- 21 Jun, 2024 5 commits
-
-
Frank Li authored
Convert fsl-fman from txt to yaml format and split it fsl,fman.yam, fsl,fman-port.yaml, fsl-muram.yaml, fsl-mdio.yaml. Addition changes: fsl,fman.yaml: - Fixed interrupts in example. - Fixed ethernet@e8000 miss } in example. - ptp-timer add label in example. - Ref to new fsl,fman*.yaml. - Reorder property in example. - Keep only one example. - Add const for #address-cells and #size-cells. - Use defined interrupt type. - ptp example use node name phc. fsl,fman-port: - Keep only one example. fsl,fman-mdio: - Add little-endian property. - Add ref to mdio.yaml. - Remove suppress-preamble. - Add #address-cells and #size-cells in example. - Remove clock-frequency, which already describe in mmio.yaml. fsl,muram.yaml: - Add reg property. - Remove range property. - Use reg instead of range in example. Signed-off-by:
Frank Li <Frank.Li@nxp.com> Reviewed-by:
Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Link: https://patch.msgid.link/20240618-ls_fman-v2-2-f00a82623d8e@nxp.comSigned-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Frank Li authored
Convert ptp-qoirq from txt to yaml format. Additional change: - Fixed example interrupts proptery. Need only 1 irq by check MPC8313 spec. - Move Reference clock context under clk,sel. - Interrupts is not required property. - Use low case for hex value. - Check reference manual of MPC8313, p1010 and so on, which dts use more than 1 irqs. Only 1 irq for each ptp device. Check driver code (drivers/ptp/ptp_qoriq.c) and only 1 irq used. So original description is wrong. - Remove comments for compatible string. Signed-off-by:
-
Breno Leitao authored
Embedding net_device into structures prohibits the usage of flexible arrays in the net_device structure. For more details, see the discussion at [1]. Un-embed the net_devices from struct mt76_dev by converting them into pointers, and allocating them dynamically. Use the leverage alloc_netdev_dummy() to allocate the net_device object at mt76_dma_init(). The free of the device occurs at mt76_dma_cleanup(). Link: https://lore.kernel.org/all/20240229225910.79e224cf@kernel.org/ [1] Reviewed-by:
Simon Horman <horms@kernel.org> Signed-off-by:
Breno Leitao <leitao@debian.org> Acked-by:
Kalle Valo <kvalo@kernel.org> Link: https://patch.msgid.link/20240619105311.3144908-1-leitao@debian.orgSigned-off-by:
-
Bartosz Golaszewski authored
These functions are only used within the compilation unit they're defined in so there's no reason to export them. Signed-off-by:
Bartosz Golaszewski <bartosz.golaszewski@linaro.org> Reviewed-by:
-
Uwe Kleine-König authored
These drivers don't use the driver_data member of struct i2c_device_id, so don't explicitly initialize this member. This prepares putting driver_data in an anonymous union which requires either no initialization or named designators. But it's also a nice cleanup on its own. Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@baylibre.com> Reviewed-by:
-
- 20 Jun, 2024 35 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/netdev/netJakub Kicinski authored
Cross-merge networking fixes after downstream PR. Conflicts: drivers/net/ethernet/broadcom/bnxt/bnxt.c 1e796211 ("bnxt_en: Restore PTP tx_avail count in case of skb_pad() error") 165f8769 ("bnxt_en: add timestamping statistics support") No adjacent changes. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/netdev/netLinus Torvalds authored
Pull networking fixes from Jakub Kicinski: "Including fixes from wireless, bpf and netfilter. Happy summer solstice! The line count is a bit inflated by a selftest and update to a driver's FW interface header, in reality this is slightly below average for us. We are expecting one driver fix from Intel, but there are no big known issues. Current release - regressions: - ipv6: bring NLM_DONE out to a separate recv() again Current release - new code bugs: - wifi: cfg80211: wext: set ssids=NULL for passive scans via old wext API Previous releases - regressions: - wifi: mac80211: fix monitor channel setting with chanctx emulation (probably most awaited of the fixes in this PR, tracked by Thorsten) - usb: