1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright(c) 2015-2016 Intel Corporation.
4 Intel(R) QuickAssist (QAT) Crypto Poll Mode Driver
5 ==================================================
7 QAT documentation consists of three parts:
9 * Details of the symmetric crypto service below.
10 * Details of the `compression service <http://dpdk.org/doc/guides/compressdevs/qat_comp.html>`_
11 in the compressdev drivers section.
12 * Details of building the common QAT infrastructure and the PMDs to support the
13 above services. See :ref:`building_qat` below.
16 Symmetric Crypto Service on QAT
17 -------------------------------
19 The QAT crypto PMD provides poll mode crypto driver support for the following
20 hardware accelerator devices:
22 * ``Intel QuickAssist Technology DH895xCC``
23 * ``Intel QuickAssist Technology C62x``
24 * ``Intel QuickAssist Technology C3xxx``
25 * ``Intel QuickAssist Technology D15xx``
31 The QAT PMD has support for:
35 * ``RTE_CRYPTO_CIPHER_3DES_CBC``
36 * ``RTE_CRYPTO_CIPHER_3DES_CTR``
37 * ``RTE_CRYPTO_CIPHER_AES128_CBC``
38 * ``RTE_CRYPTO_CIPHER_AES192_CBC``
39 * ``RTE_CRYPTO_CIPHER_AES256_CBC``
40 * ``RTE_CRYPTO_CIPHER_AES128_CTR``
41 * ``RTE_CRYPTO_CIPHER_AES192_CTR``
42 * ``RTE_CRYPTO_CIPHER_AES256_CTR``
43 * ``RTE_CRYPTO_CIPHER_SNOW3G_UEA2``
44 * ``RTE_CRYPTO_CIPHER_NULL``
45 * ``RTE_CRYPTO_CIPHER_KASUMI_F8``
46 * ``RTE_CRYPTO_CIPHER_DES_CBC``
47 * ``RTE_CRYPTO_CIPHER_AES_DOCSISBPI``
48 * ``RTE_CRYPTO_CIPHER_DES_DOCSISBPI``
49 * ``RTE_CRYPTO_CIPHER_ZUC_EEA3``
53 * ``RTE_CRYPTO_AUTH_SHA1_HMAC``
54 * ``RTE_CRYPTO_AUTH_SHA224_HMAC``
55 * ``RTE_CRYPTO_AUTH_SHA256_HMAC``
56 * ``RTE_CRYPTO_AUTH_SHA384_HMAC``
57 * ``RTE_CRYPTO_AUTH_SHA512_HMAC``
58 * ``RTE_CRYPTO_AUTH_AES_XCBC_MAC``
59 * ``RTE_CRYPTO_AUTH_SNOW3G_UIA2``
60 * ``RTE_CRYPTO_AUTH_MD5_HMAC``
61 * ``RTE_CRYPTO_AUTH_NULL``
62 * ``RTE_CRYPTO_AUTH_KASUMI_F9``
63 * ``RTE_CRYPTO_AUTH_AES_GMAC``
64 * ``RTE_CRYPTO_AUTH_ZUC_EIA3``
66 Supported AEAD algorithms:
68 * ``RTE_CRYPTO_AEAD_AES_GCM``
74 * Only supports the session-oriented API implementation (session-less APIs are not supported).
75 * SNOW 3G (UEA2), KASUMI (F8) and ZUC (EEA3) supported only if cipher length and offset fields are byte-multiple.
76 * SNOW 3G (UIA2) and ZUC (EIA3) supported only if hash length and offset fields are byte-multiple.
77 * No BSD support as BSD QAT kernel driver not available.
78 * ZUC EEA3/EIA3 is not supported by dh895xcc devices
79 * Maximum additional authenticated data (AAD) for GCM is 240 bytes long.
80 * Queue pairs are not thread-safe (that is, within a single queue pair, RX and TX from different lcores is not supported).
83 Extra notes on KASUMI F9
84 ~~~~~~~~~~~~~~~~~~~~~~~~
86 When using KASUMI F9 authentication algorithm, the input buffer must be
87 constructed according to the
88 `3GPP KASUMI specification <http://cryptome.org/3gpp/35201-900.pdf>`_
89 (section 4.4, page 13). The input buffer has to have COUNT (4 bytes),
90 FRESH (4 bytes), MESSAGE and DIRECTION (1 bit) concatenated. After the DIRECTION
91 bit, a single '1' bit is appended, followed by between 0 and 7 '0' bits, so that
92 the total length of the buffer is multiple of 8 bits. Note that the actual
93 message can be any length, specified in bits.
