- 02 Jul, 2020 4 commits
-
-
Florian Westphal authored
The script generates two random files that are then sent via tcp and mptcp connections. In order to compare throughput over consecutive runs add an option to provide the file size on the command line: "-f 128000". Also add an option, -t, to enable tcp tests. This is useful to compare throughput of mptcp connections and tcp connections. Example: run tests with a 4mb file size, 300ms delay 0.01% loss, default gso/tso/gro settings and with large write/blocking io: mptcp_connect.sh -t -f $((4 * 1024 * 1024)) -d 300 -l 0.01% -r 0 -e "" -m mmap Signed-off-by:
Florian Westphal <fw@strlen.de> Reviewed-by:
Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Danny Lin authored
Similar to fq_codel and the other qdiscs that can set as default, fq_pie is also suitable for general use without explicit configuration, which makes it a valid choice for this. This is useful in situations where a painless out-of-the-box solution for reducing bufferbloat is desired but fq_codel is not necessarily the best choice. For example, fq_pie can be better for DASH streaming, but there could be more cases where it's the better choice of the two simple AQMs available in the kernel. Signed-off-by:
Danny Lin <danny@kdrag0n.dev> Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Tony Nguyen says: ==================== Intel Wired LAN Driver Updates 2020-07-01 This series contains updates to all Intel drivers, but a majority of the changes are to the i40e driver. Jeff converts 'fall through' comments to the 'fallthrough;' keyword for all Intel drivers. Removed unnecessary delay in the ixgbe ethtool diagnostics test. Arkadiusz implements Total Port Shutdown for i40e. This is the revised patch based on Jakub's feedback from an earlier submission of this patch, where additional code comments and description was needed to describe the functionality. Wei Yongjun fixes return error code for iavf_init_get_resources(). Magnus optimizes XDP code in i40e; starting with AF_XDP zero-copy transmit completion path. Then by only executing a division when necessary in the napi_poll data path. Move the check for transmit ring full outside the send loop to increase performance. Ciara add XDP ring statistics to i40e and the ability to dump these statistics and descriptors. Tony fixes reporting iavf statistics. Radoslaw adds support for 2.5 and 5 Gbps by implementing the newer ethtool ksettings API in ixgbe. ==================== Signed-off-by: -
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Tony Nguyen says: ==================== 100GbE Intel Wired LAN Driver Updates 2020-07-01 This series contains updates to the ice driver only. Jacob implements a devlink region for device capabilities. Bruce removes structs containing only one-element arrays that are either unused or only used for indexing. Instead, use pointer arithmetic or other indexing to access the elements. Converts "C struct hack" variable-length types to the preferred C99 flexible array member. ==================== Signed-off-by:
-
- 01 Jul, 2020 36 commits
-
-
Bruce Allan authored
Convert the pre-C90-extension "C struct hack" method (using a single- element array at the end of a structure for implementing variable-length types) to the preferred use of C99 flexible array member. Additional code cleanups were done near areas affected by this change. Signed-off-by:
Bruce Allan <bruce.w.allan@intel.com> Tested-by:
Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by:
Tony Nguyen <anthony.l.nguyen@intel.com>
-
Bruce Allan authored
There are a number of structures that consist of a one-element array as the only struct member. Some of those are unused so remove them. Others are used to index into a buffer/array consisting of a variable number of a different data or structure type. Those are unnecessary since we can use simple pointer arithmetic or index directly into the buffer to access individual elements of the buffer/array. Additional code cleanups were done near areas affected by this change. Signed-off-by:
-
Jarod Wilson authored