ax88179_178a: bring back reset on init, if PHY is disconnected - bpf: fix UML x86_64 compile failure with BPF - bpf: avoid splat in pskb_pull_reason(), sanity check added can be hit with malicious BPF - eth: mvpp2: use slab_build_skb() for packets in slab, driver was missed during API refactoring - wifi: iwlwifi: add missing unlock of mvm mutex Previous releases - always broken: - ipv6: add a number of missing null-checks for in6_dev_get(), in case IPv6 disabling races with the datapath - bpf: fix reg_set_min_max corruption of fake_reg - sched: act_ct: add netns as part of the key of tcf_ct_flow_table" * tag 'net-6.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (63 commits) net: usb: rtl8150 fix unintiatilzed variables in rtl8150_get_link_ksettings selftests: virtio_net: add forgotten config options bnxt_en: Restore PTP tx_avail count in case of skb_pad() error bnxt_en: Set TSO max segs on devices with limits bnxt_en: Update firmware interface to 1.10.3.44 net: stmmac: Assign configured channel value to EXTTS event net: do not leave a dangling sk pointer, when socket creation fails net/tcp_ao: Don't leak ao_info on error-path ice: Fix VSI list rule with ICE_SW_LKUP_LAST type ipv6: bring NLM_DONE out to a separate recv() again selftests: add selftest for the SRv6 End.DX6 behavior with netfilter selftests: add selftest for the SRv6 End.DX4 behavior with netfilter netfilter: move the sysctl nf_hooks_lwtunnel into the netfilter core seg6: fix parameter passing when calling NF_HOOK() in End.DX4 and End.DX6 behaviors netfilter: ipset: Fix suspicious rcu_dereference_protected() selftests: openvswitch: Set value to nla flags. octeontx2-pf: Fix linking objects into multiple modules octeontx2-pf: Add error handling to VLAN unoffload handling virtio_net: fixing XDP for fully checksummed packets handling virtio_net: checksum offloading handling fix ... -
git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/soundLinus Torvalds authored
Pull sound fixes from Takashi Iwai: "Lots of small HD-audio quirks and fixes (mostly Realtek codec and Cirrus stuff). Also a small MIDI 2.0 fix and a fix for missing module description are included" * tag 'sound-6.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound: ALSA: hda: cs35l56: Select SERIAL_MULTI_INSTANTIATE ALSA: hda/realtek: Add more codec ID to no shutup pins list sound/oss/dmasound: add missing MODULE_DESCRIPTION() macro ALSA: hda/realtek: Add quirk for Lenovo Yoga Pro 7 14ARP8 ALSA: hda/realtek: Enable headset mic on IdeaPad 330-17IKB 81DM ALSA: hda: tas2781: Component should be unbound before deconstruction ALSA: hda: cs35l41: Component should be unbound before deconstruction ALSA: hda: cs35l56: Component should be unbound before deconstruction ALSA/hda: intel-dsp-config: Document AVS as dsp_driver option ALSA: hda/realtek: Support Lenovo Thinkbook 13x Gen 4 ALSA: hda/realtek: Support Lenovo Thinkbook 16P Gen 5 ALSA: hda: cs35l41: Support Lenovo Thinkbook 13x Gen 4 ALSA: hda: cs35l41: Support Lenovo Thinkbook 16P Gen 5 ALSA: hda/realtek: Remove Framework Laptop 16 from quirks ALSA: hda/realtek: Limit mic boost on N14AP7 ALSA: hda/realtek: fix mute/micmute LEDs don't work for ProBook 445/465 G11. ALSA: seq: ump: Fix missing System Reset message handling ALSA: hda: cs35l41: Possible null pointer dereference in cs35l41_hda_unbind() ALSA: hda: cs35l56: Fix lifecycle of codec pointer
-
git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfdLinus Torvalds authored
Pull mfd fix from Lee Jones: - Fix AXP717 PMIC probe and by extension its consumers * tag 'mfd-fixes-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd: mfd: axp20x: AXP717: Fix missing IRQ status registers range
-
Oliver Neukum authored
This functions retrieves values by passing a pointer. As the function that retrieves them can fail before touching the pointers, the variables must be initialized. Fixes: 1da177e4 ("Linux-2.6.12-rc2") Reported-by: syzbot+5186630949e3c55f0799@syzkaller.appspotmail.com Signed-off-by:
Oliver Neukum <oneukum@suse.com> Link: https://lore.kernel.org/r/20240619132816.11526-1-oneukum@suse.comSigned-off-by:
-
Jiri Pirko authored
