- 01 Feb, 2024 40 commits
-
-
David Arinzon authored
This patch reduces some of the lines by removing newlines where more variables or print strings can be pushed back to the previous line while still adhering to the styling guidelines. Signed-off-by:
David Arinzon <darinzon@amazon.com> Signed-off-by:
Paolo Abeni <pabeni@redhat.com>
-
David Arinzon authored
Fail queue size calculation when the device returns maximum TX/RX queue sizes that are smaller than the allowed minimum. Signed-off-by:
Osama Abboud <osamaabb@amazon.com> Signed-off-by:
-
David Arinzon authored
The netif_* functions are used by the driver to log events into the kernel ring (dmesg) similar to the netdev_* ones. Unlike the latter, the netif_* function family allow the user to choose what events get logged using ethtool: sudo ethtool -s [interface] msglvl [msg_type] on By default the events which get logged are slow-path related and aren't printed often (e.g. interface up related prints). This patch removes the NETIF_MSG_TX_DONE type (called every TX completion polling) from the defaults and adds NETIF_MSG_IFDOWN instead as it makes more sensible defaults. This patch also transforms ena_down() print from netif_info into netif_dbg (same as the analogue print in ena_up()) as it suits it better. Signed-off-by:
Shay Agroskin <shayagr@amazon.com> Signed-off-by:
-
David Arinzon authored
Move skb_tx_timestamp() closer to the actual time the driver sends the packets to the device. Signed-off-by:
-
David Arinzon authored
The function responsible for polling TX completions might not receive the CPU resources it needs due to higher priority tasks running on the requested core. The driver might not be able to recognize such cases, but it can use its state to suspect that they happened. If both conditions are met: - napi hasn't been executed more than the TX completion timeout value - napi is scheduled (meaning that we've received an interrupt) Then it's more likely that the napi handler isn't scheduled because of an overloaded CPU. It was decided that for this case, the driver would wait twice as long as the regular timeout before scheduling a reset. The driver uses ENA_REGS_RESET_SUSPECTED_POLL_STARVATION reset reason to indicate this case to the device. This patch also adds more information to the ena_tx_timeout() callback. This function is called by the kernel when it detects that a specific TX queue has been closed for too long. Signed-off-by:
-
David Arinzon authored
The print was re-worded to a more informative one. Signed-off-by:
Shahar Itzko <itzko@amazon.com> Signed-off-by:
-
David Arinzon authored
The functionality was added to allow the drivers to create an SQ and CQ of different sizes. When the RX/TX SQ and CQ have the same size, such update isn't necessary as the device can safely assume it doesn't override unprocessed completions. However, if the SQ is larger than the CQ, the device might "have" more completions it wants to update about than there's room in the CQ. There's no support for different SQ and CQ sizes, therefore, removing the API and its usage. '____cacheline_aligned' compiler attribute was added to 'struct ena_com_io_cq' to ensure that the removal of the 'cq_head_db_reg' field doesn't change the cache-line layout of this struct. Signed-off-by:
-
David Arinzon authored
Dynamic Interrupt Moderation (DIM) is a technique designed to balance the need for timely data processing with the desire to minimize CPU overhead. Instead of generating an interrupt for every received packet, the system can dynamically adjust the rate at which interrupts are generated based on the incoming traffic patterns. Enabling DIM by default to improve the user experience. DIM can be turned on/off through ethtool: `ethtool -C <interface> adaptive-rx <on/off>` Signed-off-by:
Arthur Kiyanovski <akiyano@amazon.com> Signed-off-by:
-
David Arinzon authored
A few changes for better readability and style 1. Adding / Removing newlines 2. Removing an unnecessary and confusing comment 3. Using an existing variable rather than re-checking a field Signed-off-by:
-
David Arinzon authored
This patch contains more details about the functionality of RX copybreak. Signed-off-by:
-
David Arinzon authored
Remove io_sq->header_addr field because it is no longer in use. LLQ was updated to support a bounce buffer so there is no need in saving the header address of the sq. Signed-off-by:
Nati Koler <nkoler@amazon.com> Signed-off-by:
-
