- 27 Mar, 2019 21 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Jeff Kirsher says: ==================== 100GbE Intel Wired LAN Driver Updates 2019-03-26 This series contains more updates to the ice driver only. Jeremiah provides his first patch to the Linux kernel to clean up un-necessary newlines in driver log messages. Mitch updates the ice driver to use existing status codes in the iavf driver so that when errors occur, it will not report nonsensical results. Adds support for VF admin queue interrupts by programming the VPINT_MBX_CTL register array. Brett adds a check for a bit that we set while preparing for a reset, to ensure we are prepared to do a proper reset. Also implemented PCI error handling operations. Went through and audited the hot path with pahole and made modifications based on the results since 2 structures were taking up more space than necessary due to cache alignment issues. Fixed an issue where when flow control was disabled, the state of flow control was being displayed as "Unknown". Anirudh fixes adaptive interrupt moderation changes by adding code that was missed, that should have been added in the initial patch to add that support. Cleaned up a function prototype that was never implemented. Did additional code cleanup by removing unneeded braces and redundant code comments. Akeem fixes an issue that occurs when the VF is attempting to remove the default LAN/MAC address, which is programmed by the administrator by updating the error message to explicitly say that the VF cannot change the MAC programmed by the PF. Preethi fixes the driver to not fall into the error path when a added filter already exists, but instead continue to process the rest of the function and add appropriate checks after adding MAC filters. ==================== Signed-off-by:David S. Miller <davem@davemloft.net>
-
David S. Miller authored
Maxime Chevallier says: ==================== net: mvpp2: Classifier updates and cleanups In preparation for the future addition of classification offload support, this series cleans-up the current code dealing with the classification engines. As of today, the classification code only deals with RSS, which is a very narrow use-case when considering all of the classifier's features. This lead to design and naming decisions that don't really stand when considering using more classification features. This series therefore includes quite a lot a renaming of functions and macros, and tries to make the naming schemes more consistent and the code more readable. The debugfs interface that allows reading the various Parsing and Classification engines has been cleaned-up and made more generic, allowing to read the Flow Table and C2 engine hit counters on a per-entry basis. The Classifier itself has been made more robust by introducing the lu_type field in classification lookups, that prevents false-positive matches in the future. We also initialise the various engine in a more extensive and less error-prone way by assining default values to all entries in the C2 and Flow table. This is a pretty big series considering it's mostly cleanup, but my goal is to make the future series that will contain new big features easier to review, focusing on the real logic. Besides the debugfs interface, this series doesn't intend to introduce any new features or change in behaviours. ==================== Signed-off-by: -
Maxime Chevallier authored
The C2 classification engine has a 256 entry TCAM, used for ternary matches on an 8 byte Header Extracted Key. For now, we compute the various indices for classification and RSS that use this engine thanks to a set of macros. This commit mainly renames the macros used to make it clear that they should be used with the C2 engine, but also make use of the full 256 entries in the engine. For now, the C2 entries are only used for RSS. These entries are put at the end of the TCAM range, in case we want to add higher priority matches later on. Signed-off-by:
Maxime Chevallier <maxime.chevallier@bootlin.com> Signed-off-by:
-
Maxime Chevallier authored
When classifying a packet pertaining to a given flow, the classifier will issue multiple lookup commands until it finds one with the 'last' bit set. It expects all prorities to be assign continuously (although not necessarily in an ordered fashion) from 0 to the number of lookups. We can initialize this once, and make sure unused lookups are given an empty port map. This avoids having to maintain priorities and the information of which lookup is the last. Signed-off-by:
-
Maxime Chevallier authored
