NVIDIA DOCA RegEx Programming Guide
This document provides developers with instructions on building and developing applications that wish to use RegEx pattern matching.
DOCA RegEx is a library that provides RegEx pattern matching to DOCA applications. It provides access to the regular expression processor (RXP) , a high-performance, hardware-accelerated RegEx engine available on the NVIDIA® BlueField® DPUs, and can utilize software-based engines when required.
Using DOCA RegEx, developers can easily execute complex regular expression operations in an optimized, hardware-accelerated way.
This document is intended for software developers wishing to accelerate their regular expressions operations.
DOCA RegEx-based applications can run either on the host machine or on the DPU target. The RegEx engine is enabled by default on the DPU. However, to enable RegEx offloading on the host, run:
host> sudo /etc/init.d/openibd stop
host> sudo echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages
dpu> echo 1 > /sys/bus/pci/devices/0000\:03\:00.0/regex/pf/regex_en
dpu> cat /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages 400
# Make sure to allocate 200 additional hugepages
dpu> sudo echo 600 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages
dpu> systemctl restart mlx-regex
# Verify the service is properly running
dpu> systemctl status mlx-regex
host> sudo /etc/init.d/openibd start
DOCA RegEx provides a flexible API for programming regular expression databases, enqueuing jobs and dequeuing results. The API operates asynchronously allowing many pattern matching operations to be executed in parallel.
3.1. Rule Compilation
Regular expressions are provided as "compiled" rule files to the library, and must therefore be externally compiled by a "compiler" prior to loading by the library. For hardware acceleration, the external compiler is termed " rxpc" (RXP compiler) and generates RXP object format (ROF) binary files that represent the compiled regular expressions.
3.2. RegEx Implementations
The library itself is designed to support multiple RegEx engine implementations. Currently, only hardware devices are supported. Software devices will be introduced in in a later release.
3.3. Huge Job Emulation
The library includes a facility to accept job lengths that are greater than the maximum size supported by an engine. The library fragments incoming jobs into smaller fragments and processes them sequentially looking for potential matches. The "huge job emulation" mechanism takes data from the end of the previous fragment and appends it to the start of the next fragment (the "size" of the overlap) to find additional matches. See the doca_regex_property_huge_job_emulation_overlap_set API call for more information.
This section details the specific structures and API operations related to the DOCA RegEx library.
The pkg-config (*.pc file) for the RegEx library is named doca-regex.
4.1. doca_regex_job_request
This structure contains information on the job to be submitted to DOCA RegEx.
struct doca_regex_job_request {
uint64_t id;
uint16_t rule_group_ids[4];
struct doca_buf const *buffer;
};
Where:
id– a user-defined field used to correlate the matches with the enqueued jobrule_group_ids– an array of IDs which can be used to select which group of rules are used to process this job. Set each value to a non-zero value to enable group selection, or to 0 to ignore it.buffer– a pointer to a buffer containing the data to be scanned
4.2. doca_regex_job_response
When a job response is dequeued, this structure is populated with any match information.
struct doca_regex_job_response {
uint64_t id;
uint64_t status_flags;
uint32_t detected_matches;
uint32_t num_matches;
struct doca_regex_match *matches;
struct doca_regex_mempool *matches_mempool;
};
Where:
id– theidvalue as supplied by the user during enqueue. See doca_regex_job_request for more information.status_flags– a bit-masked field for zero or more status flags. Seedoca_regex_status_flagfor more information.detected_matches– the total number of detected matchesnum_matches– the total number of matches returned in this response (may be fewer thandetected_matches)matches– a linked list of match structures (num_matchesin length)-
matches_mempool- thisdoca_regex-owned pointer to the mempool of matches is provided to release the results after any required processing.Note:matches_mempoolcontains all matches present, these must be released back to the mempool after you have finished processing them. Seedoca_regex_mempool_obj_putin the NVIDIA DOCA Libraries API Reference Manual for more information.