One may use tools/testing/selftests/drivers/net/virtio_net/config for example for vng build command like this one: $ vng -v -b -f tools/testing/selftests/drivers/net/virtio_net/config In that case, the needed kernel config options are not turned on. Add the missed kernel config options. Reported-by:
Matthieu Baerts <matttbe@kernel.org> Closes: https://lore.kernel.org/netdev/1a63f209-b1d4-4809-bc30-295a5cafa296@kernel.org/ Fixes: ccfaed04 ("selftests: virtio_net: add initial tests") Signed-off-by:
Jiri Pirko <jiri@nvidia.com> Reviewed-by:
Xuan Zhuo <xuanzhuo@linux.alibaba.com> Acked-by:
Michael S. Tsirkin <mst@redhat.com> Link: https://lore.kernel.org/r/20240619061748.1869404-1-jiri@resnulli.usSigned-off-by:
-
Jakub Kicinski authored
Michael Chan says: ==================== bnxt_en: Bug fixes for net The first firmware interface update is needed by the second patch to limit the number of TSO segments on the 5760X chips. The third patch fixes the TX error path for PTP packets. ==================== Link: https://lore.kernel.org/r/20240618215313.29631-1-michael.chan@broadcom.comSigned-off-by:
-
Pavan Chebbi authored
The current code only restores PTP tx_avail count when we get DMA mapping errors. Fix it so that the PTP tx_avail count will be restored for both DMA mapping errors and skb_pad() errors. Otherwise PTP TX timestamp will not be available after a PTP packet hits the skb_pad() error. Fixes: 83bb623c ("bnxt_en: Transmit and retrieve packet timestamps") Reviewed-by:
Andy Gospodarek <andrew.gospodarek@broadcom.com> Signed-off-by:
Pavan Chebbi <pavan.chebbi@broadcom.com> Signed-off-by:
Michael Chan <michael.chan@broadcom.com> Reviewed-by:
-
Michael Chan authored
Firmware will now advertise a non-zero TSO max segments if the device has a limit. 0 means no limit. The latest 5760X chip (early revs) has a limit of 2047 that cannot be exceeded. If exceeded, the chip will send out just a small number of segments. Call netif_set_tso_max_segs() if the device has a limit. Fixes: 2012a6ab ("bnxt_en: Add 5760X (P7) PCI IDs") Reviewed-by:
Ajit Khaparde <ajit.khaparde@broadcom.com> Reviewed-by:
Somnath Kotur <somnath.kotur@broadcom.com> Signed-off-by:
-
Michael Chan authored
The relevant change is the max_tso_segs value returned by firmware in the HWRM_FUNC_QCAPS response. This value will be used in the next patch to cap the TSO segments. Signed-off-by:
-
Jackie Jone authored
To facilitate running PHY parametric tests, add support for the SIOCSMIIREG ioctl. This allows a userspace application to write to the PHY registers to enable the test modes. Signed-off-by:
Jackie Jone <jackie.jone@alliedtelesis.co.nz> Acked-by:
Jacob Keller <jacob.e.keller@intel.com> Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel) Signed-off-by:
Tony Nguyen <anthony.l.nguyen@intel.com> Link: https://lore.kernel.org/r/20240618213330.982046-1-anthony.l.nguyen@intel.comSigned-off-by:
Paolo Abeni <pabeni@redhat.com>
-
Paolo Abeni authored
Jiawen Wu says: ==================== add flow director for txgbe Add flow director support for Wangxun 10Gb NICs. v2 -> v3: https://lore.kernel.org/all/20240605020852.24144-1-jiawenwu@trustnetic.com/ - Wrap the code at 80 chars where possible. (Jakub Kicinski) - Add function description address on kernel-doc. (Jakub Kicinski) - Correct return code. (Simon Horman) - Remove redundant size check. (Hariprasad Kelam) v1 -> v2: https://lore.kernel.org/all/20240529093821.27108-1-jiawenwu@trustnetic.com/ - Fix build warnings reported by kernel test robot. ==================== Link: https://lore.kernel.org/r/20240618101609.3580-1-jiawenwu@trustnetic.comSigned-off-by:
-
Jiawen Wu authored
Add flow director filter match and miss statistics to ethtool -S. And change the number of queues when using flow director for ehtool -l. Signed-off-by:
Jiawen Wu <jiawenwu@trustnetic.com> Reviewed-by:
-
Jiawen Wu authored
Support the addition and deletion of Flow Director filters. Supported fields: src-ip, dst-ip, src-port, dst-port Supported flow-types: tcp4, udp4, sctp4, ipv4 Signed-off-by:
-
Jiawen Wu authored
Add flow director ATR filter. ATR mode is enabled by default to filter TCP packets. Signed-off-by:
-
Oleksij Rempel authored
Assign the configured channel value to the EXTTS event in the timestamp interrupt handler. Without assigning the correct channel, applications like ts2phc will refuse to accept the event, resulting in errors such as: ... ts2phc[656.834]: config item end1.ts2phc.pin_index is 0 ts2phc[656.834]: config item end1.ts2phc.channel is 3 ts2phc[656.834]: config item end1.ts2phc.extts_polarity is 2 ts2phc[656.834]: config item end1.ts2phc.extts_correction is 0 ... ts2phc[656.862]: extts on unexpected channel ts2phc[658.141]: extts on unexpected channel ts2phc[659.140]: extts on unexpected channel Fixes: f4da5652 ("net: stmmac: Add support for external trigger timestamping") Cc: stable@vger.kernel.org Signed-off-by:
Oleksij Rempel <o.rempel@pengutronix.de> Reviewed-by:
Wojciech Drewek <wojciech.drewek@intel.com> Link: https://lore.kernel.org/r/20240618073821.619751-1-o.rempel@pengutronix.deSigned-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nfPaolo Abeni authored
Pablo Neira Ayuso says: ==================== Netfilter fixes for net The following patchset contains Netfilter fixes for net: Patch #1 fixes the suspicious RCU usage warning that resulted from the recent fix for the race between namespace cleanup and gc in ipset left out checking the pernet exit phase when calling rcu_dereference_protected(), from Jozsef Kadlecsik. Patch #2 fixes incorrect input and output netdevice in SRv6 prerouting hooks, from Jianguo Wu. Patch #3 moves nf_hooks_lwtunnel sysctl toggle to the netfilter core. The connection tracking system is loaded on-demand, this ensures availability of this knob regardless. Patch #4-#5 adds selftests for SRv6 netfilter hooks also from Jianguo Wu. netfilter pull request 24-06-19 * tag 'nf-24-06-19' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf: selftests: add selftest for the SRv6 End.DX6 behavior with netfilter selftests: add selftest for the SRv6 End.DX4 behavior with netfilter netfilter: move the sysctl nf_hooks_lwtunnel into the netfilter core seg6: fix parameter passing when calling NF_HOOK() in End.DX4 and End.DX6 behaviors netfilter: ipset: Fix suspicious rcu_dereference_protected() ==================== Link: https://lore.kernel.org/r/20240619170537.2846-1-pablo@netfilter.orgSigned-off-by:
-
Ignat Korchagin authored
It is possible to trigger a use-after-free by: * attaching an fentry probe to __sock_release() and the probe calling the bpf_get_socket_cookie() helper * running traceroute -I 1.1.1.1 on a freshly booted VM A KASAN enabled kernel will log something like below (decoded and stripped): ================================================================== BUG: KASAN: slab-use-after-free in __sock_gen_cookie (./arch/x86/include/asm/atomic64_64.h:15 ./include/linux/atomic/atomic-arch-fallback.h:2583 ./include/linux/atomic/atomic-instrumented.h:1611 net/core/sock_diag.c:29) Read of size 8 at addr ffff888007110dd8 by task traceroute/299 CPU: 2 PID: 299 Comm: traceroute Tainted: G E 6.10.0-rc2+ #2 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014 Call Trace: <TASK> dump_stack_lvl (lib/dump_stack.c:117 (discriminator 1)) print_report (mm/kasan/report.c:378 mm/kasan/report.c:488) ? __sock_gen_cookie (./arch/x86/include/asm/atomic64_64.h:15 ./include/linux/atomic/atomic-arch-fallback.h:2583 ./include/linux/atomic/atomic-instrumented.h:1611 net/core/sock_diag.c:29) kasan_report (mm/kasan/report.c:603) ? __sock_gen_cookie (./arch/x86/include/asm/atomic64_64.h:15 ./include/linux/atomic/atomic-arch-fallback.h:2583 ./include/linux/atomic/atomic-instrumented.h:1611 net/core/sock_diag.c:29) kasan_check_range (mm/kasan/generic.c:183 mm/kasan/generic.c:189) __sock_gen_cookie (./arch/x86/include/asm/atomic64_64.h:15 ./include/linux/atomic/atomic-arch-fallback.h:2583 ./include/linux/atomic/atomic-instrumented.h:1611 net/core/sock_diag.c:29) bpf_get_socket_ptr_cookie (./arch/x86/include/asm/preempt.h:94 ./include/linux/sock_diag.h:42 net/core/filter.c:5094 net/core/filter.c:5092) bpf_prog_875642cf11f1d139___sock_release+0x6e/0x8e bpf_trampoline_6442506592+0x47/0xaf __sock_release (net/socket.c:652) __sock_create (net/socket.c:1601) ... Allocated by task 299 on cpu 2 at 78.328492s: kasan_save_stack (mm/kasan/common.c:48) kasan_save_track (mm/kasan/common.c:68) __kasan_slab_alloc (mm/kasan/common.c:312 mm/kasan/common.c:338) kmem_cache_alloc_noprof (mm/slub.c:3941 mm/slub.c:4000 mm/slub.c:4007) sk_prot_alloc (net/core/sock.c:2075) sk_alloc (net/core/sock.c:2134) inet_create (net/ipv4/af_inet.c:327 net/ipv4/af_inet.c:252) __sock_create (net/socket.c:1572) __sys_socket (net/socket.c:1660 net/socket.c:1644 net/socket.c:1706) __x64_sys_socket (net/socket.c:1718) do_syscall_64 (arch/x86/entry/common.c:52 arch/x86/entry/common.c:83) entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130) Freed by task 299 on cpu 2 at 78.328502s: kasan_save_stack (mm/kasan/common.c:48) kasan_save_track (mm/kasan/common.c:68) kasan_save_free_info (mm/kasan/generic.c:582) poison_slab_object (mm/kasan/common.c:242) __kasan_slab_free (mm/kasan/common.c:256) kmem_cache_free (mm/slub.c:4437 mm/slub.c:4511) __sk_destruct (net/core/sock.c:2117 net/core/sock.c:2208) inet_create (net/ipv4/af_inet.c:397 net/ipv4/af_inet.c:252) __sock_create (net/socket.c:1572) __sys_socket (net/socket.c:1660 net/socket.c:1644 net/socket.c:1706) __x64_sys_socket (net/socket.c:1718) do_syscall_64 (arch/x86/entry/common.c:52 arch/x86/entry/common.c:83) entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130) Fix this by clearing the struct socket reference in sk_common_release() to cover all protocol families create functions, which may already attached the reference to the sk object with sock_init_data(). Fixes: c5dbb89f ("bpf: Expose bpf_get_socket_cookie to tracing programs") Suggested-by:Kuniyuki Iwashima <kuniyu@amazon.com> Signed-off-by:
Ignat Korchagin <ignat@cloudflare.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/netdev/20240613194047.36478-1-kuniyu@amazon.com/T/Reviewed-by:
D. Wythe <alibuda@linux.alibaba.com> Link: https://lore.kernel.org/r/20240617210205.67311-1-ignat@cloudflare.comSigned-off-by:
-
Simon Trimmer authored
The ACPI IDs used in the CS35L56 HDA drivers are all handled by the serial multi-instantiate driver which starts multiple Linux device instances from a single ACPI Device() node. As serial multi-instantiate is not an optional part of the system add it as a dependency in Kconfig so that it is not overlooked. Signed-off-by:
Simon Trimmer <simont@opensource.cirrus.com> Link: https://lore.kernel.org/20240619161602.117452-1-simont@opensource.cirrus.comSigned-off-by:
Takashi Iwai <tiwai@suse.de>
-
Jakub Kicinski authored
Shannon Nelson says: ==================== ionic: rework fix for doorbell miss A latency test in a scaled out setting (many VMs with many queues) has uncovered an issue with our missed doorbell fix from commit b69585bf ("ionic: missed doorbell workaround") As a refresher, the Elba ASIC has an issue where once in a blue moon it might miss/drop a queue doorbell notification from the driver. This can result in Tx timeouts and potential Rx buffer misses. The basic problem with the original solution is that we're delaying things with a timer for every single queue, periodically using mod_timer() to reset to reset the alarm, and mod_timer() becomes a more and more expensive thing as there are more and more VFs and queues each with their own timer. A ping-pong latency test tends to exacerbate the effect such that every napi is doing a mod_timer() in every cycle. An alternative has been worked out to replace this using periodic workqueue items outside the napi cycle to request a napi_schedule driven by a single delayed-workqueue per device rather than a timer for every queue. Also, now that newer firmware is actually reporting its ASIC type, we can restrict this to the appropriate chip. The testing scenario used 128 VFs in UP state, 16 queues per VF, and latency tests were done using TCP_RR with adaptive interrupt coalescing enabled, running on 1 VF. We would see 99th percentile latencies of up to 900us range, with some max fliers as much as 4ms. With these fixes the 99th percentile latencies are typically well under 50us with the occasional max under 500us. v1: https://lore.kernel.org/netdev/20240610230706.34883-1-shannon.nelson@amd.com/ ==================== Link: https://lore.kernel.org/r/20240619003257.6138-1-shannon.nelson@amd.comSigned-off-by:
-
Brett Creeley authored
If the doorbell workaround isn't required for a certain asic_type then there is no need to run the associated code. Since newer FW versions are finally reporting their asic_type we can use a flag to determine whether or not to do the workaround. Signed-off-by:
Brett Creeley <brett.creeley@amd.com> Signed-off-by:
Shannon Nelson <shannon.nelson@amd.com> Link: https://lore.kernel.org/r/20240619003257.6138-9-shannon.nelson@amd.comSigned-off-by:
-
Brett Creeley authored
We only support (u16)-1 size for rx_copybreak, so we can reduce the field size and move a couple other fields around to save a little space in the ionic_lif struct. Before: /* size: 17440, cachelines: 273, members: 56 */ /* sum members: 17403, holes: 9, sum holes: 37 */ /* last cacheline: 32 bytes */ After: /* size: 17424, cachelines: 273, members: 56 */ /* sum members: 17401, holes: 7, sum holes: 23 */ /* last cacheline: 16 bytes */ Signed-off-by:
-
Shannon Nelson authored
Check the deadline against the last time run and only schedule a new napi if we haven't been run recently. Signed-off-by:
-
Shannon Nelson authored
Add a work item for each queue that will be run on the queue's preferred cpu and will schedule another napi. This napi is run in case the device missed a doorbell and didn't process a packet. This is a problem for the Elba asic that happens very rarely. Signed-off-by:
-
Shannon Nelson authored
Add the first queued work for checking on the missed doorbell. This is a delayed work item that reschedules itself every cycle starting at probe. Signed-off-by:
-
Shannon Nelson authored
Instead of using the system's default workqueue, add a private workqueue for the device to use for its little jobs. This is to better support the new work items we will be adding in the next patches for PF and VF specific jobs, without inundating the system workqueue in a couple of customer cases where our devices get scaled out to 100-200 VFs. Signed-off-by:
-
Brett Creeley authored
Currently the driver either sets the initial interrupt affinity for its adminq and tx/rx queues on probe or resets it on various down/up/reconfigure flows. If any user and/or user process (i.e. irqbalance) changes IRQ affinity for any of the driver's interrupts that will be reset to driver defaults whenever any down/up/reconfigure operation happens. This is incorrect and is fixed by making 2 changes: 1. Allocate an array of cpumasks that's only allocated on probe and destroyed on remove. 2. Update the cpumask(s) for interrupts that are in use by registering for affinity notifiers. Signed-off-by:
-
Shannon Nelson authored
Remove the timer-per-queue mechanics from the missed doorbell check in preparation for the new missed doorbell fix. Signed-off-by:
-
Jakub Kicinski authored
Petr Machata says: ==================== mlxsw: Use page pool for Rx buffers allocation Amit Cohen writes: After using NAPI to process events from hardware, the next step is to use page pool for Rx buffers allocation, which is also enhances performance. To simplify this change, first use page pool to allocate one continuous buffer for each packet, later memory consumption can be improved by using fragmented buffers. This set significantly enhances mlxsw driver performance, CPU can handle about 370% of the packets per second it previously handled. The next planned improvement is using XDP to optimize telemetry. Patch set overview: Patches #1-#2 are small preparations for page pool usage Patch #3 initializes page pool, but do not use it Patch #4 converts the driver to use page pool for buffers allocations Patch #5 is an optimization for buffer access Patch #6 cleans up an unused structure Patch #7 uses napi_consume_skb() as part of Tx completion ==================== Link: https://lore.kernel.org/r/cover.1718709196.git.petrm@nvidia.comSigned-off-by:
-
Amit Cohen authored
