[PATCH v2 0/5] Fix probe failed when modprobe modules

[PATCH v2 0/5] Fix probe failed when modprobe modules

[Li Zetao]

This patchset fixes similar issue, the root cause of the
problem is that the virtqueues are not stopped on error
handling path.

Changes since v1:
- Modify the description error of the test case and fixes tag
  information.
- Add patch to fix virtio_gpu module.

v1 at:
https://lore.kernel.org/all/20221128021005.232105-1-lizetao1@huawei.com/

Li Zetao (5):
  9p: Fix probe failed when modprobe 9pnet_virtio
  virtio-mem: Fix probe failed when modprobe virtio_mem
  virtio-input: Fix probe failed when modprobe virtio_input
  virtio-blk: Fix probe failed when modprobe virtio_blk
  drm/virtio: Fix probe failed when modprobe virtio_gpu

 drivers/block/virtio_blk.c           | 1 +
 drivers/gpu/drm/virtio/virtgpu_kms.c | 1 +
 drivers/virtio/virtio_input.c        | 1 +
 drivers/virtio/virtio_mem.c          | 1 +
 net/9p/trans_virtio.c                | 1 +
 5 files changed, 5 insertions(+)

-- 
2.25.1

[PATCH v2 1/5] 9p: Fix probe failed when modprobe 9pnet_virtio

[Li Zetao]

When doing the following test steps, an error was found:
  step 1: modprobe 9pnet_virtio succeeded
    # modprobe 9pnet_virtio      <-- OK

  step 2: fault injection in sysfs_create_file()
    # modprobe -r 9pnet_virtio   <-- OK
    # ...
      FAULT_INJECTION: forcing a failure.
      name failslab, interval 1, probability 0, space 0, times 0
      CPU: 0 PID: 3790 Comm: modprobe Tainted: G        W
      6.1.0-rc6-00285-g6a1e40c4b995-dirty #108
      Hardware name: QEMU Standard PC (i440FX + PIIX, 1996)
      Call Trace:
       <TASK>
       ...
       should_failslab+0xa/0x20
       ...
       sysfs_create_file_ns+0x130/0x1d0
       p9_virtio_probe+0x662/0xb30 [9pnet_virtio]
       virtio_dev_probe+0x608/0xae0
       ...
       </TASK>
      9pnet_virtio: probe of virtio3 failed with error -12

  step 3: modprobe 9pnet_virtio failed
    # modprobe 9pnet_virtio       <-- failed
      9pnet_virtio: probe of virtio3 failed with error -2

The root cause of the problem is that the virtqueues are not
stopped on the error handling path when sysfs_create_file()
fails in p9_virtio_probe(), resulting in an error "-ENOENT"
returned in the next modprobe call in setup_vq().

virtio_pci_modern_device uses virtqueues to send or
receive message, and "queue_enable" records whether the
queues are available. In vp_modern_find_vqs(), all queues
will be selected and activated, but once queues are enabled
there is no way to go back except reset.

Fix it by reset virtio device on error handling path. After
virtio_find_single_vq() succeeded, all virtqueues should be
stopped on error handling path.

Fixes: ea52bf8eda98 ("9p/trans_virtio: reset virtio device on remove")
Signed-off-by: Li Zetao <lizetao1@huawei.com>
Reviewed-by: Christian Schoenebeck <linux_oss@crudebyte.com>
---
v1 -> v2: modify the description error of the test case in step 3 and
modify the fixes tag information.

 net/9p/trans_virtio.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/net/9p/trans_virtio.c b/net/9p/trans_virtio.c
index e757f0601304..39933187284b 100644
--- a/net/9p/trans_virtio.c
+++ b/net/9p/trans_virtio.c
@@ -668,6 +668,7 @@ static int p9_virtio_probe(struct virtio_device *vdev)
 out_free_tag:
 	kfree(tag);
 out_free_vq:
+	virtio_reset_device(vdev);
 	vdev->config->del_vqs(vdev);
 out_free_chan:
 	kfree(chan);
-- 
2.25.1

[PATCH v2 2/5] virtio-mem: Fix probe failed when modprobe virtio_mem

[Li Zetao]