95 Once this buffer is passed this way, when creating the crypto operation,
96 length of data to authenticate "op.sym.auth.data.length" must be the length
97 of all the items described above, including the padding at the end.
98 Also, offset of data to authenticate "op.sym.auth.data.offset"
99 must be such that points at the start of the COUNT bytes.
108 A QAT device can host multiple acceleration services:
110 * symmetric cryptography
113 These services are provided to DPDK applications via PMDs which register to
114 implement the corresponding cryptodev and compressdev APIs. The PMDs use
115 common QAT driver code which manages the QAT PCI device. They also depend on a
116 QAT kernel driver being installed on the platform, see :ref:`qat_kernel` below.
119 Configuring and Building the DPDK QAT PMDs
120 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
123 Further information on configuring, building and installing DPDK is described
124 `here <http://dpdk.org/doc/guides/linux_gsg/build_dpdk.html>`_.
127 Quick instructions for QAT cryptodev PMD are as follows:
129 .. code-block:: console
131 cd to the top-level DPDK directory
133 sed -i 's,\(CONFIG_RTE_LIBRTE_PMD_QAT_SYM\)=n,\1=y,' build/.config
136 Quick instructions for QAT compressdev PMD are as follows:
138 .. code-block:: console
140 cd to the top-level DPDK directory
148 These are the build configuration options affecting QAT, and their default values:
150 .. code-block:: console
152 CONFIG_RTE_LIBRTE_PMD_QAT=y
153 CONFIG_RTE_LIBRTE_PMD_QAT_SYM=n
154 CONFIG_RTE_PMD_QAT_MAX_PCI_DEVICES=48
155 CONFIG_RTE_PMD_QAT_COMP_SGL_MAX_SEGMENTS=16
157 CONFIG_RTE_LIBRTE_PMD_QAT must be enabled for any QAT PMD to be built.
159 The QAT cryptodev PMD has an external dependency on libcrypto, so is not
160 built by default. CONFIG_RTE_LIBRTE_PMD_QAT_SYM should be enabled to build it.
162 The QAT compressdev PMD has no external dependencies, so needs no configuration
163 options and is built by default.
165 The number of VFs per PF varies - see table below. If multiple QAT packages are
166 installed on a platform then CONFIG_RTE_PMD_QAT_MAX_PCI_DEVICES should be
167 adjusted to the number of VFs which the QAT common code will need to handle.
168 Note, there is a separate config item for max cryptodevs CONFIG_RTE_CRYPTO_MAX_DEVS,
169 if necessary this should be adjusted to handle the total of QAT and other devices
170 which the process will use.
172 QAT allocates internal structures to handle SGLs. For the compression service
173 CONFIG_RTE_PMD_QAT_COMP_SGL_MAX_SEGMENTS can be changed if more segments are needed.
174 An extra (max_inflight_ops x 16) bytes per queue_pair will be used for every increment.
177 Device and driver naming
178 ~~~~~~~~~~~~~~~~~~~~~~~~
180 * The qat cryptodev driver name is "crypto_qat".
181 The "rte_cryptodev_devices_get()" returns the devices exposed by this driver.
183 * Each qat crypto device has a unique name, in format
184 "<pci bdf>_<service>", e.g. "0000:41:01.0_qat_sym".
185 This name can be passed to "rte_cryptodev_get_dev_id()" to get the device_id.
189 The qat crypto driver name is passed to the dpdk-test-crypto-perf tool in the "-devtype" parameter.
191 The qat crypto device name is in the format of the slave parameter passed to the crypto scheduler.
193 * The qat compressdev driver name is "qat".
194 The rte_compressdev_devices_get() returns the devices exposed by this driver.
196 * Each qat compression device has a unique name, in format
197 <pci bdf>_<service>, e.g. "0000:41:01.0_qat_comp".
198 This name can be passed to rte_compressdev_get_dev_id() to get the device_id.
202 Dependency on the QAT kernel driver
203 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
205 To use QAT an SRIOV-enabled QAT kernel driver is required. The VF
206 devices created and initialised by this driver will be used by the QAT PMDs.