4.3. doca_regex_match
When a job response is dequeued, this structure is populated with any match information.
struct doca_regex_match {
struct doca_regex_match *next;
uint32_t match_start;
uint32_t rule_id;
uint32_t length;
};
Where:
next– as matches are linked together using a linked list, this is the pointer to the next match in the linked listmatch_start– the index relative to the start of the job of this matchrule_id– the ID of the rule that generated this matchlength– the length of the matched value
4.4. Instance Construction/Destruction API
This section details API calls related to the creation and destruction of DOCA RegEx instances.
4.4.1. doca_regex_create
Creates a DOCA RegEx instance.
struct doca_regex *doca_regex_create(void);
This function returns doca_regex object on success. NULL otherwise.
4.4.2. doca_regex_destroy
Destroys a previously created DOCA RegEx instance.
void doca_regex_destroy(struct doca_regex *regex);
Where:
regex [in]– a pointer to a previously created DOCA RegEx instance
4.5. RegEx Device Management
DOCA RegEx uses doca_dev devices to facilitate hardware accelerated RegEx implementations that perform pattern matching. Currently, DOCA RegEx supports the use of hardware devices.
RegEx devices are separate user-managed objects that must be created and registered with DOCA RegEx as either hardware or software devices.
4.5.1. doca_regex_dev_add
This function is used to add or attach a new RegEx device.
int doca_regex_dev_add(struct doca_regex *regex, struct doca_dev *dev);
Where:
regex [in]– instance pointer of the RegEx instancedev [in]– instance pointer to the DOCA device
The function returns 0 on success and a negative status code on failure.
4.5.2. doca_regex_dev_rm
This function is used to remove or detach a previously added/attached RegEx device.
int doca_regex_dev_rm(struct doca_regex *regex, struct doca_dev *dev);
Where:
regex [in]– instance pointer of the RegEx instancedev [in]– instance pointer to the DOCA device
The function returns 0 on success and a negative status code on failure.
4.6. DOCA RegEx Setup
This section details the API calls required to setup DOCA RegEx with memory to store received matches, adjust the number of queue pairs, etc.
4.6.1. doca_regex_num_qps_set
Specifies the number of queue pairs to use for this DOCA RegEx instance. This function should only be called when the instance is not running.
int doca_regex_num_qps_set(struct doca_regex *regex, uint16_t num_qps);
Where:
regex [in]– the DOCA RegEx instancenum_qps [in]– the number of queue pairs to assign to the instance. Default is 0.
The function returns 0 on success and a negative status code on failure.
4.6.2. doca_regex_property_matches_memory_pool_size_set
This function allocates memory pools for the required number of matches. The num_mem_pool_elements value provides the total number of matches in the pool. It should be noted that these are shared across all jobs on the thread. For example, if you wanted 1000 jobs in flight with each job capable of returning 15 matches, a value of 15k should be provided.
doca_error_t doca_regex_property_matches_memory_pool_size_set(struct doca_regex *regex, uint16_t num_mem_pool_elements);
Where:
regex [in]– the DOCA RegEx instancenum_mem_pool_elements [in]– the total number of matches the application expects to see in a dequeue result
This function returns DOCA_SUCCESS, or the relevant DOCA error code.
4.7. Configuration Options
DOCA RegEx has several options that alter its mode of operation and control certain features. This section details those API calls and their related impact.
4.7.1. doca_regex_property_huge_job_emulation_overlap_set
This API call enables the Huge Job Emulation functionality of the DOCA RegEx instance, allowing it to find matches in data that exceeds the maximum job length of a particular RegEx device. For example, the BlueField RXP hardware device has a maximum job size of 16KB.
