Created attachment 341 [details] vmcore-dmesg 系统频繁 crash 系统版本：Anolis OS release 8.4 kernel 版本：4.19.91-26.an8.x86_64 vmcore-dmesg 如附件所示
Created attachment 342 [details] vmcore-dmesg
[ 187.572539] watchdog: BUG: soft lockup - CPU#0 stuck for 22s! [swapper/0:0] [ 187.572940] Modules linked in: cls_bpf(E) sch_ingress(E) xt_TPROXY(E) nf_tproxy_ipv6(E) nf_tproxy_ipv4(E) xt_CT(E) veth(E) bpfilter(E) xt_socket(E) nf_socket_ipv4(E) nf_socket_ipv6(E) ip6table_filter(E) ip6table_raw(E) ip6table_mangle(E) ip6_tables(E) iptable_filter(E) iptable_raw(E) iptable_mangle(E) iptable_nat(E) nft_chain_route_ipv6(E) ip6t_MASQUERADE(E) nft_chain_route_ipv4(E) nft_chain_nat_ipv6(E) nf_nat_ipv6(E) ipt_MASQUERADE(E) xt_conntrack(E) xt_comment(E) nft_counter(E) xt_mark(E) nft_compat(E) nft_chain_nat_ipv4(E) nf_nat_ipv4(E) nf_nat(E) nf_tables(E) nfnetlink(E) dm_mod(E) 8021q(E) sch_netem(E) garp(E) mrp(E) overlay(E) rpcrdma(E) intel_rapl_msr(E) intel_rapl_common(E) sunrpc(E) rdma_ucm(E) ib_uverbs(E) ib_srpt(E) ib_isert(E) isst_if_common(E) iscsi_target_mod(E) target_core_mod(E) ib_iser(E) [ 187.572961] rdma_cm(E) skx_edac(E) iw_cm(E) ib_cm(E) nfit(E) libiscsi(E) scsi_transport_iscsi(E) x86_pkg_temp_thermal(E) intel_powerclamp(E) coretemp(E) ipmi_ssif(E) iTCO_wdt(E) iTCO_vendor_support(E) kvm_intel(E) bnxt_re(E) kvm(E) irqbypass(E) crct10dif_pclmul(E) crc32_pclmul(E) ib_core(E) ghash_clmulni_intel(E) pcbc(E) joydev(E) mei_me(E) aesni_intel(E) glue_helper(E) pcspkr(E) mousedev(E) i2c_i801(E) lpc_ich(E) mei(E) ioatdma(E) dca(E) wmi(E) ipmi_si(E) ipmi_devintf(E) pcc_cpufreq(E) ipmi_msghandler(E) acpi_pad(E) acpi_power_meter(E) sch_fq_codel(E) bridge(E) stp(E) llc(E) toa(OE) nf_conntrack(E) nf_defrag_ipv6(E) nf_defrag_ipv4(E) ip_tables(E) xfs(E) libcrc32c(E) sd_mod(E) sg(E) nvme(E) crc32c_intel(E) nvme_core(E) i2c_algo_bit(E) drm_kms_helper(E) syscopyarea(E) sysfillrect(E) sysimgblt(E) [ 187.572985] fb_sys_fops(E) ttm(E) ahci(E) drm(E) libahci(E) bnxt_en(E) i2c_core(E) libata(E) [ 187.572990] CPU: 0 PID: 0 Comm: swapper/0 Kdump: loaded Tainted: G W OE 4.19.91-26.an8.x86_64 #1 [ 187.572991] Hardware name: Inspur SA5212M5/YZMB-00882-10C, BIOS 4.1.8 05/08/2020 [ 187.572992] RIP: 0010:__netdev_pick_tx+0x19c/0x220 [ 187.572994] Code: 48 8d 04 83 0f b7 98 54 08 00 00 44 0f b7 a8 52 08 00 00 e9 e5 fe ff ff 83 e8 01 0f b7 d0 44 39 ea 73 07 01 d8 e9 08 ff ff ff <44> 29 ea 44 39 ea 73 f8 89 d0 01 d8 e9 f7 fe ff ff 4c 89 e6 4c 89 [ 187.572995] RSP: 0018:ffff963cff403bf0 EFLAGS: 00000246 ORIG_RAX: ffffffffffffff13 [ 187.572996] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000001 [ 187.572997] RDX: 0000000000000000 RSI: ffff963cf299a800 RDI: ffff963cf236e000 [ 187.572997] RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000 [ 187.572998] R10: ffff963cff403b30 R11: 0000000000000008 R12: ffff963cf299a800 [ 187.572999] R13: 0000000000000000 R14: 00000000ffffffff R15: ffff963cf299a800 [ 187.572999] FS: 0000000000000000(0000) GS:ffff963cff400000(0000) knlGS:0000000000000000 [ 187.573000] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 187.573001] CR2: 0000561975b9e478 CR3: 0000006d4c20a003 CR4: 00000000007706f0 [ 187.573001] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 187.573002] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 [ 187.573002] PKRU: 55555554 [ 187.573002] Call Trace: [ 187.573004] <IRQ> [ 187.573009] netdev_pick_tx+0xa3/0xb0 [ 187.573011] __dev_queue_xmit+0x1f1/0x8a0 [ 187.573014] __bpf_redirect+0x97/0x2c0 [ 187.573016] __netif_receive_skb_core+0xca9/0xf90 [ 187.573021] ? ip_local_deliver+0x42/0xd0 [ 187.573024] ? inet_gro_receive+0x243/0x2a0 [ 187.573025] __netif_receive_skb_one_core+0x26/0x50 [ 187.573027] netif_receive_skb_internal+0x32/0xc0 [ 187.573028] napi_gro_receive+0xb8/0xe0 [ 187.573034] bnxt_rx_pkt+0xb14/0xfa0 [bnxt_en] [ 187.573039] ? alloc_skb_with_frags+0xe0/0x1a0 [ 187.573041] bnxt_poll+0xf5/0x810 [bnxt_en] [ 187.573046] ? tick_sched_handle.isra.5+0x60/0x60 [ 187.573047] net_rx_action+0x139/0x360 [ 187.573052] __do_softirq+0xd2/0x2b5 [ 187.573056] irq_exit+0xc8/0x100 [ 187.573058] do_IRQ+0x7f/0xe0 [ 187.573060] common_interrupt+0xf/0xf [ 187.573061] </IRQ> [ 187.573065] RIP: 0010:cpuidle_enter_state+0xb9/0x320 [ 187.573066] Code: e8 4c f6 9a ff 80 7c 24 0b 00 74 17 9c 58 0f 1f 44 00 00 f6 c4 02 0f 85 3b 02 00 00 31 ff e8 6e 2f a1 ff fb 66 0f 1f 44 00 00 <48> b8 ff ff ff ff f3 01 00 00 48 2b 1c 24 ba ff ff ff 7f 48 39 c3 [ 187.573066] RSP: 0018:ffffffff90203e88 EFLAGS: 00000246 ORIG_RAX: ffffffffffffffd3 [ 187.573067] RAX: ffff963cff422640 RBX: 00000026521cd89c RCX: 000000000000001f [ 187.573068] RDX: 00000026521cd89c RSI: 000000003158af9d RDI: 0000000000000000 [ 187.573069] RBP: ffffcf22bf606270 R08: 0000000000000002 R09: 0000000000021e80 [ 187.573069] R10: 0112b141c92c930e R11: ffff963cff421624 R12: 0000000000000001 [ 187.573070] R13: ffffffff90396f78 R14: 0000000000000001 R15: 0000000000000000 [ 187.573073] ? cpuidle_enter_state+0x94/0x320 [ 187.573077] do_idle+0x210/0x250 [ 187.573078] cpu_startup_entry+0x5f/0x70 [ 187.573083] start_kernel+0x518/0x523 [ 187.573088] secondary_startup_64+0xb5/0xc0 [ 187.573090] Kernel panic - not syncing: softlockup: hung tasks [ 187.573484] CPU: 0 PID: 0 Comm: swapper/0 Kdump: loaded Tainted: G W OEL 4.19.91-26.an8.x86_64 #1 [ 187.574184] Hardware name: Inspur SA5212M5/YZMB-00882-10C, BIOS 4.1.8 05/08/2020 [ 187.574875] Call Trace: [ 187.575257] <IRQ> [ 187.575639] dump_stack+0x66/0x90 [ 187.576027] panic+0xf9/0x25c [ 187.576413] ? startup_64+0x1/0x30 [ 187.576798] ? startup_64+0x30/0x30 [ 187.577187] watchdog_timer_fn.cold.2+0x16/0x16 [ 187.577578] ? report_softlockup+0x1a0/0x1a0 [ 187.577968] __hrtimer_run_queues+0xf0/0x260 [ 187.578356] hrtimer_interrupt+0x100/0x220 [ 187.578745] ? aperfmperf_snapshot_khz+0x67/0x90 [ 187.579136] smp_apic_timer_interrupt+0x6a/0x140 [ 187.579526] apic_timer_interrupt+0xf/0x20 [ 187.579914] RIP: 0010:__netdev_pick_tx+0x19c/0x220 [ 187.580305] Code: 48 8d 04 83 0f b7 98 54 08 00 00 44 0f b7 a8 52 08 00 00 e9 e5 fe ff ff 83 e8 01 0f b7 d0 44 39 ea 73 07 01 d8 e9 08 ff ff ff <44> 29 ea 44 39 ea 73 f8 89 d0 01 d8 e9 f7 fe ff ff 4c 89 e6 4c 89 [ 187.581330] RSP: 0018:ffff963cff403bf0 EFLAGS: 00000246 ORIG_RAX: ffffffffffffff13 [ 187.582020] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000001 [ 187.582419] RDX: 0000000000000000 RSI: ffff963cf299a800 RDI: ffff963cf236e000 [ 187.582816] RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000 [ 187.583214] R10: ffff963cff403b30 R11: 0000000000000008 R12: ffff963cf299a800 [ 187.583612] R13: 0000000000000000 R14: 00000000ffffffff R15: ffff963cf299a800 [ 187.584011] ? apic_timer_interrupt+0xa/0x20 [ 187.584401] netdev_pick_tx+0xa3/0xb0 [ 187.584788] __dev_queue_xmit+0x1f1/0x8a0 [ 187.585177] __bpf_redirect+0x97/0x2c0 [ 187.585564] __netif_receive_skb_core+0xca9/0xf90 [ 187.585954] ? ip_local_deliver+0x42/0xd0 [ 187.586342] ? inet_gro_receive+0x243/0x2a0 [ 187.586731] __netif_receive_skb_one_core+0x26/0x50 [ 187.587121] netif_receive_skb_internal+0x32/0xc0 [ 187.587512] napi_gro_receive+0xb8/0xe0 [ 187.587900] bnxt_rx_pkt+0xb14/0xfa0 [bnxt_en] [ 187.588291] ? alloc_skb_with_frags+0xe0/0x1a0 [ 187.588682] bnxt_poll+0xf5/0x810 [bnxt_en] [ 187.589071] ? tick_sched_handle.isra.5+0x60/0x60 [ 187.589462] net_rx_action+0x139/0x360 [ 187.589850] __do_softirq+0xd2/0x2b5 [ 187.590237] irq_exit+0xc8/0x100 [ 187.590626] do_IRQ+0x7f/0xe0 [ 187.591011] common_interrupt+0xf/0xf [ 187.591397] </IRQ> [ 187.591777] RIP: 0010:cpuidle_enter_state+0xb9/0x320 [ 187.592168] Code: e8 4c f6 9a ff 80 7c 24 0b 00 74 17 9c 58 0f 1f 44 00 00 f6 c4 02 0f 85 3b 02 00 00 31 ff e8 6e 2f a1 ff fb 66 0f 1f 44 00 00 <48> b8 ff ff ff ff f3 01 00 00 48 2b 1c 24 ba ff ff ff 7f 48 39 c3 [ 187.593190] RSP: 0018:ffffffff90203e88 EFLAGS: 00000246 ORIG_RAX: ffffffffffffffd3 [ 187.593882] RAX: ffff963cff422640 RBX: 00000026521cd89c RCX: 000000000000001f [ 187.594283] RDX: 00000026521cd89c RSI: 000000003158af9d RDI: 0000000000000000 [ 187.594681] RBP: ffffcf22bf606270 R08: 0000000000000002 R09: 0000000000021e80 [ 187.595079] R10: 0112b141c92c930e R11: ffff963cff421624 R12: 0000000000000001 [ 187.595478] R13: ffffffff90396f78 R14: 0000000000000001 R15: 0000000000000000 [ 187.595878] ? cpuidle_enter_state+0x94/0x320 [ 187.596268] do_idle+0x210/0x250 [ 187.596653] cpu_startup_entry+0x5f/0x70 [ 187.597040] start_kernel+0x518/0x523 [ 187.597430] secondary_startup_64+0xb5/0xc0 [ 187.703923] Kernel Offset: 0xe000000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff)
[ 187.573009] netdev_pick_tx+0xa3/0xb0 [ 187.573011] __dev_queue_xmit+0x1f1/0x8a0 [ 187.573014] __bpf_redirect+0x97/0x2c0 [ 187.573016] __netif_receive_skb_core+0xca9/0xf90 RX softirq中在搞发包，__bpf_redirect()，看这个现场是网络TC相关的，是不是配置了TC相关的bpf程序，建议去掉观察一下。
两次都打在同一个位置 netdev_pick_tx+0xa3/0xb0 上，在这里存在锁竞争之类的场景。
Created attachment 343 [details] bnxt-testko-v1 bnxt驱动-测试版本v1
根据dmesg的日志: > ... > [ 187.572992] RIP: 0010:__netdev_pick_tx+0x19c/0x220 > ... 结合kernel-debuginfo，定位到代码为net/core/dev.c:2858，代码逻辑如下： > static u16 skb_tx_hash(const struct net_device *dev, > const struct net_device *sb_dev, > struct sk_buff *skb) > { > u32 hash; > u16 qoffset = 0; > u16 qcount = dev->real_num_tx_queues; > > if (dev->num_tc) { > u8 tc = netdev_get_prio_tc_map(dev, skb->priority); > > qoffset = sb_dev->tc_to_txq[tc].offset; > qcount = sb_dev->tc_to_txq[tc].count; > } > > if (skb_rx_queue_recorded(skb)) { > hash = skb_get_rx_queue(skb); > while (unlikely(hash >= qcount)) > hash -= qcount; > return hash + qoffset; > } > > return (u16) reciprocal_scale(skb_get_hash(skb), qcount) + qoffset; > } 其中第2858行，对应代码"hash -= qcount;" 猜测是因为qcount为0，导致函数一直死循环，最终触发softlockup。 