208 Instructions for installation are below, but first an explanation of the
209 relationships between the PF/VF devices and the PMDs visible to
212 Each QuickAssist PF device exposes a number of VF devices. Each VF device can
213 enable one cryptodev PMD and/or one compressdev PMD.
214 These QAT PMDs share the same underlying device and pci-mgmt code, but are
215 enumerated independently on their respective APIs and appear as independent
216 devices to applications.
220 Each VF can only be used by one DPDK process. It is not possible to share
221 the same VF across multiple processes, even if these processes are using
222 different acceleration services.
224 Conversely one DPDK process can use one or more QAT VFs and can expose both
225 cryptodev and compressdev instances on each of those VFs.
228 Available kernel drivers
229 ~~~~~~~~~~~~~~~~~~~~~~~~
231 Kernel drivers for each device are listed in the following table. Scroll right
232 to check that the driver and device supports the service you require.
235 .. _table_qat_pmds_drivers:
237 .. table:: QAT device generations, devices and drivers
239 +-----+----------+---------------+---------------+------------+--------+------+--------+--------+-----------+-------------+
240 | Gen | Device | Driver/ver | Kernel Module | Pci Driver | PF Did | #PFs | VF Did | VFs/PF | cryptodev | compressdev |
241 +=====+==========+===============+===============+============+========+======+========+========+===========+=============+
242 | 1 | DH895xCC | linux/4.4+ | qat_dh895xcc | dh895xcc | 435 | 1 | 443 | 32 | Yes | No |
243 +-----+----------+---------------+---------------+------------+--------+------+--------+--------+-----------+-------------+
244 | " | " | 01.org/4.2.0+ | " | " | " | " | " | " | Yes | No |
245 +-----+----------+---------------+---------------+------------+--------+------+--------+--------+-----------+-------------+
246 | 2 | C62x | linux/4.5+ | qat_c62x | c6xx | 37c8 | 3 | 37c9 | 16 | Yes | No |
247 +-----+----------+---------------+---------------+------------+--------+------+--------+--------+-----------+-------------+
248 | " | " | 01.org/4.2.0+ | " | " | " | " | " | " | Yes | Yes |
249 +-----+----------+---------------+---------------+------------+--------+------+--------+--------+-----------+-------------+
250 | 2 | C3xxx | linux/4.5+ | qat_c3xxx | c3xxx | 19e2 | 1 | 19e3 | 16 | Yes | No |
251 +-----+----------+---------------+---------------+------------+--------+------+--------+--------+-----------+-------------+
252 | " | " | 01.org/4.2.0+ | " | " | " | " | " | " | Yes | Yes |
253 +-----+----------+---------------+---------------+------------+--------+------+--------+--------+-----------+-------------+
254 | 2 | D15xx | p | qat_d15xx | d15xx | 6f54 | 1 | 6f55 | 16 | Yes | No |
255 +-----+----------+---------------+---------------+------------+--------+------+--------+--------+-----------+-------------+
258 The ``Driver`` column indicates either the Linux kernel version in which
259 support for this device was introduced or a driver available on Intel's 01.org
260 website. There are both linux and 01.org kernel drivers available for some
261 devices. p = release pending.
263 If you are running on a kernel which includes a driver for your device, see
264 `Installation using kernel.org driver`_ below. Otherwise see
265 `Installation using 01.org QAT driver`_.
268 Installation using kernel.org driver
269 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
271 The examples below are based on the C62x device, if you have a different device
272 use the corresponding values in the above table.
274 In BIOS ensure that SRIOV is enabled and either:
277 * Enable VT-d and set ``"intel_iommu=on iommu=pt"`` in the grub file.
279 Check that the QAT driver is loaded on your system, by executing::
283 You should see the kernel module for your device listed, e.g.::
286 intel_qat 82336 1 qat_c62x
288 Next, you need to expose the Virtual Functions (VFs) using the sysfs file system.
290 First find the BDFs (Bus-Device-Function) of the physical functions (PFs) of
295 You should see output similar to::
297 1a:00.0 Co-processor: Intel Corporation Device 37c8
298 3d:00.0 Co-processor: Intel Corporation Device 37c8
299 3f:00.0 Co-processor: Intel Corporation Device 37c8
301 Enable the VFs for each PF by echoing the number of VFs per PF to the pci driver::
303 echo 16 > /sys/bus/pci/drivers/c6xx/0000:1a:00.0/sriov_numvfs
304 echo 16 > /sys/bus/pci/drivers/c6xx/0000:3d:00.0/sriov_numvfs
305 echo 16 > /sys/bus/pci/drivers/c6xx/0000:3f:00.0/sriov_numvfs
307 Check that the VFs are available for use. For example ``lspci -d:37c9`` should
308 list 48 VF devices available for a ``C62x`` device.
