, simply set the following in the fio configuration file (see example_config.fio in the same directory as this README).
ioengine=<path to fio_plugin binary>
To select NVMe devices, you simply pass an identifier as the filename in the format
domain.bus.slot.func/namespace
Remember that NVMe namespaces start at 1, not 0! Also, the notation uses '.' throughout, not ':'. For example - 0000.04.00.0/1.
Currently the fio_plugin is limited to a single thread, so only one job is supported.
When testing random workloads, it is recommended to set norandommap=1. fio's random map processing consumes extra CPU cycles which will degrade performance over time with the fio_plugin since all I/O are submitted and completed on a single CPU core.