- 11 Mar, 2024 5 commits
-
-
Herve Codina authored
QMC channels support runtime timeslots changes but nothing is done at the QMC HDLC driver to handle these changes. Use existing IFACE ioctl in order to configure the timeslots to use. Signed-off-by:
Herve Codina <herve.codina@bootlin.com> Reviewed-by:
Christophe Leroy <christophe.leroy@csgroup.eu> Acked-by:
Jakub Kicinski <kuba@kernel.org> Reviewed-by:
Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Andy Shevchenko authored
These helpers scatters or gathers a bitmap with the help of the mask position bits parameter. bitmap_scatter() does the following: src: 0000000001011010 |||||| +------+||||| | +----+|||| | |+----+||| | || +-+|| | || | || mask: ...v..vv...v..vv ...0..11...0..10 dst: 0000001100000010 and bitmap_gather() performs this one: mask: ...v..vv...v..vv src: 0000001100000010 ^ ^^ ^ 0 | || | 10 | || > 010 | |+--> 1010 | +--> 11010 +----> 011010 dst: 0000000000011010 bitmap_gather() can the seen as the reverse bitmap_scatter() operation. Signed-off-by:Yury Norov <yury.norov@gmail.com> Signed-off-by:
-
Herve Codina authored
After contributing the driver, add myself as the maintainer for the Freescale QMC HDLC driver. Signed-off-by:
-
Herve Codina authored
The QMC HDLC driver provides support for HDLC using the QMC (QUICC Multichannel Controller) to transfer the HDLC data. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queueDavid S. Miller authored
Tony Nguyen says: ==================== ethtool: ice: Support for RSS settings to GTP Takeru Hayasaka enables RSS functionality for GTP packets on ice driver with ethtool. A user can include TEID and make RSS work for GTP-U over IPv4 by doing the following:`ethtool -N ens3 rx-flow-hash gtpu4 sde` In addition to gtpu(4|6), we now support gtpc(4|6),gtpc(4|6)t,gtpu(4|6)e, gtpu(4|6)u, and gtpu(4|6)d. gtpc(4|6): Used for GTP-C in IPv4 and IPv6, where the GTP header format does not include a TEID. gtpc(4|6)t: Used for GTP-C in IPv4 and IPv6, with a GTP header format that includes a TEID. gtpu(4|6): Used for GTP-U in both IPv4 and IPv6 scenarios. gtpu(4|6)e: Used for GTP-U with extended headers in both IPv4 and IPv6. gtpu(4|6)u: Used when the PSC (PDU session container) in the GTP-U extended header includes Uplink, applicable to both IPv4 and IPv6. gtpu(4|6)d: Used when the PSC in the GTP-U extended header includes Downlink, for both IPv4 and IPv6. ==================== Signed-off-by:
-
- 09 Mar, 2024 4 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linuxJakub Kicinski authored
Saeed Mahameed says: ==================== Support Multi-PF netdev (Socket Direct) This series adds support for combining multiple devices (PFs) of the same port under one netdev instance. Passing traffic through different devices belonging to different NUMA sockets saves cross-numa traffic and allows apps running on the same netdev from different numas to still feel a sense of proximity to the device and achieve improved performance. We achieve this by grouping PFs together, and creating the netdev only once all group members are probed. Symmetrically, we destroy the netdev once any of the PFs is removed. The channels are distributed between all devices, a proper configuration would utilize the correct close numa when working on a certain app/cpu. We pick one device to be a primary (leader), and it fills a special role. The other devices (secondaries) are disconnected from the network in the chip level (set to silent mode). All RX/TX traffic is steered through the primary to/from the secondaries. Currently, we limit the support to PFs only, and up to two devices (sockets). * tag 'mlx5-socket-direct-v3' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux: Documentation: networking: Add description for multi-pf netdev net/mlx5: Enable SD feature net/mlx5e: Block TLS device offload on combined SD netdev net/mlx5e: Support per-mdev queue