310 To complete the installation follow the instructions in
311 `Binding the available VFs to the DPDK UIO driver`_.
315 If the QAT kernel modules are not loaded and you see an error like ``Failed
316 to load MMP firmware qat_895xcc_mmp.bin`` in kernel logs, this may be as a
317 result of not using a distribution, but just updating the kernel directly.
319 Download firmware from the `kernel firmware repo
320 <http://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git/tree/>`_.
322 Copy qat binaries to ``/lib/firmware``::
324 cp qat_895xcc.bin /lib/firmware
325 cp qat_895xcc_mmp.bin /lib/firmware
327 Change to your linux source root directory and start the qat kernel modules::
329 insmod ./drivers/crypto/qat/qat_common/intel_qat.ko
330 insmod ./drivers/crypto/qat/qat_dh895xcc/qat_dh895xcc.ko
335 If you see the following warning in ``/var/log/messages`` it can be ignored:
336 ``IOMMU should be enabled for SR-IOV to work correctly``.
339 Installation using 01.org QAT driver
340 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
342 Download the latest QuickAssist Technology Driver from `01.org
343 <https://01.org/packet-processing/intel%C2%AE-quickassist-technology-drivers-and-patches>`_.
344 Consult the *Getting Started Guide* at the same URL for further information.
346 The steps below assume you are:
348 * Building on a platform with one ``C62x`` device.
349 * Using package ``qat1.7.l.4.2.0-000xx.tar.gz``.
350 * On Fedora26 kernel ``4.11.11-300.fc26.x86_64``.
352 In the BIOS ensure that SRIOV is enabled and VT-d is disabled.
354 Uninstall any existing QAT driver, for example by running:
356 * ``./installer.sh uninstall`` in the directory where originally installed.
359 Build and install the SRIOV-enabled QAT driver::
364 # Copy the package to this location and unpack
365 tar zxof qat1.7.l.4.2.0-000xx.tar.gz
367 ./configure --enable-icp-sriov=host
370 You can use ``cat /sys/kernel/debug/qat<your device type and bdf>/version/fw`` to confirm the driver is correctly installed and is using firmware version 4.2.0.
371 You can use ``lspci -d:37c9`` to confirm the presence of the 16 VF devices available per ``C62x`` PF.
373 Confirm the driver is correctly installed and is using firmware version 4.2.0::
375 cat /sys/kernel/debug/qat<your device type and bdf>/version/fw
378 Confirm the presence of 48 VF devices - 16 per PF::
383 To complete the installation - follow instructions in `Binding the available VFs to the DPDK UIO driver`_.
387 If using a later kernel and the build fails with an error relating to
388 ``strict_stroul`` not being available apply the following patch:
392 /QAT/QAT1.6/quickassist/utilities/downloader/Target_CoreLibs/uclo/include/linux/uclo_platform.h
393 + #if LINUX_VERSION_CODE >= KERNEL_VERSION(3,18,5)
394 + #define STR_TO_64(str, base, num, endPtr) {endPtr=NULL; if (kstrtoul((str), (base), (num))) printk("Error strtoull convert %s\n", str); }
396 #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,38)
397 #define STR_TO_64(str, base, num, endPtr) {endPtr=NULL; if (strict_strtoull((str), (base), (num))) printk("Error strtoull convert %s\n", str); }
399 #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,25)
400 #define STR_TO_64(str, base, num, endPtr) {endPtr=NULL; strict_strtoll((str), (base), (num));}