This function is provided with a size parameter that indicates the size of overlap to use in the Huge Job Emulation algorithm. This algorithm breaks up the incoming job data into fragments. Therefore, the overlap size causes data from the previous fragment to be prepended to the start of the next fragment. As this overlap impacts performance (job data may get searched multiple times) the overlap size should be kept to a minimum value that still guarantees that matches are found.
doca_error_t doca_regex_property_huge_job_emulation_overlap_set (struct doca_regex *regex, uint16_t nb_overlap_bytes);
Where:
regex [in]– the DOCA RegEx instancenb_overlap_bytes [in]– the number of bytes for the overlap functionality to use
This function returns DOCA_SUCCESS or the relevant DOCA error code.
4.8. Programming RegEx
As part of initialization, the RegEx devices must be programmed with compiled regular expressions. This compilation process takes place offline and generates a compiled file that can be given to a selected device.
4.8.1. doca_regex_property_hardware_binary_rules_set
This function programs the registered hardware RegEx device using rules that are already loaded into memory as pointers to arrays of bytes.
int doca_regex_property_hardware_binary_rules_set(struct doca_regex *regex,
uint8_t const *rules_buffer,
uint32_t rules_buffer_len) ;
Where:
regex [in]– the DOCA RegEx instancerules_buffer [in]– a pointer to a buffer of pre-compiled binary rules data suitable for use by the selected hardware devicerules_buffer_len [in]– the size, in bytes, of the hardware-specific pre-compiled binary rules data
The function returns 0 on success of writing at least one rule to a device and a negative status code on failure.
4.9. Executing Jobs and Receiving Matches
The DOCA RegEx API provides an asynchronous method for enqueuing job data and dequeuing detected matches.
4.9.1. doca_regex_enqueue
This function enqueues a job to the DOCA RegEx instance.
int doca_regex_enqueue(struct doca_regex *regex, uint16_t qid,
struct doca_regex_job_request const *job,
bool allow_aggregation);
Where:
regex [in]– the DOCA RegEx instanceqid [in]– the ID of the queue in which to enqueue the jobjob [in]– a DOCA RegEx job to be enqueued. The caller retains ownership of the data.allow_aggregation [in]– when set, the RegEx device may choose to not begin processing this job immediately to maximise overall efficiency and throughput. When not set, the RegEx engine must begin processing immediately, potentially reducing latency.This allows an application to favor either throughput or latency. If in doubt, it is recommended to favor throughput.
The function returns:
- 0 – device busy, wait until one or more results are dequeued before enqueuing more jobs
- 1 – job enqueued successfully
- Negative POSIX status code upon failure
4.9.2. doca_regex_dequeue
This function dequeues any matches from a previously enqueued job.
int doca_regex_dequeue(struct doca_regex *regex, uint16_t qid,
struct doca_regex_job_response *responses,
uint8_t max_results);
Where:
regex [in]– the DOCA RegEx instanceqid [in]– the ID of the queue in which to dequeue the results dataresponses [out]– response structuresmax_results [in]– maximum number of results to return. The responses array must have capacity for at least this many elements.
The function returns 0 or a positive integer representing the number of results dequeued, or a negative status code on failure.
After completing the processing of any RegEx matches, you must return each one of them to the mempool. These matches are present in the matches_mempool field off the doca_regex_job_response when dequeuing a result. See doca_regex_mempool_obj_put in the NVIDIA DOCA Libraries API Reference Manual for more information.
Notice
This document is provided for information purposes only and shall not be regarded as a warranty of a certain functionality, condition, or quality of a product. NVIDIA Corporation nor any of its direct or indirect subsidiaries and affiliates (collectively: “NVIDIA”) make no representations or warranties, expressed or implied, as to the accuracy or completeness of the information contained in this document and assume no responsibility for any errors contained herein. NVIDIA shall have no liability for the consequences or use of such information or for any infringement of patents or other rights of third parties that may result from its use. This document is not a commitment to develop, release, or deliver any Material (defined below), code, or functionality.
NVIDIA reserves the right to make corrections, modifications, enhancements, improvements, and any other changes to this document, at any time without notice.