Paolo Abeni authored
Souradeep Chakrabarti says: ==================== net: mana: Assigning IRQ affinity on HT cores This patch set introduces a new helper function irq_setup(), to optimize IRQ distribution for MANA network devices. The patch set makes the driver working 15% faster than with cpumask_local_spread(). ==================== Link: https://lore.kernel.org/r/1706509267-17754-1-git-send-email-schakrabarti@linux.microsoft.comSigned-off-by:
-
Souradeep Chakrabarti authored
Existing MANA design assigns IRQ to every CPU, including sibling hyper-threads. This may cause multiple IRQs to be active simultaneously in the same core and may reduce the network performance. Improve the performance by assigning IRQ to non sibling CPUs in local NUMA node. The performance improvement we are getting using ntttcp with following patch is around 15 percent against existing design and approximately 11 percent, when trying to assign one IRQ in each core across NUMA nodes, if enough cores are present. The change will improve the performance for the system with high number of CPU, where number of CPUs in a node is more than 64 CPUs. Nodes with 64 CPUs or less than 64 CPUs will not be affected by this change. The performance study was done using ntttcp tool in Azure. The node had 2 nodes with 32 cores each, total 128 vCPU and number of channels were 32 for 32 RX rings. The below table shows a comparison between existing design and new design: IRQ node-num core-num CPU performance(%) 1 0 | 0 0 | 0 0 | 0-1 0 2 0 | 0 0 | 1 1 | 2-3 3 3 0 | 0 1 | 2 2 | 4-5 10 4 0 | 0 1 | 3 3 | 6-7 15 5 0 | 0 2 | 4 4 | 8-9 15 ... ... 25 0 | 0 12| 24 24| 48-49 12 ... 32 0 | 0 15| 31 31| 62-63 12 33 0 | 0 16| 0 32| 0-1 10 ... 64 0 | 0 31| 31 63| 62-63 0 Signed-off-by:
Souradeep Chakrabarti <schakrabarti@linux.microsoft.com> Signed-off-by:
-
Yury Norov authored
Souradeep investigated that the driver performs faster if IRQs are spread on CPUs with the following heuristics: 1. No more than one IRQ per CPU, if possible; 2. NUMA locality is the second priority; 3. Sibling dislocality is the last priority. Let's consider this topology: Node 0 1 Core 0 1 2 3 CPU 0 1 2 3 4 5 6 7 The most performant IRQ distribution based on the above topology and heuristics may look like this: IRQ Nodes Cores CPUs 0 1 0 0-1 1 1 1 2-3 2 1 0 0-1 3 1 1 2-3 4 2 2 4-5 5 2 3 6-7 6 2 2 4-5 7 2 3 6-7 The irq_setup() routine introduced in this patch leverages the for_each_numa_hop_mask() iterator and assigns IRQs to sibling groups as described above. According to [1], for NUMA-aware but sibling-ignorant IRQ distribution based on cpumask_local_spread() performance test results look like this: ./ntttcp -r -m 16 NTTTCP for Linux 1.4.0 --------------------------------------------------------- 08:05:20 INFO: 17 threads created 08:05:28 INFO: Network activity progressing... 08:06:28 INFO: Test run completed. 08:06:28 INFO: Test cycle finished. 08:06:28 INFO: ##### Totals: ##### 08:06:28 INFO: test duration :60.00 seconds 08:06:28 INFO: total bytes :630292053310 08:06:28 INFO: throughput :84.04Gbps 08:06:28 INFO: retrans segs :4 08:06:28 INFO: cpu cores :192 08:06:28 INFO: cpu speed :3799.725MHz 08:06:28 INFO: user :0.05% 08:06:28 INFO: system :1.60% 08:06:28 INFO: idle :96.41% 08:06:28 INFO: iowait :0.00% 08:06:28 INFO: softirq :1.94% 08:06:28 INFO: cycles/byte :2.50 08:06:28 INFO: cpu busy (all) :534.41% For NUMA- and sibling-aware IRQ distribution, the same test works 15% faster: ./ntttcp -r -m 16 NTTTCP for Linux 1.4.0 --------------------------------------------------------- 08:08:51 INFO: 17 threads created 08:08:56 INFO: Network activity progressing... 08:09:56 INFO: Test run completed. 08:09:56 INFO: Test cycle finished. 08:09:56 INFO: ##### Totals: ##### 08:09:56 INFO: test duration :60.00 seconds 08:09:56 INFO: total bytes :741966608384 08:09:56 INFO: throughput :98.93Gbps 08:09:56 INFO: retrans segs :6 08:09:56 INFO: cpu cores :192 08:09:56 INFO: cpu speed :3799.791MHz 08:09:56 INFO: user :0.06% 08:09:56 INFO: system :1.81% 08:09:56 INFO: idle :96.18% 08:09:56 INFO: iowait :0.00% 08:09:56 INFO: softirq :1.95% 08:09:56 INFO: cycles/byte :2.25 08:09:56 INFO: cpu busy (all) :569.22% [1] https://lore.kernel.org/all/20231211063726.GA4977@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net/Signed-off-by:
Yury Norov <yury.norov@gmail.com> Co-developed-by:
-
Yury Norov authored
Now we can simplify code that allocates cpumasks for local needs. Signed-off-by:
