@@ -643,6 +643,11 @@ static void hv_arch_irq_unmask(struct irq_data *data)
pbus = pdev->bus;
hbus = container_of(pbus->sysdata, struct hv_pcibus_device, sysdata);
int_desc = data->chip_data;
+ if (!int_desc) {
+ dev_warn(&hbus->hdev->device, "%s() can not unmask irq %u\n",
+ __func__, data->irq);
+ return;
+ }
spin_lock_irqsave(&hbus->retarget_msi_interrupt_lock, flags);
@@ -1911,12 +1916,6 @@ static void hv_compose_msi_msg(struct irq_data *data, struct msi_msg *msg)
hv_pci_onchannelcallback(hbus);
spin_unlock_irqrestore(&channel->sched_lock, flags);
- if (hpdev->state == hv_pcichild_ejecting) {
- dev_err_once(&hbus->hdev->device,
- "the device is being ejected\n");
- goto enable_tasklet;
- }
-
udelay(100);
}
When the host tries to remove a PCI device, the host first sends a PCI_EJECT message to the guest, and the guest is supposed to gracefully remove the PCI device and send a PCI_EJECTION_COMPLETE message to the host; the host then sends a VMBus message CHANNELMSG_RESCIND_CHANNELOFFER to the guest (when the guest receives this message, the device is already unassigned from the guest) and the guest can do some final cleanup work; if the guest fails to respond to the PCI_EJECT message within one minute, the host sends the VMBus message CHANNELMSG_RESCIND_CHANNELOFFER and removes the PCI device forcibly. In the case of fast device addition/removal, it's possible that the PCI device driver is still configuring MSI-X interrupts when the guest receives the PCI_EJECT message; the channel callback calls hv_pci_eject_device(), which sets hpdev->state to hv_pcichild_ejecting, and schedules a work hv_eject_device_work(); if the PCI device driver is calling pci_alloc_irq_vectors() -> ... -> hv_compose_msi_msg(), we can break the while loop in hv_compose_msi_msg() due to the updated hpdev->state, and leave data->chip_data with its default value of NULL; later, when the PCI device driver calls request_irq() -> ... -> hv_irq_unmask(), the guest crashes in hv_arch_irq_unmask() due to data->chip_data being NULL. Fix the issue by not testing hpdev->state in the while loop: when the guest receives PCI_EJECT, the device is still assigned to the guest, and the guest has one minute to finish the device removal gracefully. We don't really need to (and we should not) test hpdev->state in the loop. Fixes: de0aa7b2f97d ("PCI: hv: Fix 2 hang issues in hv_compose_msi_msg()") Signed-off-by: Dexuan Cui <decui@microsoft.com> --- drivers/pci/controller/pci-hyperv.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) With the below debug code: @@ -643,6 +643,9 @@ static void hv_arch_irq_unmask(struct irq_data *data) pbus = pdev->bus; hbus = container_of(pbus->sysdata, struct hv_pcibus_device, sysdata); int_desc = data->chip_data; + if (!int_desc) + dev_warn(&hbus->hdev->device, "%s() can not unmask irq %u\n", + __func__, data->irq); spin_lock_irqsave(&hbus->retarget_msi_interrupt_lock, flags); @@ -1865,6 +1868,11 @@ static void hv_compose_msi_msg(struct irq_data *data, struct msi_msg *msg) goto free_int_desc; } + printk("%s: line %d: irq=%u\n", __func__, __LINE__, data->irq); + { + static bool delayed; //remove the device within the 10s. + if (!delayed) { delayed = true; mdelay(10000); } + } ret = vmbus_sendpacket_getid(hpdev->hbus->hdev->channel, &ctxt.int_pkts, size, (unsigned long)&ctxt.pci_pkt, &trans_id, VM_PKT_DATA_INBAND, I'm able to repro the below panic: [ 23.258674] hv_pci b92a0085-468b-407a-a88a-d33fac8edc75: PCI VMBus probing: Using version 0x10004 [ 23.271313] hv_pci b92a0085-468b-407a-a88a-d33fac8edc75: PCI host bridge to bus 468b:00 [ 23.274554] pci_bus 468b:00: root bus resource [mem 0xfe0000000-0xfe00fffff window] [ 23.277733] pci_bus 468b:00: No busn resource found for root bus, will use [bus 00-ff] [ 23.283845] pci 468b:00:02.0: [15b3:1016] type 00 class 0x020000 [ 23.289796] pci 468b:00:02.0: reg 0x10: [mem 0xfe0000000-0xfe00fffff 64bit pref] [ 23.296463] pci 468b:00:02.0: enabling Extended Tags ... [ 23.331300] pci_bus 468b:00: busn_res: [bus 00-ff] end is updated to 00 [ 23.334130] pci 468b:00:02.0: BAR 0: assigned [mem 0xfe0000000-0xfe00fffff 64bit pref] [ 23.507985] mlx5_core 468b:00:02.0: no default pinctrl state [ 23.510834] mlx5_core 468b:00:02.0: enabling device (0000 -> 0002) [ 23.516843] mlx5_core 468b:00:02.0: firmware version: 14.25.8102 [ 23.745069] hv_compose_msi_msg: line 1871: irq=24 [ 33.685554] hv_pci b92a0085-468b-407a-a88a-d33fac8edc75: the device is being ejected [ 33.690855] hv_compose_msi_msg: line 1871: irq=25 [ 33.694797] hv_compose_msi_msg: line 1871: irq=26 [ 33.698884] hv_compose_msi_msg: line 1871: irq=27 [ 33.702910] hv_compose_msi_msg: line 1871: irq=28 [ 33.705726] hv_compose_msi_msg: line 1871: irq=29 [ 33.709644] hv_compose_msi_msg: line 1871: irq=29 [ 33.712182] hv_pci b92a0085-468b-407a-a88a-d33fac8edc75: hv_arch_irq_unmask() can not unmask irq 29 [ 33.716625] BUG: kernel NULL pointer dereference, address: 0000000000000008 ... [ 33.737426] Workqueue: events work_for_cpu_fn [ 33.739562] RIP: 0010:hv_irq_unmask+0xc2/0x400 [pci_hyperv] ... [ 33.778511] Call Trace: [ 33.779533] <TASK> [ 33.780428] unmask_irq.part.0+0x23/0x40 [ 33.781994] irq_enable+0x60/0x70 [ 33.783336] __irq_startup+0x5b/0x80 [ 33.784772] irq_startup+0x75/0x140 [ 33.786175] __setup_irq+0x3ae/0x840 [ 33.787586] request_threaded_irq+0x112/0x180 [ 33.789298] mlx5_irq_alloc+0x111/0x310 [mlx5_core] [ 33.791464] irq_pool_request_vector+0x72/0x80 [mlx5_core] [ 33.794449] mlx5_ctrl_irq_request+0xc9/0x160 [mlx5_core] [ 33.797454] mlx5_eq_table_create+0x9e/0xb30 [mlx5_core] [ 33.802127] mlx5_load+0x54/0x3b0 [mlx5_core] [ 33.804157] mlx5_init_one+0x1e6/0x550 [mlx5_core] [ 33.806347] probe_one+0x2e5/0x460 [mlx5_core] [ 33.808664] local_pci_probe+0x4b/0xb0 [ 33.810377] work_for_cpu_fn+0x1a/0x30 [ 33.812275] process_one_work+0x21f/0x430 [ 33.814700] worker_thread+0x1fa/0x3c0