Currently, as part of Tx completion, the driver calls dev_kfree_skb_any() to free the SKB. For this flow, the correct function is napi_consume_skb(). This function and dev_consume_skb_any() were added to be used for consumed SKBs, which were not dropped, so the skb:kfree_skb tracepoint is not triggered, and we can get better diagnostics about dropped packets. Signed-off-by:
Amit Cohen <amcohen@nvidia.com> Reviewed-by:
Ido Schimmel <idosch@nvidia.com> Signed-off-by:
Petr Machata <petrm@nvidia.com> Reviewed-by:
-
Amit Cohen authored
The previous patch used page pool to allocate buffers for RDQ. With this change, 'elem_info->u.rdq.skb' is not used anymore, as we do not allocate SKB before getting the packet, we hold page pointer and build the SKB around it once packet is received. Remove the union and store SKB pointer for SDQ only. Signed-off-by:
-
Amit Cohen authored
Before accessing data buffer, call net_prefetch() to load it into the cache. This change improves driver performance, CPU can handle about 7.1% more packets per second. Signed-off-by:
-
Amit Cohen authored
As part of driver init, all Rx queues are filled with buffers for hardware usage. Later, when a packet is received, a new buffer should be allocated to be used by hardware instead of the received buffer. Packet's processing time includes allocation time, which can be improved using page pool. Using page pool, DMA mapping is done only for first allocation of buffers. As subsequent buffers allocation avoid DMA mapping, it results in performance improvement. The purpose of page pool is to allocate pages fast from cache without locking. This lockless guarantee naturally comes from running under a NAPI. Use page pool to allocate the data buffer only, so hardware will use it to fill the packet. At completion time, attach the data buffer (now filled with packet payload) to new SKB which is allocated around the received buffer. SKB building at completion time prevents cache miss for each packet, as now the SKB is allocated right before packets will be handled by networking stack. Page pool for each Rx queue enhances Rx side performance by reclaiming buffers back to each queue specific pool. This change significantly improves driver performance, CPU can handle about 345% of the packets per second it previously handled. Signed-off-by:
-
Amit Cohen authored
Next patch will use page pool to allocate buffers for RDQ. Initialize page pool for each CQ, which is mapped 1:1 to RDQ. Page pool for each Rx queue enhances Rx side performance by reclaiming buffers back to each queue specific pool. When only one NAPI instance is the consumer of pages from page pool, it is recommended to pass it as part of 'page_pool_params', then page pool APIs will be done without special locks. mlxsw driver holds NAPI instance per CQ, so add page pool per CQ and use the existing NAPI instance. For now, pages are not allocated from the pool, next patch will use it. Some notes regarding 'page_pool_params': * Use PP_FLAG_DMA_MAP to allow page pool handles DMA mapping, for now do not use sync flag, as only the device writes to this memory and we read it only when it finishes writing there. This will probably be changed when we will support XDP. * Define 'order' according to maximum MTU and take into account software overhead. Some round up are done, which means that we allocate more pages than we really need. This can be improved later by using fragmented buffers. * Use pool_size = MLXSW_PCI_WQE_COUNT. This will be the size of 'ptr_ring', and should be the maximum amount of packets that page pool will allocate memory for. In our case, this is the queue size, defined as MLXSW_PCI_WQE_COUNT. Signed-off-by:
-
Amit Cohen authored
Next patches will add support for page pool in mlxsw driver. Page pool will be used to allocate buffers for RDQ and will use NAPI instance of the appropriate CQ (RDQ is mapped 1:1 to CQ). To allow pool initialization as part of CQ init, when NAPI is initialized, page_pool structure will be as part of CQ structure. Later, the allocations for RDQ will be done from the pool in the appropriate CQ. To allow access to the appropriate pool, set CQ pointer as part of RDQ initialization. Signed-off-by:
-