-
Yury Norov authored
Similarly to cpumask_weight_and(), cpumask_weight_andnot() is a handy helper that may help to avoid creating an intermediate mask just to calculate number of bits that set in a 1st given mask, and clear in 2nd one. Signed-off-by:
Jacob Keller <jacob.e.keller@intel.com> Signed-off-by:
-
Philippe Schenker authored
This commit introduces support for the KSZ8567, a robust 7-port Ethernet switch. The KSZ8567 features two RGMII/MII/RMII interfaces, each capable of gigabit speeds, complemented by five 10/100 Mbps MAC/PHYs. Signed-off-by:
Philippe Schenker <philippe.schenker@impulsing.ch> Acked-by:
Arun Ramadoss <arun.ramadoss@microchip.com> Reviewed-by:
Andrew Lunn <andrew@lunn.ch> Reviewed-by:
Florian Fainelli <florian.fainelli@broadcom.com> Link: https://lore.kernel.org/r/20240130083419.135763-2-dev@pschenker.chSigned-off-by:
-
Philippe Schenker authored
This commit adds the dt-binding for KSZ8567, a robust 7-port Ethernet switch. The KSZ8567 features two RGMII/MII/RMII interfaces, each capable of gigabit speeds, complemented by five 10/100 Mbps MAC/PHYs. This binding is necessary to set specific capabilities for this switch chip that are necessary due to the ksz dsa driver only accepting specific chip ids. The KSZ8567 is very similar to KSZ9567 however only containing 100 Mbps phys on its downstream ports. Signed-off-by:
Conor Dooley <conor.dooley@microchip.com> Reviewed-by:
-
Krzysztof Kozlowski authored
dtschema package defines firmware-name as string-array, so individual bindings should not make it a string but instead just narrow the number of expected firmware file names. Signed-off-by:
Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Acked-by:
Alex Elder <elder@linaro.org> Link: https://lore.kernel.org/r/20240129142121.102450-1-krzysztof.kozlowski@linaro.orgSigned-off-by:
-
Jakub Kicinski authored
Donald Hunter says: ==================== tools/net/ynl: Add features for tc family Add features to ynl for tc and update the tc spec to use them. Patch 1 adds an option to output json instead of python pretty printing. Patch 2, 3 adds support and docs for sub-messages in nested attribute spaces that reference keys from a parent space. Patches 4 and 7-9 refactor ynl in support of nested struct definitions Patch 5 implements sub-message encoding for write ops. Patch 6 adds logic to set default zero values for binary blobs Patches 10, 11 adds support and docs for nested struct definitions Patch 12 updates the ynl doc generator to include type information for struct members. Patch 13 updates the tc spec - still a work in progress but more complete ==================== Link: https://lore.kernel.org/r/20240129223458.52046-1-donald.hunter@gmail.comSigned-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Donald Hunter authored
Fill in many of the gaps in the tc netlink spec, including stats attrs, classes and actions. Many documentation strings have also been added. This is still a work in progress, albeit fairly complete: - there are still many attributes left as binary blobs. - actions have not had much testing Signed-off-by:
Donald Hunter <donald.hunter@gmail.com> Reviewed-by:
Jiri Pirko <jiri@nvidia.com> Link: https://lore.kernel.org/r/20240129223458.52046-14-donald.hunter@gmail.comSigned-off-by:
-
Donald Hunter authored
Extend the ynl doc generator to include type information for struct members, ignoring the pad type. Signed-off-by:
-
Donald Hunter authored
Add a description and example of nested struct definitions to the netlink raw documentation. Signed-off-by:
-
Donald Hunter authored
Make it possible for struct definitions to reference other struct definitions ofr binary members. For example, the tbf qdisc uses this struct definition for its parms attribute: - name: tc-tbf-qopt type: struct members: - name: rate type: binary struct: tc-ratespec - name: peakrate type: binary struct: tc-ratespec - name: limit type: u32 - name: buffer type: u32 - name: mtu type: u32 This adds the necessary schema changes and adds nested struct encoding and decoding to ynl. Signed-off-by: -
Donald Hunter authored
The formatted_string() class method was in NlAttr so that it could be accessed by NlAttr.as_struct(). Now that as_struct() has been removed, move formatted_string() to YnlFamily as an internal helper method. Signed-off-by:
Breno Leitao <leitao@debian.org> Reviewed-by:
-
Donald Hunter authored
