【hns-0416】net: hns3: fix kernel crash when unload VF while it is being reset

[Fred Kimmy]

[Bug Description]
When fully configure VLANs for a VF, then unload the VF while
triggering a reset to PF, will cause a kernel crash because the
irq is already uninit.

[78895.670263] ------------[ cut here ]------------
[78895.674867] kernel BUG at drivers/pci/msi.c:378!
[78895.679467] Internal error: Oops - BUG: 0 [#1] SMP
[78895.684239] Modules linked in: binfmt_misc hns_roce_hw_v2 ib_uverbs ib_umad ib_core hclgevf 8021q garp mrp stp llc nls_iso8859_1 dm_multipath scsi_dh_rdac scsi_dh_emc scsi_dh_alua ipmi_ssif joydev input_leds arm_spe_pmu efi_pstore hisi_sec2 authenc hisi_hpre hisi_zip hisi_qm uacce hisi_dma uio_pdrv_genirq uio hisi_trng_v2 acpi_ipmi ipmi_si ipmi_devintf ipmi_msghandler cppc_cpufreq sch_fq_codel ip_tables x_tables autofs4 btrfs blake2b_generic raid10 raid456 async_raid6_recov async_memcpy async_pq async_xor async_tx xor xor_neon raid6_pq libcrc32c raid1 raid0 multipath linear ses enclosure realtek hid_generic hibmc_drm drm_vram_helper drm_ttm_helper ttm i2c_algo_bit drm_kms_helper syscopyarea crct10dif_ce sysfillrect ghash_ce sysimgblt sha2_ce fb_sys_fops hisi_sas_v3_hw cec sha256_arm64 hns3 rc_core usbhid hisi_sas_main sha1_ce hclge hid ixgbe drm libsas xhci_pci hnae3 xhci_pci_renesas ahci megaraid_sas scsi_transport_sas xfrm_algo mdio spi_dw_mmio gpio_dwapb spi_dw aes_neon_bs
[78895.684337] aes_neon_blk aes_ce_blk crypto_simd cryptd aes_ce_cipher [last unloaded: ib_core]
[78895.778896] CPU: 8 PID: 1671944 Comm: kworker/8:1 Not tainted 5.11.0-27-generic #29~20.04.1-Ubuntu
[78895.787813] Hardware name: Huawei TaiShan 2280 V2/BC82AMDC, BIOS 2280-V2 CS V5.B211.01 11/10/2021
[78895.796645] Workqueue: hclgevf hclgevf_service_task [hclgevf]
[78895.802374] pstate: 00400009 (nzcv daif +PAN -UAO -TCO BTYPE=--)
[78895.808354] pc : free_msi_irqs+0x188/0x1a8
[78895.812433] lr : free_msi_irqs+0x178/0x1a8
[78895.816511] sp : ffff800033eabba0
[78895.819811] x29: ffff800033eabba0 x28: 0000000000020c00
[78895.825100] x27: 00000000000000e1 x26: 000047c8bdaf3151
[78895.830388] x25: ffff0041c20e0430 x24: ffff0041c20e03f8
[78895.835677] x23: 0000000000027008 x22: ffff0040a1a48000
[78895.840965] x21: ffff0040a1a482f0 x20: 0000000000000000
[78895.846254] x19: ffff0040feac9580 x18: 0000000000000000
[78895.851542] x17: 0000000000000000 x16: ffffdef63b555010
[78895.856830] x15: 0000000000000000 x14: 0000000000000000
[78895.862117] x13: ffffffffffffff00 x12: ffffffffffffffff
[78895.867406] x11: 0000000000000040 x10: ffffdef63cd680a0
[78895.872694] x9 : ffffdef63aec8464 x8 : ffff20400745c250
[78895.877982] x7 : 0000000000000000 x6 : 0000000000000000
[78895.883270] x5 : ffff20400745c228 x4 : ffff20400745c298
[78895.888558] x3 : 0000000000000000 x2 : 0000000000000000
[78895.893845] x1 : 0000000000000549 x0 : 0000000000000001
[78895.899133] Call trace:
[78895.901569] free_msi_irqs+0x188/0x1a8
[78895.905300] pci_disable_msix+0xec/0x118
[78895.909207] pci_free_irq_vectors+0x20/0x38
[78895.913371] hclgevf_uninit_msi+0x48/0x60 [hclgevf]
[78895.918231] hclgevf_reset_rebuild+0x178/0x398 [hclgevf]
[78895.923520] hclgevf_reset_service_task+0x354/0x600 [hclgevf]
[78895.929242] hclgevf_service_task+0x1b8/0x2e0 [hclgevf]
[78895.934445] process_one_work+0x1fc/0x4d0
[78895.938437] worker_thread+0x148/0x510
[78895.942170] kthread+0xf4/0x120
[78895.945297] ret_from_fork+0x10/0x18
[78895.948859] Code: 72001c1f 54ffff00 a90363f7 f90023f9 (d4210000)
[78895.954925] ---[ end trace d83c0afbdc9d99a2 ]---
[78895.960560] ------------[ cut here ]------------

[Steps to Reproduce]
1)echo 1 > /sys/class/net/<ethx>/device/sriov_numvfs
2)ip link set dev <vf_name> up
3）config full vlan_id 0~4094
ip link add dev <vf_name>.<vlan_id> link <vf_name> type vlan id <vlan_id>
ifconfig <vf_name>.<vlan_id> up
4） enable VF spoofchk function
ip link set <pf_name> vf <vf_id> spoofchk on
5) resest vf and clear pf
echo 1 > /sys/class/net/<pf_name>/device/reset
echo 0 > /sys/class/net/<pf_name>/device/sriov_numvfs
6) resest vf and clear pf
echo 1 > /sys/class/net/<pf_name>/device/reset
echo 1 > /sys/class/net/<pf_name>/device/sriov_numvfs
7) re-test step 3

[Actual Results]
call trace

[Expected Results]
no call trace

[Reproducibility]

[Additional information]
(Firmware version, kernel version, affected hardware, etc. if required2022011211577):
OS:Ubuntu 20.04.3 LTS
DRV(driver version):Linux tx 5.11.0-27-generic #29~20.04.1-Ubuntu SMP Wed Aug 11 15:58:08 UTC 2021 aarch64 aarch64 aarch64 GNU/Linux

[Resolution]

net: hns3: fix kernel crash when unload VF while it is being reset

https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/drivers/net/ethernet/hisilicon?id=e140c7983e3054be0652bf914f4454f16c5520b0

[Ike Panhc (ikepanhc)]

e140c7983e30 <email address hidden> 2021-11-10 14:20:43 +0000 net: hns3: fix kernel crash when unload VF while it is being reset

This patch has been merged into mainline kernel since v5.16. I will try to backport to Ubuntu kernels.

[Ike Panhc (ikepanhc)]

This patch has been applied to Ubuntu 5.15 kernel since 5.15.0-12.12. When focal HWE kernel rolls to 5.15, this issue will be fixed.

[Ike Panhc (ikepanhc)]

focal HWE kernel has been rolled to 5.15.

$ rmadison linux-generic-hwe-20.04 | grep focal-updates
 linux-generic-hwe-20.04 | 5.13.0.1026.29~20.04.1 | focal-updates | riscv64
 linux-generic-hwe-20.04 | 5.15.0.41.44~20.04.13 | focal-updates | amd64, arm64, armhf, ppc64el, s390x