At the moment, bonding xfrm crypto offload can only be set up if the bonding module is loaded with active-backup mode already set. We need to be able to make this work with bonds set to AB after the bonding driver has already been loaded. So what's done here is: 1) move #define BOND_XFRM_FEATURES to net/bonding.h so it can be used by both bond_main.c and bond_options.c 2) set BOND_XFRM_FEATURES in bond_dev->hw_features universally, rather than only when loading in AB mode 3) wire up xfrmdev_ops universally too 4) disable BOND_XFRM_FEATURES in bond_dev->features if not AB 5) exit early (non-AB case) from bond_ipsec_offload_ok, to prevent a performance hit from traversing into the underlying drivers 5) toggle BOND_XFRM_FEATURES in bond_dev->wanted_features and call netdev_change_features() from bond_option_mode_set() In my local testing, I can change bonding modes back and forth on the fly, have hardware offload work when I'm in AB, and see no performance penalty to non-AB software encryption, despite having xfrm bits all wired up for all modes now. Fixes: 18cb261a ("bonding: support hardware encryption offload to slaves") Reported-by:
Huy Nguyen <huyn@mellanox.com> CC: Saeed Mahameed <saeedm@mellanox.com> CC: Jay Vosburgh <j.vosburgh@gmail.com> CC: Veaceslav Falico <vfalico@gmail.com> CC: Andy Gospodarek <andy@greyhouse.net> CC: "David S. Miller" <davem@davemloft.net> CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com> CC: Jakub Kicinski <kuba@kernel.org> CC: Steffen Klassert <steffen.klassert@secunet.com> CC: Herbert Xu <herbert@gondor.apana.org.au> CC: netdev@vger.kernel.org CC: intel-wired-lan@lists.osuosl.org Signed-off-by:
Jarod Wilson <jarod@redhat.com> Signed-off-by:
-
Jacob Keller authored
Add a new devlink region used for capturing a snapshot of the device capabilities buffer which is reported by the firmware over the AdminQ. This information can useful in debugging driver and firmware interactions. Signed-off-by:
Jacob Keller <jacob.e.keller@intel.com> Tested-by:
-
David S. Miller authored
Alex Elder says: ==================== net: ipa: endpoint configuration updates This series updates code that configures IPA endpoints. The changes made mainly affect access to registers that are valid only for RX, or only for TX endpoints. The first three patches avoid writing endpoint registers if they are not defined to be valid. The fourth patch slightly modifies the parameters for the offset macros used for these endpoint registers, to make it explicit when only some endpoints are valid. The last patch just tweaks one line of code so it uses a convention used everywhere else in the driver. Version 2 of this series eliminates some of the "assert()" comments that Jakub inquired about. The ones removed will actually go away in an upcoming (not-yet-posted) patch series anyway. ==================== Signed-off-by: -
Alex Elder authored
The convention throughout the IPA driver is to directly use single-bit field mask values, rather than using (for example) u32_encode_bits() to set or clear them. Fix the one place that doesn't follow that convention, which sets HOL_BLOCK_EN_FMASK in ipa_endpoint_init_hol_block_enable(). Signed-off-by:
Alex Elder <elder@linaro.org> Signed-off-by:
-
Alex Elder authored
A handful of registers are valid only for RX endpoints, and some others are valid only for TX endpoints. For these endpoints, add a comment above their defined offset macro that indicates the endpoints to which they apply. Extend the endpoint parameter naming convention as well, to make these constraints more explicit. Signed-off-by:
-
Alex Elder authored
The INIT_MODE endpoint configuration register is only valid for TX endpoints. Rather than writing a zero to that register for RX endpoints, avoid writing the register at all. Signed-off-by:
-
Alex Elder authored
The INIT_HDR_METADATA_MASK endpoint configuration register is only valid for RX endpoints. Rather than writing a zero to that register for TX endpoints, avoid writing the register at all. Signed-off-by:
-
Alex Elder authored
The INIT_HOL_BLOCK_EN and INIT_HOL_BLOCK_TIMER endpoint registers are only valid for RX endpoints. Have ipa_endpoint_modem_hol_block_clear_all() skip writing these registers for TX endpoints. Signed-off-by:
-
David S. Miller authored
