aboutsummaryrefslogtreecommitdiff
path: root/drivers/net/ethernet/broadcom
AgeCommit message (Collapse)AuthorFilesLines
2024-02-07net: bcmgenet: remove eee_enabled/eee_active in bcmgenet_get_eee()Russell King (Oracle)2-7/+6
bcmgenet_get_eee() sets edata->eee_active and edata->eee_enabled from its own copy, and then calls phy_ethtool_get_eee() which in turn will call genphy_c45_ethtool_get_eee(). genphy_c45_ethtool_get_eee() will overwrite eee_enabled and eee_active with its own interpretation from the PHYs settings and negotiation result. Therefore, setting these members in bcmgenet_get_eee() is redundant, and can be removed. This also makes priv->eee.eee_active unnecessary, so remove this and use a local variable where appropriate. Reviewed-by: Andrew Lunn <[email protected]> Reviewed-by: Florian Fainelli <[email protected]> Signed-off-by: Russell King (Oracle) <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-02-06tg3: convert EEE handling to use linkmode bitmapsHeiner Kallweit1-13/+14
Convert EEE handling to use linkmode bitmaps. This prepares for removing the legacy bitmaps from struct ethtool_keee. No functional change intended. Note: The change to mii_eee_cap1_mod_linkmode_t(tp->eee.advertised, val) in tg3_phy_autoneg_cfg() isn't completely obvious, but it doesn't change the current functionality. Signed-off-by: Heiner Kallweit <[email protected]> Reviewed-by: Andrew Lunn <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-02-01Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/netJakub Kicinski1-1/+1
Cross-merge networking fixes after downstream PR. No conflicts or adjacent changes. Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-31ethtool: add suffix _u32 to legacy bitmap members of struct ethtool_keeeHeiner Kallweit4-28/+28
This is in preparation of using the existing names for linkmode bitmaps. Suggested-by: Andrew Lunn <[email protected]> Signed-off-by: Heiner Kallweit <[email protected]> Reviewed-by: Andrew Lunn <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-31ethtool: replace struct ethtool_eee with a new struct ethtool_keee on kernel ↵Heiner Kallweit10-26/+26
side In order to pass EEE link modes beyond bit 32 to userspace we have to complement the 32 bit bitmaps in struct ethtool_eee with linkmode bitmaps. Therefore, similar to ethtool_link_settings and ethtool_link_ksettings, add a struct ethtool_keee. In a first step it's an identical copy of ethtool_eee. This patch simply does a s/ethtool_eee/ethtool_keee/g for all users. No functional change intended. Suggested-by: Andrew Lunn <[email protected]> Signed-off-by: Heiner Kallweit <[email protected]> Reviewed-by: Andrew Lunn <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-26bnx2x: Fix firmware version string character countsKees Cook3-12/+13
A potential string truncation was reported in bnx2x_fill_fw_str(), when a long bp->fw_ver and a long phy_fw_ver might coexist, but seems unlikely with real-world hardware. Use scnprintf() to indicate the intent that truncations are tolerated. While reading this code, I found a collection of various buffer size counting issues. None looked like they might lead to a buffer overflow with current code (the small buffers are 20 bytes and might only ever consume 10 bytes twice with a trailing %NUL). However, early truncation (due to a %NUL in the middle of the string) might be happening under likely rare conditions. Regardless fix the formatters and related functions: - Switch from a separate strscpy() to just adding an additional "%s" to the format string that immediately follows it in bnx2x_fill_fw_str(). - Use sizeof() universally instead of using unbound defines. - Fix bnx2x_7101_format_ver() and bnx2x_null_format_ver() to report the number of characters written, not including the trailing %NUL (as already done with the other firmware formatting functions). - Require space for at least 1 byte in bnx2x_get_ext_phy_fw_version() for the trailing %NUL. - Correct the needed buffer size in bnx2x_3_seq_format_ver(). Reported-by: kernel test robot <[email protected]> Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/ Cc: Ariel Elior <[email protected]> Cc: Sudarsana Kalluru <[email protected]> Cc: Manish Chopra <[email protected]> Signed-off-by: Kees Cook <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-26bnxt_en: Make PTP timestamp HWRM more silentBreno Leitao1-1/+1