counter net/mlx5e: Support cross-vhca RSS net/mlx5e: Let channels be SD-aware net/mlx5e: Create EN core HW resources for all secondary devices net/mlx5e: Create single netdev per SD group net/mlx5: SD, Add debugfs net/mlx5: SD, Add informative prints in kernel log net/mlx5: SD, Implement steering for primary and secondaries net/mlx5: SD, Implement devcom communication and primary election net/mlx5: SD, Implement basic query and instantiation net/mlx5: SD, Introduce SD lib net/mlx5: Add MPIR bit in mcam_access_reg ==================== Link: https://lore.kernel.org/r/20240307084229.500776-1-saeed@kernel.orgSigned-off-by:
-
Jakub Kicinski authored
Merge tag 'for-net-next-2024-03-08' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next Luiz Augusto von Dentz says: ==================== bluetooth-next pull request for net-next: - hci_conn: Only do ACL connections sequentially - hci_core: Cancel request on command timeout - Remove CONFIG_BT_HS - btrtl: Add the support for RTL8852BT/RTL8852BE-VT - btusb: Add support Mediatek MT7920 - btusb: Add new VID/PID 13d3/3602 for MT7925 - Add new quirk for broken read key length on ATS2851 * tag 'for-net-next-2024-03-08' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next: (52 commits) Bluetooth: hci_sync: Fix UAF in hci_acl_create_conn_sync Bluetooth: Fix eir name length Bluetooth: ISO: Align broadcast sync_timeout with connection timeout Bluetooth: Add new quirk for broken read key length on ATS2851 Bluetooth: mgmt: remove NULL check in add_ext_adv_params_complete() Bluetooth: mgmt: remove NULL check in mgmt_set_connectable_complete() Bluetooth: btusb: Add support Mediatek MT7920 Bluetooth: btmtk: Add MODULE_FIRMWARE() for MT7922 Bluetooth: btnxpuart: Fix btnxpuart_close Bluetooth: ISO: Clean up returns values in iso_connect_ind() Bluetooth: fix use-after-free in accessing skb after sending it Bluetooth: af_bluetooth: Fix deadlock Bluetooth: bnep: Fix out-of-bound access Bluetooth: btusb: Fix memory leak Bluetooth: msft: Fix memory leak Bluetooth: hci_core: Fix possible buffer overflow Bluetooth: btrtl: fix out of bounds memory access Bluetooth: hci_h5: Add ability to allocate memory for private data Bluetooth: hci_sync: Fix overwriting request callback Bluetooth: hci_sync: Use QoS to determine which PHY to scan ... ==================== Link: https://lore.kernel.org/r/20240308181056.120547-1-luiz.dentz@gmail.comSigned-off-by:
-
Jakub Kicinski authored
Merge tag 'ieee802154-for-net-next-2024-03-07' of git://git.kernel.org/pub/scm/linux/kernel/git/wpan/wpan-next Stefan Schmidt says: ==================== pull-request: ieee802154-next 2024-03-07 Various cross tree patches for ieee802154v drivers and a resource leak fix for ieee802154 llsec. Andy Shevchenko changed GPIO header usage for at86rf230 and mcr20a to only include needed headers. Bo Liu converted the at86rf230, mcr20a and mrf24j40 driver regmap support to use the maple tree register cache. Fedor Pchelkin fixed a resource leak in the llsec key deletion path. Ricardo B. Marliere made wpan_phy_class const. Tejun Heo removed WQ_UNBOUND from a workqueue call in ca8210. * tag 'ieee802154-for-net-next-2024-03-07' of git://git.kernel.org/pub/scm/linux/kernel/git/wpan/wpan-next: ieee802154: cfg802154: make wpan_phy_class constant ieee802154: mcr20a: Remove unused of_gpio.h ieee802154: at86rf230: Replace of_gpio.h by proper one mac802154: fix llsec key resources release in mac802154_llsec_key_del ieee802154: ca8210: Drop spurious WQ_UNBOUND from alloc_ordered_workqueue() call net: ieee802154: mrf24j40: convert to use maple tree register cache net: ieee802154: mcr20a: convert to use maple tree register cache net: ieee802154: at86rf230: convert to use maple tree register cache ==================== Link: https://lore.kernel.org/r/20240307195105.292085-1-stefan@datenfreihafen.orgSigned-off-by:
-
Colin Ian King authored