根据代码drivers/net/ethernet/broadcom/bnxt/bnxt.c:5857，即函数bnxt_setup_msix()的逻辑: > static void bnxt_setup_msix(struct bnxt *bp) > { > const int len = sizeof(bp->irq_tbl[0].name); > struct net_device *dev = bp->dev; > int tcs, i; > > tcs = netdev_get_num_tc(dev); > if (tcs > 1) { > int i, off, count; > > for (i = 0; i < tcs; i++) { > count = bp->tx_nr_rings_per_tc; > off = i * count; > netdev_set_tc_queue(dev, i, count, off); > } > } > ... > } 当tcs为1时，sb_dev->tc_to_txq[tc].count将不会被设置，其值为默认值0，导致skb_tx_hash()中发生死循环。 目前在bnxt_setup_msix()中手动添加逻辑，使得当tcs为1时，也将sb_dev->tc_to_txq[tc].count设置为1.即： > static void bnxt_setup_msix(struct bnxt *bp) > { > const int len = sizeof(bp->irq_tbl[0].name); > struct net_device *dev = bp->dev; > int tcs, i; > > tcs = netdev_get_num_tc(dev); > if (tcs > 1) { > int i, off, count; > > for (i = 0; i < tcs; i++) { > count = bp->tx_nr_rings_per_tc; > off = i * count; > netdev_set_tc_queue(dev, i, count, off); > } > } else if (tcs == 1) { > netdev_set_tc_queue(dev, 0, 1, 0); > } > ... > } 重新编译的ko已上传至该bugzilla附件：bnxt-testko-v1 麻烦帮忙验证一下
Created attachment 344 [details] bnxt-testko-v2
使用了更加合理的修改方式，上传了v2版本，见附件bnxt-testko-v2 请使用该ko进行测试。 修改方式如下： > static void bnxt_setup_msix(struct bnxt *bp) > { > const int len = sizeof(bp->irq_tbl[0].name); > struct net_device *dev = bp->dev; > int tcs, i; > > tcs = netdev_get_num_tc(dev); > if (tcs > 1) { > int i, off, count; > > for (i = 0; i < tcs; i++) { > count = bp->tx_nr_rings_per_tc; > off = i * count; > netdev_set_tc_queue(dev, i, count, off); > } > } else if (tcs == 1) { > WARN_ONCE(bp->tx_nr_rings_per_tc == 0, "unexpected bp->tx_nr_rings_per_tc: 0 !!!"); > netdev_set_tc_queue(dev, 0, bp->tx_nr_rings_per_tc, 0); > } > .. > }
测试方法： 1. 使用浏览器下载bnxt-testko-v2.tar.gz （请不要直接wget，亲测有些问题） 2. 上传该文件至测试机器 3. 解包获取bnxt_en.ko: tar xf bnxt-testko-v2.tar.gz 4. 备份原有ko： mv /usr/lib/modules/4.19.91-26.an8.x86_64/kernel/drivers/net/ethernet/broadcom/bnxt/bnxt_en.ko /usr/lib/modules/4.19.91-26.an8.x86_64/kernel/drivers/net/ethernet/broadcom/bnxt/bnxt_en.ko.bak 5. 覆盖原有ko： mv bnxt_en.ko /usr/lib/modules/4.19.91-26.an8.x86_64/kernel/drivers/net/ethernet/broadcom/bnxt/bnxt_en.ko 6. reboot进行测试
还在 crash ， vmcore-dmesg 见附件，分别是两次的 crash 信息
Created attachment 345 [details] vmcore-dmesg-8-2
Created attachment 346 [details] vmcore-dmesg-8-2
> if (skb_rx_queue_recorded(skb)) { > hash = skb_get_rx_queue(skb); > while (unlikely(hash >= qcount)) > hash -= qcount; > return hash + qoffset; > } > 是否在 qcount 为 0 时，直接 BUGON crash 掉
(In reply to maqiao_mq from comment #9) > 测试方法： > 1. 使用浏览器下载bnxt-testko-v2.tar.gz （请不要直接wget，亲测有些问题） > 2. 上传该文件至测试机器 > 3. 解包获取bnxt_en.ko: tar xf bnxt-testko-v2.tar.gz > 4. 备份原有ko： > mv > /usr/lib/modules/4.19.91-26.an8.x86_64/kernel/drivers/net/ethernet/broadcom/ > bnxt/bnxt_en.ko > /usr/lib/modules/4.19.91-26.an8.x86_64/kernel/drivers/net/ethernet/broadcom/ > bnxt/bnxt_en.ko.bak > > 5. 覆盖原有ko： > mv bnxt_en.ko > /usr/lib/modules/4.19.91-26.an8.x86_64/kernel/drivers/net/ethernet/broadcom/ > bnxt/bnxt_en.ko > > 6. reboot进行测试 建议在驱动版本号上做个修改，或者打印，确保使用的是新驱动做的验证
