OPAE Installation Guide¶
System compatibility¶
The OPAE SDK has been tested on the following configurations.
- Hardware: Intel(R) FPGA Programmable Acceleration Cards: Arria(R) 10 GX, N3000.
- Operating System: Tested on Fedora 32, with Linux kernel 5.8.
- Arria® 10 GX FPGA FIM version: 1.0.3 (1.0 Production)
How to download the OPAE SDK¶
OPAE SDK releases are available on GitHub. Source code for the OPAE DFL device driver for Linux is also available on GitHub.
Install the Fedora 32¶
Download the Fedora 32 (x86_64 version) installation file in fedora, and install the Fedora 32 in yourserver. You can choose Fedora Workstation or Fedora server.
Build the kernel and DFL drivers¶
For building the OPAE kernel and kernel driver, the kernel development environment is required. So before you build the kernel, you must install the required packages. Run the following commands:
$ sudo yum install gcc gcc-c++ make kernel-headers kernel-devel elfutils-libelf-devel ncurses-devel openssl-devel bison flex
Download the OPAE upstream kernel tree from github.
$ git clone https://github.com/OPAE/linux-dfl.git -b fpga-upstream-dev-5.8.0
Configure the kernel.
$ cd linux-dfl
$ cp /boot/config-`uname -r` .config
$ cat configs/n3000_d5005_defconfig >> .config
$ echo 'CONFIG_LOCALVERSION="-dfl"' >> .config
$ echo 'CONFIG_LOCALVERSION_AUTO=y' >> .config
$ make olddefconfig
Compile and install the new kernel.
$ make -j
$ sudo make modules_install
$ sudo make install
When installed finished, reboot your system. When the system login again, check the kernel version is correctly or not.
[figo@localhost linux-dfl]$ uname -a
Linux localhost.localdomain 5.8.0-rc1-dfl-g73e16386cda0 #6 SMP Wed Aug 19 08:38:32 EDT 2020 x86_64 x86_64 x86_64 GNU/Linux
And also you can check the OPAE dfl drivers have auto-loaded or not.
[figo@localhost linux-dfl]$ lsmod | grep fpga
ifpga_sec_mgr 20480 1 intel_m10_bmc_secure
fpga_region 20480 3 dfl_fme_region,dfl_fme,dfl
fpga_bridge 24576 4 dfl_fme_region,fpga_region,dfl_fme,dfl_fme_br
fpga_mgr 16384 4 dfl_fme_region,fpga_region,dfl_fme_mgr,dfl_fme
[figo@localhost linux-dfl]$ lsmod | grep dfl
dfl_eth_group 36864 0
dfl_fme_region 20480 0
dfl_emif 16384 0
dfl_n3000_nios 20480 0
dfl_fme_br 16384 0
dfl_fme_mgr 20480 1
dfl_fme 49152 0
dfl_afu 36864 0
dfl_pci 20480 0
dfl 40960 7 dfl_pci,dfl_fme,dfl_fme_br,dfl_eth_group,dfl_n3000_nios,dfl_afu,dfl_emif
fpga_region 20480 3 dfl_fme_region,dfl_fme,dfl
fpga_bridge 24576 4 dfl_fme_region,fpga_region,dfl_fme,dfl_fme_br
fpga_mgr 16384 4 dfl_fme_region,fpga_region,dfl_fme_mgr,dfl_fme
Build the OPAE-SDK¶
Before you build the kernel, you must install the required packages. Run the following commands:
$ sudo yum install cmake libuuid libuuid-devel json-c python3-devel python3-libs json-c-devel hwloc-devel uuid python3-pip python3-virtualenv tbb-devel rpm-build
Download the OPAE-SDK source code from github.
$ git clone https://github.com/OPAE/opae-sdk.git
Compile and build the OPAE-SDK.