There is a spelling mistake in an error message. Fix it. Signed-off-by:
Colin Ian King <colin.i.king@gmail.com> Acked-by:
-
- 08 Mar, 2024 31 commits
-
-
Eric Dumazet authored
There is no point cloning an skb and having to free the clone if the receive queue of the raw socket is full. Signed-off-by:
Eric Dumazet <edumazet@google.com> Reviewed-by:
Willem de Bruijn <willemb@google.com> Link: https://lore.kernel.org/r/20240307163020.2524409-1-edumazet@google.comSigned-off-by:
-
Eric Dumazet authored
There is no point cloning an skb and having to free the clone if the receive queue of the raw socket is full. Signed-off-by:
-
Ido Schimmel authored
The only error that can happen during a nexthop dump is insufficient space in the skb caring the netlink messages (EMSGSIZE). If this happens and some messages were already filled in, the nexthop code returns the skb length to signal the netlink core that more objects need to be dumped. After commit b5a89915 ("netlink: handle EMSGSIZE errors in the core") there is no need to handle this error in the nexthop code as it is now handled in the core. Simplify the code and simply return the error to the core. No regressions in nexthop tests: # ./fib_nexthops.sh Tests passed: 234 Tests failed: 0 Signed-off-by:
Ido Schimmel <idosch@nvidia.com> Reviewed-by:
David Ahern <dsahern@kernel.org> Link: https://lore.kernel.org/r/20240307154727.3555462-1-idosch@nvidia.comSigned-off-by:
-
Eric Dumazet authored
Similar to skb_unref(), add skb_data_unref() to save an expensive atomic operation (and cache line dirtying) when last reference on shinfo->dataref is released. I saw this opportunity on hosts with RAW sockets accidentally bound to UDP protocol, forcing an skb_clone() on all received packets. These RAW sockets had their receive queue full, so all clone packets were immediately dropped. When UDP recvmsg() consumes later the original skb, skb_release_data() is hitting atomic_sub_return() quite badly, because skb->clone has been set permanently. Note that this patch helps TCP TX performance, because TCP stack also use (fast) clones. This means that at least one of the two packets (the main skb or its clone) will no longer have to perform this atomic operation in skb_release_data(). Signed-off-by:
-
Jakub Kicinski authored
Merge tag 'wireless-next-2024-03-08' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next Kalle Valo says: ==================== wireless-next patches for v6.9 The fourth "new features" pull request for v6.9 with changes both in stack and in drivers. The theme in this pull request is to fix sparse warnings but we still have some left in wireless subsystem. Otherwise quite normal. Major changes: rtw89 * NL80211_EXT_FEATURE_SCAN_RANDOM_SN support * NL80211_EXT_FEATURE_SET_SCAN_DWELL support rtw88 * support for more rtw8811cu and rtw8821cu devices mt76 * mt76x2u: add Netgear WNDA3100v3 USB * mt7915: newer ADIE version support * mt7925: radio temperature sensor support * mt7996: remove GCMP IGTK offload * tag 'wireless-next-2024-03-08' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next: (125 commits) wifi: rtw89: wow: move release offload packet earlier for WoWLAN mode wifi: rtw89: wow: set security engine options for 802.11ax chips only wifi: rtw89: update suspend/resume for different generation wifi: rtw89: wow: update config mac function with different generation wifi: rtw89: update DMA function with different generation wifi: rtw89: wow: update WoWLAN status register for different generation wifi: rtw89: wow: update WoWLAN reason register for different chips wifi: brcm80211: handle pmk_op allocation failure wifi: rtw89: coex: Add coexistence policy to decrease WiFi packet CRC-ERR wifi: rtw89: coex: When Bluetooth not available don't set power/gain wifi: rtw89: coex: add return value to ensure H2C command is success or not wifi: rtw89: coex: Reorder H2C command index to align with firmware wifi: rtw89: coex: add BTC ctrl_info version 7 and related logic wifi: rtw89: coex: add init_info H2C command format version 7 wifi: rtw89: 8922a: add coexistence helpers of SW grant wifi: rtw89: mac: add coexistence helpers {cfg/get}_plt wifi: cw1200: restore endian swapping wifi: wlcore: sdio: Rate limit wl12xx_sdio_raw_{read,write}() failures warns wifi: rtlwifi: Remove rtl_intf_ops.read_efuse_byte wifi: rtw88: 8821c: Fix false alarm count ... ==================== Link: https://lore.kernel.org/r/20240308100429.B8EA2C433F1@smtp.kernel.orgSigned-off-by:
-
Luiz Augusto von Dentz authored
This fixes the following error caused by hci_conn being freed while hcy_acl_create_conn_sync is pending: ================================================================== BUG: KASAN: slab-use-after-free in hci_acl_create_conn_sync+0xa7/0x2e0 Write of size 2 at addr ffff888002ae0036 by task kworker/u3:0/848 CPU: 0 PID: 848 Comm: kworker/u3:0 Not tainted 6.8.0-rc6-g2ab3e8d6 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-1.fc38 04/01/2014 Workqueue: hci0 hci_cmd_sync_work Call Trace: <TASK> dump_stack_lvl+0x21/0x70 print_report+0xce/0x620 ? preempt_count_sub+0x13/0xc0 ? __virt_addr_valid+0x15f/0x310 ? hci_acl_create_conn_sync+0xa7/0x2e0 kasan_report+0xdf/0x110 ? hci_acl_create_conn_sync+0xa7/0x2e0 hci_acl_create_conn_sync+0xa7/0x2e0 ? __pfx_hci_acl_create_conn_sync+0x10/0x10 ? __pfx_lock_release+0x10/0x10 ? __pfx_hci_acl_create_conn_sync+0x10/0x10 hci_cmd_sync_work+0x138/0x1c0 process_one_work+0x405/0x800 ? __pfx_lock_acquire+0x10/0x10 ? __pfx_process_one_work+0x10/0x10 worker_thread+0x37b/0x670 ? __pfx_worker_thread+0x10/0x10 kthread+0x19b/0x1e0 ? kthread+0xfe/0x1e0 ? __pfx_kthread+0x10/0x10 ret_from_fork+0x2f/0x50 ? __pfx_kthread+0x10/0x10 ret_from_fork_asm+0x1a/0x30 </TASK> Allocated by task 847: kasan_save_stack+0x33/0x60 kasan_save_track+0x14/0x30 __kasan_kmalloc+0x8f/0xa0 hci_conn_add+0xc6/0x970 hci_connect_acl+0x309/0x410 pair_device+0x4fb/0x710 hci_sock_sendmsg+0x933/0xef0 sock_write_iter+0x2c3/0x2d0 do_iter_readv_writev+0x21a/0x2e0 vfs_writev+0x21c/0x7b0 do_writev+0x14a/0x180 do_syscall_64+0x77/0x150 entry_SYSCALL_64_after_hwframe+0x6c/0x74 Freed by task 847: kasan_save_stack+0x33/0x60 kasan_save_track+0x14/0x30 kasan_save_free_info+0x3b/0x60 __kasan_slab_free+0xfa/0x150 kfree+0xcb/0x250 device_release+0x58/0xf0 kobject_put+0xbb/0x160 hci_conn_del+0x281/0x570 hci_conn_hash_flush+0xfc/0x130 hci_dev_close_sync+0x336/0x960 hci_dev_close+0x10e/0x140 hci_sock_ioctl+0x14a/0x5c0 sock_ioctl+0x58a/0x5d0 __x64_sys_ioctl+0x480/0xf60 do_syscall_64+0x77/0x150 entry_SYSCALL_64_after_hwframe+0x6c/0x74 Fixes: 45340097 ("Bluetooth: hci_conn: Only do ACL connections sequentially") Signed-off-by:
Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
-
Frédéric Danis authored
According to Section 1.2 of Core Specification Supplement Part A the complete or short name strings are defined as utf8s, which should not include the trailing NULL for variable length array as defined in Core Specification Vol1 Part E Section 2.9.3. Removing the trailing NULL allows PTS to retrieve the random address based on device name, e.g. for SM/PER/KDU/BV-02-C, SM/PER/KDU/BV-08-C or GAP/BROB/BCST/BV-03-C. Fixes: f61851f6 ("Bluetooth: Fix append max 11 bytes of name to scan rsp data") Signed-off-by:
Frédéric Danis <frederic.danis@collabora.com> Signed-off-by:
-
David S. Miller authored
Jijie Shao says: ==================== There are some bugfix for the HNS3 ethernet driver There are some bugfix for the HNS3 ethernet driver ==================== Signed-off-by: -
Jian Shen authored
Add checking for vf id of mailbox, in order to avoid array out-of-bounds risk. Signed-off-by:
Jian Shen <shenjian15@huawei.com> Signed-off-by:
Jijie Shao <shaojijie@huawei.com> Reviewed-by:
Sunil Goutham <sgoutham@marvell.com> Signed-off-by:
-
Jie Wang authored