commit 056bce63c469 ("bnxt_en: Make PTP TX timestamp HWRM query silent") changed a netdev_err() to netdev_WARN_ONCE(). netdev_WARN_ONCE() is it generates a kernel WARNING, which is bad, for the following reasons: * You do not a kernel warning if the firmware queries are late * In busy networks, timestamp query failures fairly regularly * A WARNING message doesn't bring much value, since the code path is clear. (This was discussed in-depth in [1]) Transform the netdev_WARN_ONCE() into a netdev_warn_once(), and print a more well-behaved message, instead of a full WARN(). bnxt_en 0000:67:00.0 eth0: TS query for TX timer failed rc = fffffff5 [1] Link: https://lore.kernel.org/all/ZbDj%[email protected]/ Signed-off-by: Breno Leitao <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Reviewed-by: Michael Chan <[email protected]> Fixes: 056bce63c469 ("bnxt_en: Make PTP TX timestamp HWRM query silent") Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-24net: fill in MODULE_DESCRIPTION()s for Broadcom bgmacBreno Leitao5-0/+5
W=1 builds now warn if module is built without a MODULE_DESCRIPTION(). Add descriptions to the Broadcom iProc GBit driver. Signed-off-by: Breno Leitao <[email protected]> Acked-by: Florian Fainelli <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-19bnxt_en: Fix possible crash after creating sw mqprio TCsMichael Chan5-11/+17
The driver relies on netdev_get_num_tc() to get the number of HW offloaded mqprio TCs to allocate and free TX rings. This won't work and can potentially crash the system if software mqprio or taprio TCs have been setup. netdev_get_num_tc() will return the number of software TCs and it may cause the driver to allocate or free more TX rings that it should. Fix it by adding a bp->num_tc field to store the number of HW offload mqprio TCs for the device. Use bp->num_tc instead of netdev_get_num_tc(). This fixes a crash like this: BUG: kernel NULL pointer dereference, address: 0000000000000000 PGD 42b8404067 P4D 0 Oops: 0000 [#1] PREEMPT SMP NOPTI CPU: 120 PID: 8661 Comm: ifconfig Kdump: loaded Tainted: G OE 5.18.16 #1 Hardware name: Lenovo ThinkSystem SR650 V3/SB27A92818, BIOS ESE114N-2.12 04/25/2023 RIP: 0010:bnxt_hwrm_cp_ring_alloc_p5+0x10/0x90 [bnxt_en] Code: 41 5c 41 5d 41 5e c3 cc cc cc cc 41 8b 44 24 08 66 89 03 eb c6 e8 b0 f1 7d db 0f 1f 44 00 00 41 56 41 55 41 54 55 48 89 fd 53 <48> 8b 06 48 89 f3 48 81 c6 28 01 00 00 0f b6 96 13 ff ff ff 44 8b RSP: 0018:ff65907660d1fa88 EFLAGS: 00010202 RAX: 0000000000000010 RBX: ff4dde1d907e4980 RCX: f400000000000000 RDX: 0000000000000010 RSI: 0000000000000000 RDI: ff4dde1d907e4980 RBP: ff4dde1d907e4980 R08: 000000000000000f R09: 0000000000000000 R10: ff4dde5f02671800 R11: 0000000000000008 R12: 0000000088888889 R13: 0500000000000000 R14: 00f0000000000000 R15: ff4dde5f02671800 FS: 00007f4b126b5740(0000) GS:ff4dde9bff600000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000000 CR3: 000000416f9c6002 CR4: 0000000000771ee0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400 PKRU: 55555554 Call Trace: <TASK> bnxt_hwrm_ring_alloc+0x204/0x770 [bnxt_en] bnxt_init_chip+0x4d/0x680 [bnxt_en] ? bnxt_poll+0x1a0/0x1a0 [bnxt_en] __bnxt_open_nic+0xd2/0x740 [bnxt_en] bnxt_open+0x10b/0x220 [bnxt_en] ? raw_notifier_call_chain+0x41/0x60 __dev_open+0xf3/0x1b0 __dev_change_flags+0x1db/0x250 dev_change_flags+0x21/0x60 devinet_ioctl+0x590/0x720 ? avc_has_extended_perms+0x1b7/0x420 ? _copy_from_user+0x3a/0x60 inet_ioctl+0x189/0x1c0 ? wp_page_copy+0x45a/0x6e0 sock_do_ioctl+0x42/0xf0 ? ioctl_has_perm.constprop.0.isra.0+0xbd/0x120 sock_ioctl+0x1ce/0x2e0 __x64_sys_ioctl+0x87/0xc0 do_syscall_64+0x59/0x90 ? syscall_exit_work+0x103/0x130 ? syscall_exit_to_user_mode+0x12/0x30 ? do_syscall_64+0x69/0x90 ? exc_page_fault+0x62/0x150 Fixes: c0c050c58d84 ("bnxt_en: New Broadcom ethernet driver.") Reviewed-by: Damodharam Ammepalli <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-19bnxt_en: Prevent kernel warning when running offline self testMichael Chan1-0/+3