Alex Elder says: ==================== net: ipa: small improvements This series contains two patches that improve the error output that's reported when an error occurs while changing the state of a GSI channel or event ring. The first ensures all such error conditions report an error, and the second simplifies the messages a little and ensures they are all consistent. A third (independent) patch gets rid of an unused symbol in the microcontroller code. Version 2 fixes two alignment problems pointed out by checkpatch.pl, as requested by Jakub Kicinski. ==================== Signed-off-by: -
Alex Elder authored
The microcontroller shared memory area is at the beginning of the IPA resident memory. IPA_MEM_UC_OFFSET was defined as the offset within that region where it's found, but it's 0, and it's never actually used. Just get rid of the definition, and move some of the description it had to be above the definition of the ipa_uc_mem_area structure. Signed-off-by:
-
Alex Elder authored
Make minor updates to error messages reported in "gsi.c": - Use local variables to reduce multi-line function calls - Don't use parentheses in messages - Do some slight rewording in a few cases Signed-off-by:
-
Alex Elder authored
We check the state of an event ring or channel both before and after any GSI command issued that will change that state. In most--but not all--cases, if the state is something different than expected we report an error message. Add error messages where missing, so that all unexpected states provide information about what went wrong. Drop the parentheses around the state value shown in all cases. Signed-off-by:
-
David S. Miller authored
Alex Elder says: ==================== net: ipa: simple refactorizations This series makes three small changes to some endpoint configuration code. The first uses a constant to represent the frequency of an internal clock used for timers in the IPA. The second modifies a limit used so it matches Qualcomm's internal code. And the third reworks a few lines of code, eliminating a multi-line function call. ==================== Signed-off-by: -
Alex Elder authored
Reuse the "limit" local variable in ipa_endpoint_init_aggr() when setting the aggregation size limit. Simple cleanup. Signed-off-by:
-
Alex Elder authored
Halve the time limit used when aggregation is enabled on an RX endpoint, to half a millisecond. Use DIV_ROUND_CLOSEST() to compute the value that represents the time period, to get better accuracy in the event the time limit is not an even multiple of the granularity. Signed-off-by:
-
Alex Elder authored
The timer used for aggregation makes use of an internal 32 KHz clock. The granularity of the timer is programmed by a field whose value is computed by ipa_aggr_granularity_val(). Redefine the way that value is computed by using a new TIMER_FREQUENCY constant representing the underlying clock frequency. Add two BUILD_BUG_ON() calls to ensure the value used is valid. Signed-off-by:
-
David S. Miller authored
Denis Kirjanov says: ==================== xen networking: add XDP support to xen-netfront The first patch adds a new extra type to enable proper synchronization between an RX request/response pair. The second patch implements BFP interface for xen-netfront. The third patch enables extra space for XDP processing. v14: - fixed compilation warnings v13: - fixed compilation due to previous rename v12: - xen-netback: rename netfront_xdp_headroom to xdp_headroom v11: - add the new headroom constant to netif.h - xenbus_scanf check - lock a bulk of puckets in xennet_xdp_xmit() v10: - add a new xen_netif_extra_info type to enable proper synchronization between an RX request/response pair. - order local variable declarations v9: - assign an xdp program before switching to Reconfiguring - minor cleanups - address checkpatch issues v8: - add PAGE_POOL config dependency - keep the state of XDP processing in netfront_xdp_enabled - fixed allocator type in xdp_rxq_info_reg_mem_model() - minor cleanups in xen-netback v7: - use page_pool_dev_alloc_pages() on page allocation - remove the leftover break statement from netback_changed v6: - added the missing SOB line - fixed subject v5: - split netfront/netback changes - added a sync point between backend/frontend on switching to XDP - added pagepool API v4: - added verbose patch descriprion - don't expose the XDP headroom offset to the domU guest - add a modparam to netback to toggle XDP offset - don't process jumbo frames for now v3: - added XDP_TX support (tested with xdping echoserver) - added XDP_REDIRECT support (tested with modified xdp_redirect_kern) - moved xdp negotiation to xen-netback v2: - avoid data copying while passing to XDP - tell xen-netback that we need the headroom space ==================== Signed-off-by: -