Currently, the mac port is fixed to configured as full dplex mode in hclge_mac_init() when driver initialization or reset restore. Users may change the mode to half duplex with ethtool, so it may cause the user configuration dropped after reset. To fix it, don't change the duplex mode when resetting. Fixes: 2d03eacc ("net: hns3: Only update mac configuation when necessary") Signed-off-by:
Jie Wang <wangjie125@huawei.com> Signed-off-by:
-
Peiyang Wang authored
The cmdq reset command times out when all VFs are enabled and the queue is full. The hardware processing time exceeds the timeout set by the driver. In order to avoid the above extreme situations, the driver extends the reset timeout to 1 second. Signed-off-by:
Peiyang Wang <wangpeiyang1@huawei.com> Signed-off-by:
-
Jijie Shao authored
When the tc is removed during reset, hns3 driver will return a errcode. But kernel ignores this errcode, As a result, the driver status is inconsistent with the kernel status. This patch retains the deletion status when the deletion fails and continues to delete after the reset to ensure that the status of the driver is consistent with that of kernel. Signed-off-by:
-
Yonglong Liu authored
The HIP08 devices does not register the ptp devices, so the hdev->ptp is NULL, but the hardware can receive 1588 messages, and set the HNS3_RXD_TS_VLD_B bit, so, if match this case, the access of hdev->ptp->flags will cause a kernel crash: [ 5888.946472] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000018 [ 5888.946475] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000018 ... [ 5889.266118] pc : hclge_ptp_get_rx_hwts+0x40/0x170 [hclge] [ 5889.272612] lr : hclge_ptp_get_rx_hwts+0x34/0x170 [hclge] [ 5889.279101] sp : ffff800012c3bc50 [ 5889.283516] x29: ffff800012c3bc50 x28: ffff2040002be040 [ 5889.289927] x27: ffff800009116484 x26: 0000000080007500 [ 5889.296333] x25: 0000000000000000 x24: ffff204001c6f000 [ 5889.302738] x23: ffff204144f53c00 x22: 0000000000000000 [ 5889.309134] x21: 0000000000000000 x20: ffff204004220080 [ 5889.315520] x19: ffff204144f53c00 x18: 0000000000000000 [ 5889.321897] x17: 0000000000000000 x16: 0000000000000000 [ 5889.328263] x15: 0000004000140ec8 x14: 0000000000000000 [ 5889.334617] x13: 0000000000000000 x12: 00000000010011df [ 5889.340965] x11: bbfeff4d22000000 x10: 0000000000000000 [ 5889.347303] x9 : ffff800009402124 x8 : 0200f78811dfbb4d [ 5889.353637] x7 : 2200000000191b01 x6 : ffff208002a7d480 [ 5889.359959] x5 : 0000000000000000 x4 : 0000000000000000 [ 5889.366271] x3 : 0000000000000000 x2 : 0000000000000000 [ 5889.372567] x1 : 0000000000000000 x0 : ffff20400095c080 [ 5889.378857] Call trace: [ 5889.382285] hclge_ptp_get_rx_hwts+0x40/0x170 [hclge] [ 5889.388304] hns3_handle_bdinfo+0x324/0x410 [hns3] [ 5889.394055] hns3_handle_rx_bd+0x60/0x150 [hns3] [ 5889.399624] hns3_clean_rx_ring+0x84/0x170 [hns3] [ 5889.405270] hns3_nic_common_poll+0xa8/0x220 [hns3] [ 5889.411084] napi_poll+0xcc/0x264 [ 5889.415329] net_rx_action+0xd4/0x21c [ 5889.419911] __do_softirq+0x130/0x358 [ 5889.424484] irq_exit+0x134/0x154 [ 5889.428700] __handle_domain_irq+0x88/0xf0 [ 5889.433684] gic_handle_irq+0x78/0x2c0 [ 5889.438319] el1_irq+0xb8/0x140 [ 5889.442354] arch_cpu_idle+0x18/0x40 [ 5889.446816] default_idle_call+0x5c/0x1c0 [ 5889.451714] cpuidle_idle_call+0x174/0x1b0 [ 5889.456692] do_idle+0xc8/0x160 [ 5889.460717] cpu_startup_entry+0x30/0xfc [ 5889.465523] secondary_start_kernel+0x158/0x1ec [ 5889.470936] Code: 97ffab78 f9411c14 91408294 f9457284 (f9400c80) [ 5889.477950] SMP: stopping secondary CPUs [ 5890.514626] SMP: failed to stop secondary CPUs 0-69,71-95 [ 5890.522951] Starting crashdump kernel... Fixes: 0bf5eb78 ("net: hns3: add support for PTP") Signed-off-by:
Yonglong Liu <liuyonglong@huawei.com> Signed-off-by:
-
Hao Lan authored