We call bnxt_half_open_nic() to setup the chip partially to run loopback tests. The rings and buffers are initialized normally so that we can transmit and receive packets in loopback mode. That means page pool buffers are allocated for the aggregation ring just like the normal case. NAPI is not needed because we are just polling for the loopback packets. When we're done with the loopback tests, we call bnxt_half_close_nic() to clean up. When freeing the page pools, we hit a WARN_ON() in page_pool_unlink_napi() because the NAPI state linked to the page pool is uninitialized. The simplest way to avoid this warning is just to initialize the NAPIs during half open and delete the NAPIs during half close. Trying to skip the page pool initialization or skip linking of NAPI during half open will be more complicated. This fix avoids this warning: WARNING: CPU: 4 PID: 46967 at net/core/page_pool.c:946 page_pool_unlink_napi+0x1f/0x30 CPU: 4 PID: 46967 Comm: ethtool Tainted: G S W 6.7.0-rc5+ #22 Hardware name: Dell Inc. PowerEdge R750/06V45N, BIOS 1.3.8 08/31/2021 RIP: 0010:page_pool_unlink_napi+0x1f/0x30 Code: 90 90 90 90 90 90 90 90 90 90 90 0f 1f 44 00 00 48 8b 47 18 48 85 c0 74 1b 48 8b 50 10 83 e2 01 74 08 8b 40 34 83 f8 ff 74 02 <0f> 0b 48 c7 47 18 00 00 00 00 c3 cc cc cc cc 66 90 90 90 90 90 90 RSP: 0018:ffa000003d0dfbe8 EFLAGS: 00010246 RAX: ff110003607ce640 RBX: ff110010baf5d000 RCX: 0000000000000008 RDX: 0000000000000000 RSI: ff110001e5e522c0 RDI: ff110010baf5d000 RBP: ff11000145539b40 R08: 0000000000000001 R09: ffffffffc063f641 R10: ff110001361eddb8 R11: 000000000040000f R12: 0000000000000001 R13: 000000000000001c R14: ff1100014553a080 R15: 0000000000003fc0 FS: 00007f9301c4f740(0000) GS:ff1100103fd00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007f91344fa8f0 CR3: 00000003527cc005 CR4: 0000000000771ef0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 PKRU: 55555554 Call Trace: <TASK> ? __warn+0x81/0x140 ? page_pool_unlink_napi+0x1f/0x30 ? report_bug+0x102/0x200 ? handle_bug+0x44/0x70 ? exc_invalid_op+0x13/0x60 ? asm_exc_invalid_op+0x16/0x20 ? bnxt_free_ring.isra.123+0xb1/0xd0 [bnxt_en] ? page_pool_unlink_napi+0x1f/0x30 page_pool_destroy+0x3e/0x150 bnxt_free_mem+0x441/0x5e0 [bnxt_en] bnxt_half_close_nic+0x2a/0x40 [bnxt_en] bnxt_self_test+0x21d/0x450 [bnxt_en] __dev_ethtool+0xeda/0x2e30 ? native_queued_spin_lock_slowpath+0x17f/0x2b0 ? __link_object+0xa1/0x160 ? _raw_spin_unlock_irqrestore+0x23/0x40 ? __create_object+0x5f/0x90 ? __kmem_cache_alloc_node+0x317/0x3c0 ? dev_ethtool+0x59/0x170 dev_ethtool+0xa7/0x170 dev_ioctl+0xc3/0x530 sock_do_ioctl+0xa8/0xf0 sock_ioctl+0x270/0x310 __x64_sys_ioctl+0x8c/0xc0 do_syscall_64+0x3e/0xf0 entry_SYSCALL_64_after_hwframe+0x6e/0x76 Fixes: 294e39e0d034 ("bnxt: hook NAPIs to page pools") Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Ajit Khaparde <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-19bnxt_en: Fix RSS table entries calculation for P5_PLUS chipsMichael Chan2-6/+14
The existing formula used in the driver to calculate the number of RSS table entries is to round up the number of RX rings to the next integer multiples of 64 (e.g. 64, 128, 192, ..). This is incorrect. The valid values supported by the chip are 64, 128, 256, 512 only (power of 2 starting from 64). When the number of RX rings is greater than 128, the entry size will likely be wrong. Firmware will round down the invalid value (e.g. 192 rounded down to 128) provided by the driver, causing some RSS rings to not receive any packets. We already have an existing function bnxt_calc_nr_ring_pages() to do this calculation. Use it in bnxt_get_nr_rss_ctxs() to calculate the number of RSS contexts correctly for P5_PLUS chips. Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Fixes: 7b3af4f75b81 ("bnxt_en: Add RSS support for 57500 chips.") Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-19bnxt_en: Fix memory leak in bnxt_hwrm_get_rings()Michael Chan1-2/+3
bnxt_hwrm_get_rings() can abort and return error when there are not enough ring resources. It aborts without releasing the HWRM DMA buffer, causing a dma_pool_destroy warning when the driver is unloaded: bnxt_en 0000:99:00.0: dma_pool_destroy bnxt_hwrm, 000000005b089ba8 busy Fixes: f1e50b276d37 ("bnxt_en: Fix trimming of P5 RX and TX rings") Reviewed-by: Somnath Kotur <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-19bnxt_en: Wait for FLR to complete during probeMichael Chan1-0/+5