Denis Kirjanov authored
the patch basically adds the offset adjustment and netfront state reading to make XDP work on netfront side. Reviewed-by:
Paul Durrant <paul@xen.org> Signed-off-by:
Denis Kirjanov <kda@linux-powerpc.org> Signed-off-by:
-
Denis Kirjanov authored
The patch adds a basic XDP processing to xen-netfront driver. We ran an XDP program for an RX response received from netback driver. Also we request xen-netback to adjust data offset for bpf_xdp_adjust_head() header space for custom headers. synchronization between frontend and backend parts is done by using xenbus state switching: Reconfiguring -> Reconfigured- > Connected UDP packets drop rate using xdp program is around 310 kpps using ./pktgen_sample04_many_flows.sh and 160 kpps without the patch. Signed-off-by:
-
Denis Kirjanov authored
The patch adds a new extra type to be able to diffirentiate between RX responses on xen-netfront side with the adjusted offset required for XDP processing. The offset value from a guest is passed via xenstore. Signed-off-by:
-
Radoslaw Tyl authored
Added full support for new version Ethtool API. New API allow use 2500Gbase-T and 5000base-T supported and advertised link speed modes. Signed-off-by:
Radoslaw Tyl <radoslawx.tyl@intel.com> Tested-by:
-
Jeff Kirsher authored
There is a 4 seconds delay in ixgbe_diag_test() that is holding up other ioctls such as SIOCGIFCONF that Oracle database applications use. One of Oracle's product runs "ethtool -t ethX online" periodically for system monitoring and that is impacting database applications that use SIOCGIFCONF at that same time. This 4 second delay was needed in out early 1GbE parts to give the PHY time to recover from a reset. This code was carried forward to the 10 GbE driver even it was not needed for the supported PHYs in the ixgbe driver. CC: Aleksandr Loktionov <aleksandr.loktionov@intel.com> CC: Jack Vogel <jack.vogel@oracle.com> Reported-by:
Venkat Venkatsubra <venkat.x.venkatsubra@oracle.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Tested-by:
-
Tony Nguyen authored
Commit bac84861 ("iavf: Refactor the watchdog state machine") inverted the logic for when to update statistics. Statistics should be updated when no other commands are pending, instead they were only requested when a command was processed. iavf_request_stats() would see a pending request and not request statistics to be updated. This caused statistics to never be updated; fix the logic. Fixes: bac84861 ("iavf: Refactor the watchdog state machine") Signed-off-by:
-
Ciara Loftus authored
Interfaces already exist for dumping Rx and Tx descriptor information. Introduce another for doing the same for XDP descriptors. Signed-off-by:
Ciara Loftus <ciara.loftus@intel.com> Tested-by:
-
Ciara Loftus authored
Prior to this, only the Rx and Tx ring statistics were dumped. The XDP ring statistics are now dumped as well. Signed-off-by:
-
Ciara Loftus authored
Prior to this, only Rx and Tx ring statistics were accounted for. Signed-off-by:
-
Magnus Karlsson authored
Move the check if the HW Tx ring is full to outside the send loop. Currently it is checked for every single descriptor that we send. Instead, tell the send loop to only process a maximum number of packets equal to the number of available slots in the Tx ring. This way, we can remove the check inside the send loop to and gain some performance. Suggested-by:
Sridhar Samudrala <sridhar.samudrala@intel.com> Signed-off-by:
Magnus Karlsson <magnus.karlsson@intel.com> Tested-by:
-
Magnus Karlsson authored
Eliminate a division in the napi_poll data path. This division is executed even though it is only needed in the rare case when there are not enough interrupt lines so they have to be shared between queue pairs. Instead, just test for this case and only execute the division if needed. The code has been lifted from the ice driver. Signed-off-by:
-
Magnus Karlsson authored