When the hilink version is H60, the serdes serial loopback test is not supported. This patch add hilink version detection. When the version is H60, the serdes serial loopback test will be disable. Signed-off-by:
Hao Lan <lanhao@huawei.com> Signed-off-by:
-
Hao Lan authored
The hisilicon device now supports a new 200G link interface, which query from firmware in a new bit. Therefore, the HCLGE_SUPPORT_200G_R4_BIT capability bit has been added. The HCLGE_SUPPORT_200G_BIT has been renamed as HCLGE_SUPPORT_200G_R4_EXT_BIT, and the firmware has extended support for this mode. Fixes: ae6f010c ("net: hns3: add support for 200G device") Signed-off-by:
-
Jijie Shao authored
In hns3_dcbnl_ieee_delapp, should check ieee_delapp not ieee_setapp. This path fix the wrong judgment. Fixes: 0ba22bcb ("net: hns3: add support config dscp map to tc") Signed-off-by:
-
David S. Miller authored
Shannon Nelson says: ==================== ionic: putting ionic on a diet Building on the performance work done in the previous patchset [Link] https://lore.kernel.org/netdev/20240229193935.14197-1-shannon.nelson@amd.com/ this patchset puts the ionic driver on a diet, decreasing the memory requirements per queue, and simplifies a few more bits of logic. We trimmed the queue management structs and gained some ground, but the most savings came from trimming the individual buffer descriptors. The original design used a single generic buffer descriptor for Tx, Rx and Adminq needs, but the Rx and Adminq descriptors really don't need all the info that the Tx descriptors track. By splitting up the descriptor types we can significantly reduce the descriptor sizes for Rx and Adminq use. There is a small reduction in the queue management structs, saving about 3 cachelines per queuepair: ionic_qcq: Before: /* size: 2176, cachelines: 34, members: 23 */ After: /* size: 2048, cachelines: 32, members: 23 */ We also remove an array of completion descriptor pointers, or about 8 Kbytes per queue. But the biggest savings came from splitting the desc_info struct into queue specific structs and trimming out what was unnecessary. Before: ionic_desc_info: /* size: 496, cachelines: 8, members: 10 */ After: ionic_tx_desc_info: /* size: 496, cachelines: 8, members: 6 */ ionic_rx_desc_info: /* size: 224, cachelines: 4, members: 2 */ ionic_admin_desc_info: /* size: 8, cachelines: 1, members: 1 */ In a 64 core host the ionic driver will default to 64 queuepairs of 1024 descriptors for Rx, 1024 for Tx, and 80 for Adminq and Notifyq. The total memory usage for 64 queues: Before: 65 * sizeof(ionic_qcq) 141,440 + 64 * 1024 * sizeof(ionic_desc_info) 32,505,856 + 64 * 1024 * sizeof(ionic_desc_info) 32,505,856 + 64 * 1024 * 2 * sizeof(ionic_qc_info) 16,384 + 1 * 80 * sizeof(ionic_desc_info) 39,690 ---------- 65,201,038 After: 65 * sizeof(ionic_qcq) 133,120 + 64 * 1024 * sizeof(ionic_tx_desc_info) 32,505,856 + 64 * 1024 * sizeof(ionic_rx_desc_info) 14,680,064 + (removed) 0 + 1 * 80 * sizeof(ionic_admin desc_info) 640 ---------- 47,319,680 This saves us approximately 18 Mbytes per port in a 64 core machine, a 28% savings in our memory needs. In addition, this improves our simple single thread / single queue iperf case on a 9100 MTU connection from 86.7 to 95 Gbits/sec. ==================== Signed-off-by: -
Shannon Nelson authored
When possible, keep the stats struct references strictly in the error handling blocks and out of the fastpath. Reviewed-by:
Brett Creeley <brett.creeley@amd.com> Signed-off-by:
Shannon Nelson <shannon.nelson@amd.com> Signed-off-by:
-
Shannon Nelson authored
Fix up a couple of small dma_addr handling issues - don't double-count dma-map-err stat in ionic_tx_map_skb() or ionic_xdp_post_frame() - return 0 on error from both ionic_tx_map_single() and ionic_tx_map_frag() and check for !dma_addr in ionic_tx_map_skb() and ionic_xdp_post_frame() - be sure to unmap buf_info[0] in ionic_tx_map_skb() error path - don't assign rx buf->dma_addr until error checked in ionic_rx_page_alloc() - remove unnecessary dma_addr_t casts Reviewed-by: -