The first message to firmware may fail if the device is undergoing FLR. The driver has some recovery logic for this failure scenario but we must wait 100 msec for FLR to complete before proceeding. Otherwise the recovery will always fail. Fixes: ba02629ff6cb ("bnxt_en: log firmware status on firmware init failure") Reviewed-by: Damodharam Ammepalli <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-08bnxt_en: Fix RCU locking for ntuple filters in bnxt_rx_flow_steer()Michael Chan1-1/+1
Similar to the previous patch, RCU locking was released too early in bnxt_rx_flow_steer(). Fix it to unlock after reading fltr->base.sw_id to guarantee that fltr won't be freed while we are still reading it. Fixes: cb5bdd292dc0 ("bnxt_en: Add bnxt_lookup_ntp_filter_from_idx() function") Reported-by: Simon Horman <[email protected]> Link: https://lore.kernel.org/netdev/[email protected]/ Signed-off-by: Michael Chan <[email protected]> Reviewed-by: Simon Horman <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-08bnxt_en: Fix RCU locking for ntuple filters in bnxt_srxclsrldel()Michael Chan1-10/+11
After looking up an ntuple filter from a RCU hash list, the rcu_read_unlock() call should be made after reading the structure, or after determining that the filter cannot age out (by aRFS). The existing code was calling rcu_read_unlock() too early in bnxt_srxclsrldel(). As suggested by Simon Horman, change the code to handle the error case of fltr_base not found in the if condition. The code looks cleaner this way. Fixes: 8d7ba028aa9a ("bnxt_en: Add support for ntuple filter deletion by ethtool.") Suggested-by: Simon Horman <[email protected]> Reported-by: Jakub Kicinski <[email protected]> Link: https://lore.kernel.org/netdev/[email protected]/ Signed-off-by: Michael Chan <[email protected]> Reviewed-by: Simon Horman <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-08bnxt_en: Remove unneeded variable in bnxt_hwrm_clear_vnic_filter()Michael Chan1-4/+1
After recent refactoring, this function doesn't return error any more. Remove the unneeded rc variable and change the function to void. The caller is not checking for the return value. Fixes: 96c9bedc755e ("bnxt_en: Refactor L2 filter alloc/free firmware commands.") Reported-by: kernel test robot <[email protected]> Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/ Signed-off-by: Michael Chan <[email protected]> Reviewed-by: Simon Horman <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-04Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/netJakub Kicinski2-3/+5
Cross-merge networking fixes after downstream PR. Conflicts: drivers/net/ethernet/broadcom/bnxt/bnxt.c e009b2efb7a8 ("bnxt_en: Remove mis-applied code from bnxt_cfg_ntp_filters()") 0f2b21477988 ("bnxt_en: Fix compile error without CONFIG_RFS_ACCEL") https://lore.kernel.org/all/[email protected]/ Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-04bnxt_en: Fix compile error without CONFIG_RFS_ACCELMichael Chan1-0/+2
Fix the following compile error: .../bnxt.c: In function 'bnxt_cfg_ntp_filters': .../bnxt.c:14077:37: error: implicit declaration of function 'rps_may_expire_flow' [-Werror=implicit-function-declaration] 14077 | if (rps_may_expire_flow(bp->dev, fltr->base.rxq, | ^~~~~~~~~~~~~~~~~~~ bnxt_cfg_ntp_filters() is only used when CONFIG_RFS_ACCEL is enabled. User configured ntuple filters are directly added and will not go through this function. Wrap the body of bnxt_cfg_ntp_filters() with CONFIG_RFS_ACCEL. Fixes: 59cde76f33fa ("bnxt_en: Refactor filter insertion logic in bnxt_rx_flow_steer().") Reported-by: Arnd Bergmann <[email protected]> Link: https://lore.kernel.org/netdev/[email protected]/ Reported-by: kernel test robot <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Signed-off-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-04bnxt_en: Remove mis-applied code from bnxt_cfg_ntp_filters()Michael Chan1-2/+2
The 2 lines to check for the BNXT_HWRM_PF_UNLOAD_SP_EVENT bit was mis-applied to bnxt_cfg_ntp_filters() and should have been applied to bnxt_sp_task(). Fixes: 19241368443f ("bnxt_en: Send PF driver unload notification to all VFs.") Reviewed-by: Andy Gospodarek <[email protected]> Signed-off-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02net: bcmgenet: Fix FCS generation for fragmented skbuffsAdrian Cinal1-1/+3