Improve the performance of the AF_XDP zero-copy Tx completion path. When there are no XDP buffers being sent using XDP_TX or XDP_REDIRECT, we do not have go through the SW ring to clean up any entries since the AF_XDP path does not use these. In these cases, just fast forward the next-to-use counter and skip going through the SW ring. The limit on the maximum number of entries to complete is also removed since the algorithm is now O(1). To simplify the code path, the maximum number of entries to complete for the XDP path is therefore also increased from 256 to 512 (the default number of Tx HW descriptors). This should be fine since the completion in the XDP path is faster than in the SKB path that has 256 as the maximum number. This patch provides around 4% throughput improvement for the l2fwd application in xdpsock on my machine. Signed-off-by:
-
Wei Yongjun authored
Fix to return negative error code -ENOMEM from the error handling case instead of 0, as done elsewhere in this function. Fixes: b66c7bc1 ("iavf: Refactor init state machine") Signed-off-by:
Wei Yongjun <weiyongjun1@huawei.com> Tested-by:
-
Arkadiusz Kubalewski authored
After OS requests to down a link on a physical network port, the traffic is no longer being processed but the physical link with a link partner is still established. Currently there is a feature (Link down on close) which allows to physically bring the link down (after OS request). With this patch new feature with similar capability is introduced: TOTAL_PORT_SHUTDOWN Allows to physically disable the link on the NIC's port. If enabled, (after link down request from the OS) no link, traffic or led activity is possible on that port. If I40E_FLAG_TOTAL_PORT_SHUTDOWN is enabled, the I40E_FLAG_LINK_DOWN_ON_CLOSE_ENABLED must be explicitly forced to true and cannot be disabled at that time. The functionalities are exclusive in terms of configuration, but they also have similar behavior (allowing to disable physical link of the port), with following differences: - LINK_DOWN_ON_CLOSE_ENABLED is configurable at host OS run-time and is supported by whole family of 7xx Intel Ethernet Controllers - TOTAL_PORT_SHUTDOWN may be enabled only before OS loads (in BIOS) only if motherboard's BIOS and NIC's FW has support of it - when LINK_DOWN_ON_CLOSE_ENABLED is used, the link is being brought down by sending phy_type=0 to NIC's FW - when TOTAL_PORT_SHUTDOWN is used, phy_type is not altered, instead the link is being brought down by clearing bit (I40E_AQ_PHY_ENABLE_LINK) in abilities field of i40e_aq_set_phy_config structure Introduced changes: - new private flag I40E_FLAG_TOTAL_PORT_SHUTDOWN for handling the feature - probe of NVM if the feature was enabled at driver's port initialization - special handling on link-down procedure to let FW physically shutdown the port if the feature was enabled Signed-off-by:
Arkadiusz Kubalewski <arkadiusz.kubalewski@intel.com> Signed-off-by:
Aleksandr Loktionov <aleksandr.loktionov@intel.com> Tested-by:
-
Jeff Kirsher authored
Convert all the remaining 'fall through" code comments to the newer 'fallthrough;' keyword. Suggested-by:
Joe Perches <joe@perches.com> Signed-off-by:
Aaron Brown <aaron.f.brown@intel.com> Signed-off-by:
-
David S. Miller authored
Vaibhav Gupta says: ==================== net: ethernet: use generic power management Linux Kernel Mentee: Remove Legacy Power Management. The purpose of this patch series is to remove legacy power management callbacks from net ethernet drivers. The callbacks performing suspend() and resume() operations are still calling pci_save_state(), pci_set_power_state(), etc. and handling the power management themselves, which is not recommended. The conversion requires the removal of the those function calls and change the callback definition accordingly and make use of dev_pm_ops structure. All patches are compile-tested only. ==================== Signed-off-by: -
Vaibhav Gupta authored
With legacy PM, drivers themselves were responsible for managing the device's power states and takes care of register states. After upgrading to the generic structure, PCI core will take care of required tasks and drivers should do only device-specific operations. Thus, there is no need to call the PCI helper functions like pci_enable_device, which is not recommended. Hence, removed. Compile-tested only. Signed-off-by:
Vaibhav Gupta <vaibhavgupta40@gmail.com> Signed-off-by:
-