C2 TCAM entries can be invalidated to avoid unwanted matches. Make sure all entries are invalidated at init, then validate only the ones we use. Signed-off-by:
-
Maxime Chevallier authored
The Flow Table dictates what lookups will be issued for each flow type. The lookup sequence for each flow is similar, and the index of each lookup is computed by some macros. There are similar mechanisms for the C2 TCAM lookups, so in order to avoid confusion, rename the flow table index computing macros with a common prefix. The only difference in behaviour is that we now use the very first entry in the flow for the RSS lookup (the first entry was previously unused). Signed-off-by:
-
Maxime Chevallier authored
The classifier allows to combine multiple lookups in one "sequence" that is counted as a single lookup to an engine, with a single result. We don't actually use that feature, so remove any places where we set this field, so that the classifier doesn't try to interpret these fields. Signed-off-by:
-
Maxime Chevallier authored
This commit renames some of the classifier functions to follow the naming 'mvpp2_port_*' that's used for function that act on a given port. This commit is purely cosmetic. Signed-off-by:
-
Maxime Chevallier authored
Move C2 read/write helpers higher in the file to ease future work that rely on these helpers Signed-off-by:
-
Maxime Chevallier authored
When writing a C2 entry to hardware, some registers writes will only take effect when the TCAM_DATA4 register is written. This includes all C2 TCAM registers, and the C2 invalidate register. To make sure we always write C2 entries correctly, document that behaviour with a comment, and move TCAM writes to the end of the mvpp2_cls_c2_write helper. Signed-off-by:
-
Maxime Chevallier authored
The cls_table is a global read-only table containing the different parameters that are used by various tables in the classifier. It describes the links between the Header Parser, the decoding table and the flow_table. There are several possible way we want to iterate over that table, depending on wich classifier engine we want to configure. For the Header Parser, we want to iterate over each entry. For the Decoding table, we want to iterate over each entry having a unique flow_id. Finally, when configuring an ethtool flow, we want to iterate over each entry having a unique flow_id and that has a given flow_type. This commit introduces some iterator to both provide syntactic sugar and also clarify the way we want to iterate over the table. Signed-off-by:
-
Maxime Chevallier authored
PPv2's Classifier uses multiple engines to perform classification. So far, only the C2 engine is used, which has a 256 entries TCAM. So far, we only accessed the relevant entries from the C2 engines, which are the one implementing RSS. To implement and debug ntuple classification offload, beaing able to see the hit count for each C2 entry is helpful, so this commit moves the logic to a dedicated directory allowing to access each entry. Signed-off-by:
-
Maxime Chevallier authored
The Classifier flow table is the central part of the PPv2 Classifier, since it describes all classification steps performed for each flow. It has 512 entries, shared between all ports, which are divided into sequences that are pointed-to by the decoding table. Being able to see which entries in the flow table were hit is a key point when implementing and debugging classification offload. This commit allows reading each flow table entry's hit count independently, with a clear-on-read behaviour. Signed-off-by:
-
Maxime Chevallier authored
The current way to store the required private data needed to access various debugfs entries is to alloc them on the fly, share them within the entries that need to access them, and finally have one entry free that data upon closing. This leads to hard to maintain code, and is very error-prone. This commit stores all debugfs related data in the same place, making sure this is allocated only when the debugfs directory is successfully created, so that we don't waste memory when we don't use this feature. Signed-off-by:
-
Maxime Chevallier authored
The cls_flow table represent the overall configuration of the classifier, used to match the different traffic classes in the Parsing and Classification engines. This configuration is static, and applies to all PPv2 instances, we must therefore keep it const so that no modifications of this table are performed at runtime. Signed-off-by:
-
Maxime Chevallier authored