The flag DMA_TX_APPEND_CRC was only written to the first DMA descriptor in the TX path, where each descriptor corresponds to a single skbuff fragment (or the skbuff head). This led to packets with no FCS appearing on the wire if the kernel allocated the packet in fragments, which would always happen when using PACKET_MMAP/TPACKET (cf. tpacket_fill_skb() in net/af_packet.c). Fixes: 1c1008c793fa ("net: bcmgenet: add main driver file") Signed-off-by: Adrian Cinal <[email protected]> Acked-by: Doug Berger <[email protected]> Acked-by: Florian Fainelli <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2024-01-02bnxt_en: Add support for ntuple filter deletion by ethtool.Michael Chan1-0/+29
Add logic to delete a user specified ntuple filter from ethtool. Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Add support for ntuple filters added from ethtool.Michael Chan3-7/+201
Add support for adding user defined ntuple TCP/UDP filters. These filters are similar to aRFS filters except that they don't get aged. Source IP, destination IP, source port, or destination port can be unspecifed as wildcard. At least one of these tuples must be specifed. If a tuple is specified, the full mask must be specified. All ntuple related ethtool functions are now no longer compiled only for CONFIG_RFS_ACCEL. Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Add ntuple matching flags to the bnxt_ntuple_filter structure.Michael Chan3-50/+99
aRFS filters match all 5 tuples. User defined ntuple filters may specify some of the tuples as wildcards. To support that, we add the ntuple_flags to the bnxt_ntuple_filter struct to specify which tuple fields are to be matched. The matching tuple fields will then be passed to the firmware in bnxt_hwrm_cfa_ntuple_filter_alloc() to create the proper filter. Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Refactor ntuple filter removal logic in bnxt_cfg_ntp_filters().Michael Chan2-16/+22
Refactor the logic into a new function bnxt_del_ntp_filters(). The same call will be used when the user deletes an ntuple filter. The bnxt_hwrm_cfa_ntuple_filter_free() function to call fw to free the ntuple filter is exported so that the ethtool logic can call it. Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Reviewed-by: Simon Horman <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Refactor the hash table logic for ntuple filters.Michael Chan3-28/+59
Generalize the ethtool logic that walks the ntuple hash table now that we have the common bnxt_filter_base structure. This will allow the code to easily extend to cover user defined ntuple or ether filters. Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Refactor filter insertion logic in bnxt_rx_flow_steer().Michael Chan2-52/+41
Add a new function bnxt_insert_ntp_filter() to insert the ntuple filter into the hash table and other basic setup. We'll use this function to insert a user defined filter from ethtool. Also, export bnxt_lookup_ntp_filter_from_idx() and bnxt_get_ntp_filter_idx() for similar purposes. All ntuple related functions are now no longer compiled only for CONFIG_RFS_ACCEL Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Add new BNXT_FLTR_INSERTED flag to bnxt_filter_base struct.Michael Chan2-1/+11
Change the unused flag to BNXT_FLTR_INSERTED. To prepare for multiple pathways that an ntuple filter can be deleted, we add this flag. These filter structures can be retreived from the RCU hash table but only the caller that sees that the BNXT_FLTR_INSERTED flag is set can delete the filter structure and clear the flag under spinlock. Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Add bnxt_lookup_ntp_filter_from_idx() functionMichael Chan1-6/+20
Add the helper function to look up the ntuple filter from the hash index and use it in bnxt_rx_flow_steer(). The helper function will also be used by user defined ntuple filters in the next patches. Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Reviewed-by: Simon Horman <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Add function to calculate Toeplitz hashPavan Chebbi2-3/+108