$ cd opae-sdk
$ mkdir build
$ cmake .. -DCPACK_GENERATOR=RPM -DOPAE_BUILD_LEGACY=ON
$ make -j
$ make -j package_rpm
After compile successful, there are 8 rpm packages generated.
opae-2.0.0-1.x86_64.rpm
opae-devel-2.0.0-1.x86_64.rpm
opae-libs-2.0.0-1.x86_64.rpm
opae-opae.admin-2.0.0-1.x86_64.rpm
opae-PACSign-2.0.0-1.x86_64.rpm
opae-tests-2.0.0-1.x86_64.rpm
opae-tools-2.0.0-1.x86_64.rpm
opae-tools-extra-2.0.0-1.x86_64.rpm
OPAE SDK installation with rpm packages¶
The rpm packages generated in the previous step can be installed using these commands:
$ sudo yum install opae-libs-<release>.x86_64.rpm
$ sudo yum install opae-tools-<release>.x86_64.rpm
$ sudo yum install opae-tools-extra-<release>.x86_64.rpm
$ sudo yum install opae-devel-<release>.x86_64.rpm
$ sudo yum install opae-<release>.x86_64.rpm
$ sudo yum install opae-opae.admin-<release>.x86_64.rpm
$ sudo yum install opae-PACSign-<release>.x86_64.rpm
$ sudo yum install opae-tests-<release>.x86_64.rpm
When you installed the rpms, you can use fpgainfo command to check the FPGA FME infomation.
[figo@localhost install_guide]$ fpgainfo fme
Board Management Controller, MAX10 NIOS FW version: D.2.1.24
Board Management Controller, MAX10 Build version: D.2.0.7
//****** FME ******//
Object Id : 0xEF00000
PCIe s:b:d.f : 0000:08:00.0
Device Id : 0x0B30
Socket Id : 0x00
Ports Num : 01
Bitstream Id : 0x2300011001030F
Bitstream Version : 0.2.3
Pr Interface Id : f3c99413-5081-4aad-bced-07eb84a6d0bb
Boot Page : user
To uninstall the OPAE rpms, you can use this commands
$ sudo yum remove opae-libs
$ sudo yum remove opae-tools
$ sudo yum remove opae-tools-extra
$ sudo yum remove opae-devel
$ sudo yum remove opae
$ sudo yum remove opae-opae.admin
$ sudo yum remove opae-PACSign
$ sudo yum remove opae-tests
FPGA Device Access Permissions¶
Access to FPGA accelerators and devices is controlled using file access
permissions on the Intel® FPGA device files, /dev/dfl-fme.*
and
/dev/dfl-port.*
, as well as to the files reachable through
/sys/class/fpga_region/
.
In order to allow regular (non-root) users to access accelerators, you
need to grant them read and write permissions on /dev/dfl-port.*
(with *
denoting the respective socket, i.e. 0 or 1). E.g.:
$ sudo chmod a+rw /dev/dfl-port.0
Memlock limit¶
Depending on the requirements of your application, you may also want to increase the maximum amount of memory a user process is allowed to lock. The exact way to do this depends on your Linux distribution.
You can check the current memlock limit using
$ ulimit -l
A way to permanently remove the limit for locked memory for a regular user is to add the following lines to your /etc/security/limits.conf:
user1 hard memlock unlimited
user1 soft memlock unlimited
This removes the limit on locked memory for user user1
. To remove it
for all users, you can replace user1
with *
:
* hard memlock unlimited
* soft memlock unlimited
Note that settings in the /etc/security/limits.conf file don’t apply to services. To increase the locked memory limit for a service you need to modify the application’s systemd service file and add the line:
[Service]
LimitMEMLOCK=infinity
Hugepage Settings¶
Users need to configure system hugepages to reserve 2MB-hugepages or 1GB-hugepages. For example, the ‘hello_fpga’ sample requires several 2MB-hugepages. And the fpgadiag tool requires several 1GB-hugepages.
The command below reserves 20 2M-hugepages:
$ sudo sh -c 'echo 20 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages'
The command below reserves 4 1GB-hugepages:
$ sudo sh -c 'echo 4 > /sys/kernel/mm/hugepages/hugepages-1048576kB/nr_hugepages'
For x86_64 architecture processors, user can use following command to find out avaiable hugepage sizes:
$ grep pse /proc/cpuinfo | uniq
flags : ... pse ...
If this commands returns a non-empty string, 2MB pages are supported.
$ grep pse /proc/cpuinfo | uniq
flags : ... pdpe1gb ...
If this commands returns a non-empty string, 1GB pages are supported.