Customer should obtain the latest relevant information before placing orders and should verify that such information is current and complete.
NVIDIA products are sold subject to the NVIDIA standard terms and conditions of sale supplied at the time of order acknowledgement, unless otherwise agreed in an individual sales agreement signed by authorized representatives of NVIDIA and customer (“Terms of Sale”). NVIDIA hereby expressly objects to applying any customer general terms and conditions with regards to the purchase of the NVIDIA product referenced in this document. No contractual obligations are formed either directly or indirectly by this document.
NVIDIA products are not designed, authorized, or warranted to be suitable for use in medical, military, aircraft, space, or life support equipment, nor in applications where failure or malfunction of the NVIDIA product can reasonably be expected to result in personal injury, death, or property or environmental damage. NVIDIA accepts no liability for inclusion and/or use of NVIDIA products in such equipment or applications and therefore such inclusion and/or use is at customer’s own risk.
NVIDIA makes no representation or warranty that products based on this document will be suitable for any specified use. Testing of all parameters of each product is not necessarily performed by NVIDIA. It is customer’s sole responsibility to evaluate and determine the applicability of any information contained in this document, ensure the product is suitable and fit for the application planned by customer, and perform the necessary testing for the application in order to avoid a default of the application or the product. Weaknesses in customer’s product designs may affect the quality and reliability of the NVIDIA product and may result in additional or different conditions and/or requirements beyond those contained in this document. NVIDIA accepts no liability related to any default, damage, costs, or problem which may be based on or attributable to: (i) the use of the NVIDIA product in any manner that is contrary to this document or (ii) customer product designs.
No license, either expressed or implied, is granted under any NVIDIA patent right, copyright, or other NVIDIA intellectual property right under this document. Information published by NVIDIA regarding third-party products or services does not constitute a license from NVIDIA to use such products or services or a warranty or endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual property rights of the third party, or a license from NVIDIA under the patents or other intellectual property rights of NVIDIA.
Reproduction of information in this document is permissible only if approved in advance by NVIDIA in writing, reproduced without alteration and in full compliance with all applicable export laws and regulations, and accompanied by all associated conditions, limitations, and notices.
THIS DOCUMENT AND ALL NVIDIA DESIGN SPECIFICATIONS, REFERENCE BOARDS, FILES, DRAWINGS, DIAGNOSTICS, LISTS, AND OTHER DOCUMENTS (TOGETHER AND SEPARATELY, “MATERIALS”) ARE BEING PROVIDED “AS IS.” NVIDIA MAKES NO WARRANTIES, EXPRESSED, IMPLIED, STATUTORY, OR OTHERWISE WITH RESPECT TO THE MATERIALS, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE. TO THE EXTENT NOT PROHIBITED BY LAW, IN NO EVENT WILL NVIDIA BE LIABLE FOR ANY DAMAGES, INCLUDING WITHOUT LIMITATION ANY DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE, OR CONSEQUENTIAL DAMAGES, HOWEVER CAUSED AND REGARDLESS OF THE THEORY OF LIABILITY, ARISING OUT OF ANY USE OF THIS DOCUMENT, EVEN IF NVIDIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Notwithstanding any damages that customer might incur for any reason whatsoever, NVIDIA’s aggregate and cumulative liability towards customer for the products described herein shall be limited in accordance with the Terms of Sale for the product.
Trademarks
NVIDIA, the NVIDIA logo, and Mellanox are trademarks and/or registered trademarks of Mellanox Technologies Ltd. and/or NVIDIA Corporation in the U.S. and in other countries. The registered trademark Linux® is used pursuant to a sublicense from the Linux Foundation, the exclusive licensee of Linus Torvalds, owner of the mark on a world¬wide basis. Other company and product names may be trademarks of the respective companies with which they are associated.
Copyright
© 2022 NVIDIA Corporation & affiliates. All rights reserved.