For ntuple filters added by aRFS, the Toeplitz hash calculated by our NIC is available and is used to store the ntuple filter for quick retrieval. In the next patches, user defined ntuple filter support will be added and we need to calculate the same hash for these filters. The same hash function needs to be used so we can detect duplicates. Add the function bnxt_toeplitz() to calculate the Toeplitz hash for user defined ntuple filters. bnxt_toeplitz() uses the same Toeplitz key and the same key length as the NIC. bnxt_get_ntp_filter_idx() is added to return the hash index. For aRFS, the hash comes from the NIC. For user defined ntuple, we call bnxt_toeplitz() to calculate the hash index. Reviewed-by: Andy Gospodarek <[email protected]> Signed-off-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Refactor L2 filter alloc/free firmware commands.Michael Chan2-54/+112
Refactor the L2 filter alloc/free logic so that these filters can be added/deleted by the user. The bp->ntp_fltr_bmap allocated size is also increased to allow enough IDs for L2 filters. Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Re-structure the bnxt_ntuple_filter structure.Michael Chan2-26/+40
With the new bnxt_l2_filter structure, we can now re-structure the bnxt_ntuple_filter structure to point to the bnxt_l2_filter structure. We eliminate the L2 ether address info from the ntuple filter structure as we can get the information from the L2 filter structure. Note that the source L2 MAC address is no longer used. Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Reviewed-by: Simon Horman <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Add bnxt_l2_filter hash table.Michael Chan2-14/+191
The current driver only has an array of 4 additional L2 unicast addresses to support the netdev uc address list. Generalize and expand this infrastructure with an L2 address hash table so we can support an expanded list of unicast addresses (for bridges, macvlans, OVS, etc). The L2 hash table infrastructure will also allow more generalized n-tuple filter support. This patch creates the bnxt_l2_filter structure and the hash table. This L2 filter structure has the same bnxt_filter_base structure as used in the bnxt_ntuple_filter structure. All currently supported L2 filters will now have an entry in this new table. Note that L2 filters may be created for the VF. VF filters should not be freed when the PF goes down. Add some logic in bnxt_free_l2_filters() to allow keeping the VF filters or to free everything during rmmod. Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Reviewed-by: Simon Horman <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2024-01-02bnxt_en: Refactor bnxt_ntuple_filter structure.Michael Chan3-30/+44
This is in preparation to support user defined L2 (ether) filters, which will have many similarities with ntuple filters. Refactor bnxt_ntuple_filter structure to have a bnxt_filter_base structure that can be re-used by the L2 filters. Reviewed-by: Vasundhara Volam <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Reviewed-by: Simon Horman <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2023-12-21Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/netPaolo Abeni1-9/+2
Cross-merge networking fixes after downstream PR. Adjacent changes: drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c 23c93c3b6275 ("bnxt_en: do not map packet buffers twice") 6d1add95536b ("bnxt_en: Modify TX ring indexing logic.") tools/testing/selftests/net/Makefile 2258b666482d ("selftests: add vlan hw filter tests") a0bc96c0cd6e ("selftests: net: verify fq per-band packet limit") Signed-off-by: Paolo Abeni <[email protected]>
2023-12-15bnxt_en: do not map packet buffers twiceAndy Gospodarek1-9/+2
Remove double-mapping of DMA buffers as it can prevent page pool entries from being freed. Mapping is managed by page pool infrastructure and was previously managed by the driver in __bnxt_alloc_rx_page before allowing the page pool infrastructure to manage it. Fixes: 578fcfd26e2a ("bnxt_en: Let the page pool manage the DMA mapping") Reviewed-by: Somnath Kotur <[email protected]> Signed-off-by: Andy Gospodarek <[email protected]> Signed-off-by: Michael Chan <[email protected]> Reviewed-by: David Wei <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-14Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/netJakub Kicinski4-36/+42