Created attachment 348 [details] url of 4.19.91-cbp.git.5b8703df0.an8.x86_64
通过分析vmcore，确认是qcount为0导致的。 此外，分析下来，发生宕机的机器上，bnxt_en.ko还是老的，新的ko并没有被安装到内核中，导致之前修改的代码不起作用，猜测是在替换ko后没有调用dracut -f刷新initramfs导致的，这一步在上文的替换步骤中忘了。。。 我们回合了upstream上两个修复该问题的补丁，出了kernel的rpm包，麻烦直接使用该rpm包进行验证。 rpm包的下载链接，见附件：url of 4.19.91-cbp.git.5b8703df0.an8.x86_64 使用方法： 1. 下载以下3个包： - kernel-core-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm - kernel-modules-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm - kernel-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm 2. 执行安装命令： rpm -ivh kernel-core-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm kernel-modules-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm kernel-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm --force 3. reboot重启即可 附：回合补丁列表： - bnxt_en: Fix TC queue mapping. - net: Prevent infinite while loop in skb_tx_hash()
# rpm -ivh kernel-core-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm kernel-modules-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm kernel-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm --force Verifying... ################################# [100%] Preparing... ################################# [100%] Updating / installing... 1:kernel-core-4.19.91-cbp.git.5b870################################# [ 33%] 2:kernel-modules-4.19.91-cbp.git.5b################################# [ 67%] 3:kernel-4.19.91-cbp.git.5b8703df0.################################# [100%] Unable to decompress /boot/initramfs-4.19.91-cbp.git.5b8703df0.an8.x86_64.img: Unknown format
Created attachment 351 [details] url(with md5sum) of 4.19.91-cbp.git.5b8703df0.an8.x86_64
(In reply to jni_ni from comment #17) > # rpm -ivh kernel-core-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm > kernel-modules-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm > kernel-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm --force > Verifying... ################################# > [100%] > Preparing... ################################# > [100%] > Updating / installing... > 1:kernel-core-4.19.91-cbp.git.5b870################################# [ > 33%] > 2:kernel-modules-4.19.91-cbp.git.5b################################# [ > 67%] > 3:kernel-4.19.91-cbp.git.5b8703df0.################################# > [100%] > Unable to decompress > /boot/initramfs-4.19.91-cbp.git.5b8703df0.an8.x86_64.img: Unknown format [root@mq-an8 ~]# rpm -ivh kernel-core-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm kernel-modules-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm kernel-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm --force Verifying... ################################# [100%] Preparing... ################################# [100%] Updating / installing... 1:kernel-core-4.19.91-cbp.git.5b870################################# [ 33%] 2:kernel-modules-4.19.91-cbp.git.5b################################# [ 67%] 3:kernel-4.19.91-cbp.git.5b8703df0.