402 #define STR_TO_64(str, base, num, endPtr) \
406 *(num) = -(simple_strtoull((str+1), &(endPtr), (base))); \
408 *(num) = simple_strtoull((str), &(endPtr), (base)); \
418 If the build fails due to missing header files you may need to do following::
420 sudo yum install zlib-devel
421 sudo yum install openssl-devel
422 sudo yum install libudev-devel
426 If the build or install fails due to mismatching kernel sources you may need to do the following::
428 sudo yum install kernel-headers-`uname -r`
429 sudo yum install kernel-src-`uname -r`
430 sudo yum install kernel-devel-`uname -r`
433 Binding the available VFs to the DPDK UIO driver
434 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
436 Unbind the VFs from the stock driver so they can be bound to the uio driver.
438 For an Intel(R) QuickAssist Technology DH895xCC device
439 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
441 The unbind command below assumes ``BDFs`` of ``03:01.00-03:04.07``, if your
442 VFs are different adjust the unbind command below::
444 for device in $(seq 1 4); do \
445 for fn in $(seq 0 7); do \
446 echo -n 0000:03:0${device}.${fn} > \
447 /sys/bus/pci/devices/0000\:03\:0${device}.${fn}/driver/unbind; \
451 For an Intel(R) QuickAssist Technology C62x device
452 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
454 The unbind command below assumes ``BDFs`` of ``1a:01.00-1a:02.07``,
455 ``3d:01.00-3d:02.07`` and ``3f:01.00-3f:02.07``, if your VFs are different
456 adjust the unbind command below::
458 for device in $(seq 1 2); do \
459 for fn in $(seq 0 7); do \
460 echo -n 0000:1a:0${device}.${fn} > \
461 /sys/bus/pci/devices/0000\:1a\:0${device}.${fn}/driver/unbind; \
463 echo -n 0000:3d:0${device}.${fn} > \
464 /sys/bus/pci/devices/0000\:3d\:0${device}.${fn}/driver/unbind; \
466 echo -n 0000:3f:0${device}.${fn} > \
467 /sys/bus/pci/devices/0000\:3f\:0${device}.${fn}/driver/unbind; \
471 For Intel(R) QuickAssist Technology C3xxx or D15xx device
472 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
474 The unbind command below assumes ``BDFs`` of ``01:01.00-01:02.07``, if your
475 VFs are different adjust the unbind command below::
477 for device in $(seq 1 2); do \
478 for fn in $(seq 0 7); do \
479 echo -n 0000:01:0${device}.${fn} > \
480 /sys/bus/pci/devices/0000\:01\:0${device}.${fn}/driver/unbind; \
484 Bind to the DPDK uio driver
485 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
487 Install the DPDK igb_uio driver, bind the VF PCI Device id to it and use lspci
488 to confirm the VF devices are now in use by igb_uio kernel driver,
489 e.g. for the C62x device::
491 cd to the top-level DPDK directory
493 insmod ./build/kmod/igb_uio.ko
494 echo "8086 37c9" > /sys/bus/pci/drivers/igb_uio/new_id
498 Another way to bind the VFs to the DPDK UIO driver is by using the
499 ``dpdk-devbind.py`` script::
501 cd to the top-level DPDK directory
502 ./usertools/dpdk-devbind.py -b igb_uio 0000:03:01.1
507 QAT crypto PMD can be tested by running the test application::
512 ./test -l1 -n1 -w <your qat bdf>
513 RTE>>cryptodev_qat_autotest
515 QAT compression PMD can be tested by running the test application::
518 sed -i 's,\(CONFIG_RTE_COMPRESSDEV_TEST\)=n,\1=y,' build/.config
521 ./test -l1 -n1 -w <your qat bdf>
522 RTE>>compressdev_autotest
528 There are 2 sets of trace available via the dynamic logging feature:
530 * pmd.qat_dp exposes trace on the data-path.
531 * pmd.qat_general exposes all other trace.
533 pmd.qat exposes both sets of traces.
534 They can be enabled using the log-level option (where 8=maximum log level) on
535 the process cmdline, e.g. using any of the following::
537 --log-level="pmd.qat_general,8"
538 --log-level="pmd.qat_dp,8"
539 --log-level="pmd.qat,8"
543 The global RTE_LOG_DP_LEVEL overrides data-path trace so must be set to
544 RTE_LOG_DEBUG to see all the trace. This variable is in config/rte_config.h
545 for meson build and config/common_base for gnu make.
546 Also the dynamic global log level overrides both sets of trace, so e.g. no
547 QAT trace would display in this case::
549 --log-level="7" --log-level="pmd.qat_general,8"