Cross-merge networking fixes after downstream PR. Conflicts: drivers/net/ethernet/intel/iavf/iavf_ethtool.c 3a0b5a2929fd ("iavf: Introduce new state machines for flow director") 95260816b489 ("iavf: use iavf_schedule_aq_request() helper") https://lore.kernel.org/all/[email protected]/ drivers/net/ethernet/broadcom/bnxt/bnxt.c c13e268c0768 ("bnxt_en: Fix HWTSTAMP_FILTER_ALL packet timestamp logic") c2f8063309da ("bnxt_en: Refactor RX VLAN acceleration logic.") a7445d69809f ("bnxt_en: Add support for new RX and TPA_START completion types for P7") 1c7fd6ee2fe4 ("bnxt_en: Rename some macros for the P5 chips") https://lore.kernel.org/all/[email protected]/ drivers/net/ethernet/broadcom/bnxt/bnxt_ptp.c bd6781c18cb5 ("bnxt_en: Fix wrong return value check in bnxt_close_nic()") 84793a499578 ("bnxt_en: Skip nic close/open when configuring tstamp filters") https://lore.kernel.org/all/[email protected]/ drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c 3d7a3f2612d7 ("net/mlx5: Nack sync reset request when HotPlug is enabled") cecf44ea1a1f ("net/mlx5: Allow sync reset flow when BF MGT interface device is present") https://lore.kernel.org/all/[email protected]/ No adjacent changes. Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-13net: ethtool: pass a pointer to parameters to get/set_rxfh ethtool opsAhmed Zaki3-37/+38
The get/set_rxfh ethtool ops currently takes the rxfh (RSS) parameters as direct function arguments. This will force us to change the API (and all drivers' functions) every time some new parameters are added. This is part 1/2 of the fix, as suggested in [1]: - First simplify the code by always providing a pointer to all params (indir, key and func); the fact that some of them may be NULL seems like a weird historic thing or a premature optimization. It will simplify the drivers if all pointers are always present. - Then make the functions take a dev pointer, and a pointer to a single struct wrapping all arguments. The set_* should also take an extack. Link: https://lore.kernel.org/netdev/[email protected]/ [1] Suggested-by: Jakub Kicinski <[email protected]> Suggested-by: Jacob Keller <[email protected]> Signed-off-by: Ahmed Zaki <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Make PTP TX timestamp HWRM query silentPavan Chebbi1-3/+3
In a busy network, especially with flow control enabled, we may experience timestamp query failures fairly regularly. After a while, dmesg may be flooded with timestamp query failure error messages. Silence the error message from the low level hwrm function that sends the firmware message. Change netdev_err() to netdev_WARN_ONCE() if this FW call ever fails. Signed-off-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Skip nic close/open when configuring tstamp filtersPavan Chebbi2-20/+11
We don't have to close and open the nic to make sure we have valid rx timestamps. Once we have the timestamp filter applied to the HW and the timestamp_fld_format bit is cleared in the rx completion and the timestamp is non-zero, we can be sure that rx timestamp is valid data. Skip close/open when we set any timestamp filter. Reviewed-by: Andy Gospodarek <[email protected]> Signed-off-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Add support for UDP GSO on 5760X chipsMichael Chan2-3/+19
The new 5760X chips supports UDP GSO. Tested using udpgso_bench_tx. Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Somnath Kotur <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: add rx_filter_miss extended statsDamodharam Ammepalli1-0/+1
rx_filter_miss counter is newly added to the rx_port_stats_ext stats structure for newer chips. Newer firmware will return the structure size that includes this counter. Add this entry to the bnxt_port_stats_ext_arr array and the ethtool -S code will pick up this counter if it is supported. Signed-off-by: Damodharam Ammepalli <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Configure UDP tunnel TPAMichael Chan2-0/+34
On the new P7 chips, TPA for tunnel packets can be independently enabled for each VNIC. The default TPA configuration should not include UDP tunnels because the UDP ports for these tunnels are not known yet. The chip should not aggregate these UDP tunneled packets using default UDP ports until the ports are known. Add a new function bnxt_hwrm_vnic_update_tunl_tpa() to enable VXLAN and Geneve TPA if the corresponding UDP ports are known. Reviewed-by: Ajit Khaparde <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Add support for VXLAN GPEMichael Chan2-5/+40