################################# [100%] [root@mq-an8 ~]# cat /etc/os-release NAME="Anolis OS" VERSION="8.4" ID="anolis" ID_LIKE="rhel fedora centos" VERSION_ID="8.4" PLATFORM_ID="platform:an8" PRETTY_NAME="Anolis OS 8.4" ANSI_COLOR="0;31" HOME_URL="https://openanolis.cn/" [root@mq-an8 ~]# uname -a Linux mq-an8 4.19.91-26.an8.x86_64 #1 SMP Tue May 24 13:10:09 CST 2022 x86_64 x86_64 x86_64 GNU/Linux 在ECS上未复现该问题，有可能是文件在下载时损坏了，请比对一下md5值。 md5值见附件：url(with md5sum) of 4.19.91-cbp.git.5b8703df0.an8.x86_64
重启了服务器，重新下载了文件，并校对过MD5，还是报 Unable to decompress /boot/initramfs-4.19.91-cbp.git.5b8703df0.an8.x86_64.img: Unknown format tmp]# md5sum kernel-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm 98c6d43505a7960f80a9f24814bf16a8 kernel-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm tmp]# md5sum kernel-core-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm 95d4e85e0c26911062e09e4b0233d3f5 kernel-core-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm tmp]# md5sum kernel-modules-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm 29b2ec03399be4ba68904b8424901a9b kernel-modules-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm tmp]# rpm -ivh --force kernel-modules-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm Verifying... ################################# [100%] Preparing... ################################# [100%] Updating / installing... 1:kernel-modules-4.19.91-cbp.git.5b################################# [100%] tmp]# rpm -ivh --force kernel-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm Verifying... ################################# [100%] Preparing... ################################# [100%] Updating / installing... 1:kernel-4.19.91-cbp.git.5b8703df0.################################# [100%] tmp]# rpm -ivh --force kernel-core-4.19.91-cbp.git.5b8703df0.an8.x86_64.rpm Verifying... ################################# [100%] Preparing... ################################# [100%] Updating / installing... 1:kernel-core-4.19.91-cbp.git.5b870################################# [100%] Unable to decompress /boot/initramfs-4.19.91-cbp.git.5b8703df0.an8.x86_64.img: Unknown format
tmp]# cat /etc/os-release NAME="Anolis OS" VERSION="8.4" ID="anolis" ID_LIKE="rhel fedora centos" VERSION_ID="8.4" PLATFORM_ID="platform:an8" PRETTY_NAME="Anolis OS 8.4" ANSI_COLOR="0;31" HOME_URL="https://openanolis.cn/" tmp]# uname -a Linux 10.72.46.4 4.19.91-26.an8.x86_64 #1 SMP Tue May 24 13:10:09 CST 2022 x86_64 x86_64 x86_64 GNU/Linux
(In reply to maqiao_mq from comment #9) > 测试方法： > 1. 使用浏览器下载bnxt-testko-v2.tar.gz （请不要直接wget，亲测有些问题） > 2. 上传该文件至测试机器 > 3. 解包获取bnxt_en.ko: tar xf bnxt-testko-v2.tar.gz > 4. 备份原有ko： > mv > /usr/lib/modules/4.19.91-26.an8.x86_64/kernel/drivers/net/ethernet/broadcom/ > bnxt/bnxt_en.ko > /usr/lib/modules/4.19.91-26.an8.x86_64/kernel/drivers/net/ethernet/broadcom/ > bnxt/bnxt_en.ko.bak > > 5. 覆盖原有ko： > mv bnxt_en.ko > /usr/lib/modules/4.19.91-26.an8.x86_64/kernel/drivers/net/ethernet/broadcom/ > bnxt/bnxt_en.ko > > 6. reboot进行测试 rpm安装的问题一时半会不好排查，可以先用下面这个方法workaround： 在第5-6步之间，执行dracut -f刷新initramfs，然后再reboot重试
已合入修复代码 https://gitee.com/anolis/cloud-kernel/pulls/624