Refactor the _fixed_header_size() method to be _struct_size() so that naming is consistent with _encode_struct() and _decode_struct(). Signed-off-by:
-
Donald Hunter authored
_decode_fixed_header() and NlAttr.as_struct() both implemented struct decoding logic. Deduplicate the code into newly named _decode_struct() method. Signed-off-by:
-
Donald Hunter authored
Add support for defaulting binary byte arrays to all zeros as well as defaulting scalar values to 0 when encoding input parameters. Signed-off-by:
-
Donald Hunter authored
Add sub-message encoding to ynl. This makes it possible to create tc qdiscs and other polymorphic netlink objects. Signed-off-by:
-
Donald Hunter authored
Refactor the fixed header encoding into a separate _encode_struct method so that it can be reused for fixed headers in sub-messages and for encoding structs. Signed-off-by:
-
Donald Hunter authored
Update the netlink-raw docs to add a description of sub-message selector resolution to explain that selector resolution is constrained by the spec. Signed-off-by:
-
Donald Hunter authored
Sub-message selectors could only be resolved using values from the current nest level. Enable value lookup in outer scopes by using collections.ChainMap to implement an ordered lookup from nested to outer scopes. Signed-off-by:
-
Donald Hunter authored
The ynl cli currently emits python pretty printed structures which is hard to consume. Add a new --output-json argument to emit JSON. Signed-off-by:
-
David Ahern authored
Running fcnal-test.sh script with -P argument is causing test failures: $ ./fcnal-test.sh -t ping -P TEST: ping out - ns-B IP [ OK ] hit enter to continue, 'q' to quit fcnal-test.sh: line 106: [: ping: integer expression expected TEST: out, [FAIL] expected rc ping; actual rc 0 hit enter to continue, 'q' to quit The test functions use local variable 'a' for addresses and then log_test is also using 'a' without a local declaration. Fix by declaring a local variable and using 'ans' (for answer) in the read. Signed-off-by:David Ahern <dsahern@kernel.org> Link: https://lore.kernel.org/r/20240130154327.33848-1-dsahern@kernel.orgSigned-off-by:
-
Jakub Kicinski authored
Kuniyuki Iwashima says: ==================== af_unix: Remove io_uring dead code in GC. I will post another series that rewrites the garbage collector for AF_UNIX socket. This is a prep series to clean up changes to GC made by io_uring but now not necessary. ==================== Link: https://lore.kernel.org/r/20240129190435.57228-1-kuniyu@amazon.comSigned-off-by:
-
Kuniyuki Iwashima authored
Originally, the code related to garbage collection was all in garbage.c. Commit f4e65870 ("net: split out functions related to registering inflight socket files") moved some functions to scm.c for io_uring and added CONFIG_UNIX_SCM just in case AF_UNIX was built as module. However, since commit 97154bcf ("af_unix: Kconfig: make CONFIG_UNIX bool"), AF_UNIX is no longer built separately. Also, io_uring does not support SCM_RIGHTS now. Let's move the functions back to garbage.c Signed-off-by:
Kuniyuki Iwashima <kuniyu@amazon.com> Acked-by:
Jens Axboe <axboe@kernel.dk> Link: https://lore.kernel.org/r/20240129190435.57228-4-kuniyu@amazon.comSigned-off-by:
-
Kuniyuki Iwashima authored
Since commit 705318a9 ("io_uring/af_unix: disable sending io_uring over sockets"), io_uring's unix socket cannot be passed via SCM_RIGHTS, so it does not contribute to cyclic reference and no longer be candidate for garbage collection. Also, commit 6e5e6d27 ("io_uring: drop any code related to SCM_RIGHTS") cleaned up SCM_RIGHTS code in io_uring. Let's do it in AF_UNIX as well by reverting commit 0091bfc8 ("io_uring/af_unix: defer registered files gc to io_uring release") and commit 10369080 ("net: reclaim skb->scm_io_uring bit"). Signed-off-by:
-
Kuniyuki Iwashima authored
This is a prep patch for the last patch in this series so that checkpatch will not warn about BUG_ON(). Signed-off-by:
-
Kunwu Chan authored
commit 0a31bd5f ("KMEM_CACHE(): simplify slab cache creation") introduces a new macro. Use the new KMEM_CACHE() macro instead of direct kmem_cache_create to simplify the creation of SLAB caches. Signed-off-by:
Kunwu Chan <chentao@kylinos.cn> Acked-by:
Nikolay Aleksandrov <razor@blackwall.org> Reviewed-by:
-
Kunwu Chan authored
commit 0a31bd5f ("KMEM_CACHE(): simplify slab cache creation") introduces a new macro. Use the new KMEM_CACHE() macro instead of direct kmem_cache_create to simplify the creation of SLAB caches. Signed-off-by:
-