Add a new bnxt_udp_tunnels_p7 struct to support the new P7 chips that can parse VXLAN GPE packets. Add VXLAN GPE tunnel type handling to the .set_port() and .unset_port() functions. .ndo_features_check() is also enhanced to support VXLAN GPE which may encapsulate inner IP packets instead of ethernet packets. Reviewed-by: Damodharam Ammepalli <[email protected]> Reviewed-by: Ajit Khaparde <[email protected]> Reviewed-by: Somnath Kotur <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Use proper TUNNEL_DST_PORT_ALLOC* commandsMichael Chan1-2/+2
In bnxt_udp_tunnel_set_port(), use the proper ALLOC commands instead of the FREE commands for correctness. The ALLOC and FREE commands happen to be identical so this is just a cosmetic fix for correctness. Reviewed-by: Somnath Kotur <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Allocate extra QP backing store memory when RoCE FW reports itSelvin Xavier1-2/+12
The Fast QP modify destroy RoCE feature requires additional QP entries in QP context backing store. FW reports the extra count to be allocated during backing store query. Use this value and allocate extra memory. Note that this works for both the V1 and V1 backing store FW APIs. Signed-off-by: Selvin Xavier <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Support TX coalesced completion on 5760X chipsMichael Chan2-2/+12
TX coalesced completions are supported on newer chips to provide one TX completion record for multiple TX packets up to the sq_cons_idx in the completion record. This method saves PCIe bandwidth by reducing the number of TX completions. Only very minor changes are now required to support this mode with the new framework that handles TX completions based on the consumer indices. Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Prevent TX timeout with a very small TX ringMichael Chan1-1/+4
If xmit_more condition is true, the driver may set the TX_BD_FLAGS_NO_CMPL flag. If after this packet, the TX ring can no longer hold a packet with maximum fragments, we will stop the TX queue. When this happens, we must clear the TX_BD_FLAGS_NO_CMPL flag on the last packet or there will be no completion and cause TX timeout. Fixes: c1056a59aee1 ("bnxt_en: Optimize xmit_more TX path") Reviewed-by: Somnath Kotur <[email protected]> Reviewed-by: Andy Gospodarek <[email protected]> Reviewed-by: Hongguang Gao <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Fix TX ring indexing logicMichael Chan2-2/+2
Two spots were missed when modifying the TX ring indexing logic. The use of unmasked TX index in bnxt_tx_int() will cause unnecessary __bnxt_tx_int() calls. The same issue in bnxt_tx_int_xdp() can result in illegal array index. Fixes: 6d1add95536b ("bnxt_en: Modify TX ring indexing logic.") Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Fix AGG ring check logic in bnxt_check_rings()Somnath Kotur1-3/+3
_bnxt_get_max_rings() that is invoked in bnxt_check_rings() already accounts for the AGG ring(s) and gives a max value based on that. Increasing for AGG rings before calling _bnxt_get_max_rings() will result in checking for twice the number of rings than required and it can fail. Fix it by adjusting for AGG rings after calling _bnxt_get_max_rings(). Fixes: f5b29c6afe36 ("bnxt_en: Add helper to get the number of CP rings required for TX rings") Signed-off-by: Somnath Kotur <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
2023-12-12bnxt_en: Fix trimming of P5 RX and TX ringsMichael Chan1-5/+11
The recent commit to trim the RX and TX rings on P5 chips by assigning each with max CP rings divided by 2 is not correct. Max CP rings divided by 2 may be bigger than the original RX or TX and would lead to failure. In other words, we may be checking for increased RX/TX rings than required and it may fail. Fix it by calling __bnxt_trim_rings() instead that would properly trim RX and TX without the possibility of increasing their values. Fixes: f5b29c6afe36 ("bnxt_en: Add helper to get the number of CP rings required for TX rings") Reviewed-by: Pavan Chebbi <[email protected]> Signed-off-by: Michael Chan <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jakub Kicinski <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.