Shannon Nelson authored
We call ionic_rx_page_alloc() only on existing buf_info structs from ionic_rx_fill(). There's no need for the additional NULL test. Reviewed-by:
-
Shannon Nelson authored
A simple change to the struct ionic_queue layout removes some unnecessary padding and saves us a cacheline in the struct ionic_qcq layout. struct ionic_queue { Before: /* size: 256, cachelines: 4, members: 29 */ After: /* size: 192, cachelines: 3, members: 29 */ struct ionic_qcq { Before: /* size: 2112, cachelines: 33, members: 23 */ After: /* size: 2048, cachelines: 32, members: 23 */ Reviewed-by: -
Shannon Nelson authored
Rearange a few fields for better cache use and to put the flags field up into the first cacheline rather than the last. struct ionic_qcq Before: /* size: 2176, cachelines: 34, members: 23 */ After: /* size: 2112, cachelines: 33, members: 23 */ Reviewed-by: -
Shannon Nelson authored
Remove the idev field from ionic_queue, which saves us a bit of space, and add it into ionic_cq where there's room within some cacheline padding. Use this pointer rather than doing a multi level reference from lif->ionic. Suggested-by:
Neel Patel <npatel2@amd.com> Reviewed-by:
-
Shannon Nelson authored
The existing ionic_rx_frags() code is a bit of a mess and can be cleaned up by unrolling the first frag/header setup from the loop, then reworking the do-while-loop into a for-loop. We rename the function to a more descriptive ionic_rx_build_skb(). We also change a couple of related variable names for readability. Reviewed-by:
-
Shannon Nelson authored
Since the AdminQ clean is a simple action called from only one place, fold it back into the service routine. Reviewed-by:
-
Shannon Nelson authored
Make desc_info structure specific to the queue type, which allows us to cut down the Rx and AdminQ descriptor sizes by not including all the fields needed for the Tx desriptors. Before: struct ionic_desc_info { /* size: 464, cachelines: 8, members: 6 */ After: struct ionic_tx_desc_info { /* size: 464, cachelines: 8, members: 6 */ struct ionic_rx_desc_info { /* size: 224, cachelines: 4, members: 2 */ struct ionic_admin_desc_info { /* size: 8, cachelines: 1, members: 1 */ Suggested-by: -
Shannon Nelson authored
With a little simple math we don't need another struct array to find the completion structs, so we can remove the ionic_cq_info altogether. This doesn't really save anything in the ionic_cq since it gets padded out to the cacheline, but it does remove the parallel array allocation of 8 * num_descriptors, or about 8 Kbytes per queue in a default configuration. Suggested-by:
-
Shannon Nelson authored
By reworking the queue service routines to have their own servicing loops we can remove the cb pointer from desc_info to save another 8 bytes per descriptor, This simplifies some of the queue handling indirection and makes the code a little easier to follow, and keeps service code in one place rather than jumping between code files. struct ionic_desc_info Before: /* size: 472, cachelines: 8, members: 7 */ After: /* size: 464, cachelines: 8, members: 6 */ Suggested-by:
-
Shannon Nelson authored
Move the AdminQ and NotifyQ queue handling to ionic_main.c with the rest of the adminq code. Suggested-by:
-
Shannon Nelson authored
Now that we're not using desc_info pointers mapped in every q we can simplify and drop the unnecessary utility functions. Reviewed-by:
-
Shannon Nelson authored
Remove the struct pointers from desc_info to use less space. Instead of pointers in every desc_info to its descriptor, we can use the queue descriptor index to find the individual desc, desc_info, and sgl structs in their parallel arrays. struct ionic_desc_info Before: /* size: 496, cachelines: 8, members: 10 */ After: /* size: 472, cachelines: 8, members: 7 */ Suggested-by:
-