The macro definition MVPP2_N_FLOWS is ambiguous because it really represents the number of entries in the Header Parser that are used to identify the classification flows. Rename the macro to clearly state that we represent the number of flows in the Header Parser. Signed-off-by:
-
Maxime Chevallier authored
The PPv2 classifier allows to perform multiple lookups on the same engine when classifying a packet. These lookups can match similar parts of a packet header, but perform different actions upon matching. To differentiate these types of lookups, it's possible to specify a Lookup Type in the flow table entries, which will be part of the key for the lookup engines. This commit introduces the use of Lookup Types for C2 matches. Since for now we only perform C2 lookups to enable RSS, we only need one Lookup Type. Signed-off-by:
-
Maxime Chevallier authored
The Classifier flow table has 512 entries, that contains lookups commands executed consecutively for every flow. Since we have 21 different flows, we have to carefully manage the flow table use. As of today, the start index of a lookup sequence is computed directly based in the flow->id. There are 8 reserved flow ids, from 0-7, which don't have any corresponding sequence in the flow table. We can therefore ignore them when computing the index, and make so that the first non-reserved flow point to the very beginning of the flow table. Signed-off-by:
Alan Winkowski <walan@marvell.com> Signed-off-by:
-
Maxime Chevallier authored
PPv2's classifier supports extracting the MAC Destination Address from the L2 header to perform RSS and flow steering. Add the missing case when setting the Header Extracted Key fields in the flow table. Signed-off-by:
-
Maxime Chevallier authored
int is not long enough to store all netdev_features, use the correct dedicated type to store them when building the list of dev->features. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-nextDavid S. Miller authored
Alexei Starovoitov says: ==================== pull-request: bpf-next 2019-03-26 The following pull-request contains BPF updates for your *net-next* tree. The main changes are: 1) introduce bpf_tcp_check_syncookie() helper for XDP and tc, from Lorenz. 2) allow bpf_skb_ecn_set_ce() in tc, from Peter. 3) numerous bpf tc tunneling improvements, from Willem. 4) and other miscellaneous improvements from Adrian, Alan, Daniel, Ivan, Stanislav. ==================== Signed-off-by:
-
- 26 Mar, 2019 19 commits
-
-
Anirudh Venkataramanan authored
Some enums in ice_tx_desc_cmd_bits have a trailing /* 2 BITS */ comment, but the value has just one bit set (ex. ICE_TX_DESC_CMD_L4T_EOFT_SCTP has the value 0x200 (i.e. only bit 9 is set). This is confusing and misleading. So remove the comment. Signed-off-by:
Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by:
Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com>
-
Brett Creeley authored
Since the driver now hard codes the ITR granularity to 2 us in the GLINT_CTL register the comment next to ITR_GRAN_S needs to be updated. Signed-off-by:
Brett Creeley <brett.creeley@intel.com> Signed-off-by:
-
Anirudh Venkataramanan authored
Remove some redundant text in the function header for __ice_vsi_get_qs Signed-off-by:
-
Anirudh Venkataramanan authored
Single statement if conditions don't need braces. Remove it. Signed-off-by:
-
Anirudh Venkataramanan authored
Commit 37bb8390 ("ice: Move common functions out of ice_main.c part 7/7") seems to have inadvertently introduced a function prototype for ice_vsi_cfg_tc without a corresponding function implementation. Remove it. Signed-off-by:
-
Brett Creeley authored
Currently we aren't checking for the ICE_FC_NONE case for the current flow control mode. This is causing "Unknown" to be printed for the current flow control method if flow control is disabled. Fix this by adding the case for ICE_FC_NONE to print "None". Signed-off-by:
-
Brett Creeley authored
Currently the ice_q_vector structure and ice_ring_container structure are taking up more space than necessary due to cache alignment holes and unnecessary variables respectively. This is not helping the driver's performance. The following fixes were done to improve cache alignment, reduce wasted space, and increase performance. 1. Remove the ice_latency_range enum as it is unused. 2. Remove the latency_range variable in the ice_ring_container structure. 3. Change the size of the itr_idx in the ice_ring_container structure from an int to an u16. This reduced the size of ice_ring_container structure to 32 Bytes so it has no holes or padding. 4. Re-arrange the ice_q_vector structure using pahole to align members as best as possible in regards to 64 Byte cache line size. Signed-off-by:
-
Preethi Banala authored
If filter already exists, do not go through error path flow but instead continue to process rest of the function. Hence have an appropriate check after adding MAC filters. Signed-off-by:
Preethi Banala <preethi.banala@intel.com> Signed-off-by:
-
Akeem G Abodunrin authored
This patch fixes issue that occurs when VF is attempting to remove default LAN/MAC address, which is programmed by the administrator. We shouldn't return error for the call by the VF, but continue with the remaining steps to handle MAC opcode. Also update the dev_err message to explicitly say that VF can't change MAC programmed by PF. Also change "mac" to "MAC" for kernel print statements in the same file. Signed-off-by:
Akeem G Abodunrin <akeem.g.abodunrin@intel.com> Signed-off-by:
-
Mitch Williams authored
The VPINT_MBX_CTL register array must be programmed to enable VF admin queue interrupts. Without this, VFs never get interrupts on vector 0, and some VF drivers will fail to init. Signed-off-by:
Mitch Williams <mitch.a.williams@intel.com> Signed-off-by:
-
Anirudh Venkataramanan authored
commit 63f545ed ("ice: Add support for adaptive interrupt moderation") was meant to add support for adaptive interrupt moderation but there was an error on my part while formatting the patch, and thus only part of the patch ended up being submitted. This patch rectifies the error by adding the rest of the code. Fixes: 63f545ed ("ice: Add support for adaptive interrupt moderation") Signed-off-by:
-
Brett Creeley authored
This patch implements the following pci_error_handler ops: .error_detected = ice_pci_err_detected .slot_reset = ice_pci_err_slot_reset .reset_notify = ice_pci_err_reset_notify .reset_prepare = ice_pci_err_reset_prepare .reset_done = ice_pci_err_reset_done .resume = ice_pci_err_resume Signed-off-by:
-
Brett Creeley authored
Currently we check if the __ICE_PREPARED_FOR_RESET bit is set prior to calling ice_prepare_for_reset in ice_reset_subtask(), but we aren't checking that bit in ice_do_reset() before calling ice_prepare_for_reset(). This is not consistent and can cause issues if ice_prepare_for_reset() is called prior to ice_do_reset(). Fix this by checking if the __ICE_PREPARED_FOR_RESET bit is set internal to ice_prepare_for_reset(). Signed-off-by:
-
Mitch Williams authored
When communicating with the AVF driver, we need to use the status codes from virtchnl.h, not our own ice-specific codes. Without this, when an error occurs, the VF will report nonsensical results. NOTE: this depends on changes made to include/linux/avf/virtchnl.h by commit bb58fd7e ("i40e: Update status codes") Signed-off-by:
-
Jeremiah Kyle authored
Two log messages contained newlines in the middle of the message. This resulted in unexpected driver log output. This patch removes the newlines to restore consistency with the rest of the driver log messages. Signed-off-by:
Jeremiah Kyle <jeremiah.kyle@intel.com> Signed-off-by:
-
Flavio Leitner authored
When the conntrack is initialized, there is no helper attached yet so the nat info initialization (nf_nat_setup_info) skips adding the seqadj ext. A helper is attached later when the conntrack is not confirmed but is going to be committed. In this case, if NAT is needed then adds the seqadj ext as well. Fixes: 16ec3d4f ("openvswitch: Fix cached ct with helper.") Signed-off-by:
Flavio Leitner <fbl@sysclose.org> Acked-by:
Pravin B Shelar <pshelar@ovn.org> Signed-off-by:
-
Stanislav Fomichev authored
When running stacktrace_build_id_nmi, try to query kernel.perf_event_max_sample_rate sysctl and use it as a sample_freq. If there was an error reading sysctl, fallback to 5000. kernel.perf_event_max_sample_rate sysctl can drift and/or can be adjusted by the perf tool, so assuming a fixed number might be problematic on a long running machine. Signed-off-by:
Stanislav Fomichev <sdf@google.com> Signed-off-by:
Alexei Starovoitov <ast@kernel.org>
-
Ioana Ciornei authored
Take advantage of the software Rx batching by using netif_receive_skb_list instead of napi_gro_receive. Signed-off-by:
Ioana Ciornei <ioana.ciornei@nxp.com> Signed-off-by:
-
Wei Yongjun authored
Fix the return value check which testing the wrong variable in tipc_mcast_send_sync(). Fixes: c55c8eda ("tipc: smooth change between replicast and broadcast") Reported-by:
Hulk Robot <hulkci@huawei.com> Signed-off-by:
Wei Yongjun <weiyongjun1@huawei.com> Acked-by:
Jon Maloy <jon.maloy@ericsson.com> Signed-off-by:
-