When doing the following test steps, an error was found:
  step 1: modprobe virtio_mem succeeded
    # modprobe virtio_mem      <-- OK

  step 2: fault injection in virtio_mem_init()
    # modprobe -r virtio_mem   <-- OK
    # ...
      CPU: 0 PID: 1837 Comm: modprobe Not tainted
      6.1.0-rc6-00285-g6a1e40c4b995-dirty
      Hardware name: QEMU Standard PC (i440FX + PIIX, 1996)
      Call Trace:
       <TASK>
       should_fail.cold+0x5/0x1f
       ...
       virtio_mem_init_hotplug+0x9ae/0xe57 [virtio_mem]
       virtio_mem_init.cold+0x327/0x339 [virtio_mem]
       virtio_mem_probe+0x48e/0x910 [virtio_mem]
       virtio_dev_probe+0x608/0xae0
       ...
       </TASK>
      virtio_mem virtio4: could not reserve device region
      virtio_mem: probe of virtio4 failed with error -16

  step 3: modprobe virtio_mem failed
    # modprobe virtio_mem       <-- failed
      virtio_mem: probe of virtio4 failed with error -16

The root cause of the problem is that the virtqueues are not
stopped on the error handling path when virtio_mem_init()
fails in virtio_mem_probe(), resulting in an error "-ENOENT"
returned in the next modprobe call in setup_vq().

virtio_pci_modern_device uses virtqueues to send or
receive message, and "queue_enable" records whether the
queues are available. In vp_modern_find_vqs(), all queues
will be selected and activated, but once queues are enabled
there is no way to go back except reset.

Fix it by reset virtio device on error handling path. After
virtio_mem_init_vq() succeeded, all virtqueues should be
stopped on error handling path.

Fixes: 5f1f79bbc9e2 ("virtio-mem: Paravirtualized memory hotplug")
Signed-off-by: Li Zetao <lizetao1@huawei.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
---
v1 -> v2: modify the description error of the test case in step 3 and
modify the fixes tag information.

 drivers/virtio/virtio_mem.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index 0c2892ec6817..c7f09c2ce982 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -2793,6 +2793,7 @@ static int virtio_mem_probe(struct virtio_device *vdev)
 
 	return 0;
 out_del_vq:
+	virtio_reset_device(vdev);
 	vdev->config->del_vqs(vdev);
 out_free_vm:
 	kfree(vm);
-- 
2.25.1

[PATCH v2 3/5] virtio-input: Fix probe failed when modprobe virtio_input

[Li Zetao]

When doing the following test steps, an error was found:
  step 1: modprobe virtio_input succeeded
    # modprobe virtio_input      <-- OK

  step 2: fault injection in input_allocate_device()
    # modprobe -r virtio_input   <-- OK
    # ...
      CPU: 0 PID: 4260 Comm: modprobe Tainted: G        W
      6.1.0-rc6-00285-g6a1e40c4b995-dirty #109
      Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
      Call Trace:
       <TASK>
       should_fail.cold+0x5/0x1f
       ...
       kmalloc_trace+0x27/0xa0
       input_allocate_device+0x43/0x280
       virtinput_probe+0x23b/0x1648 [virtio_input]
       ...
       </TASK>
      virtio_input: probe of virtio5 failed with error -12

  step 3: modprobe virtio_input failed
    # modprobe virtio_input       <-- failed
      virtio_input: probe of virtio1 failed with error -2

The root cause of the problem is that the virtqueues are not
stopped on the error handling path when input_allocate_device()
fails in virtinput_probe(), resulting in an error "-ENOENT"
returned in the next modprobe call in setup_vq().

virtio_pci_modern_device uses virtqueues to send or
receive message, and "queue_enable" records whether the
queues are available. In vp_modern_find_vqs(), all queues
will be selected and activated, but once queues are enabled
there is no way to go back except reset.

Fix it by reset virtio device on error handling path. After
virtinput_init_vqs() succeeded, all virtqueues should be
stopped on error handling path.

Fixes: 271c865161c5 ("Add virtio-input driver.")
Signed-off-by: Li Zetao <lizetao1@huawei.com>
---
v1 -> v2: modify the description error of the test case in step 3 and
modify the fixes tag information.

 drivers/virtio/virtio_input.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/virtio/virtio_input.c b/drivers/virtio/virtio_input.c
index 3aa46703872d..f638f1cd3531 100644
--- a/drivers/virtio/virtio_input.c
+++ b/drivers/virtio/virtio_input.c
@@ -330,6 +330,7 @@ static int virtinput_probe(struct virtio_device *vdev)
 err_mt_init_slots:
 	input_free_device(vi->idev);
 err_input_alloc:
+	virtio_reset_device(vdev);
 	vdev->config->del_vqs(vdev);
 err_init_vq:
 	kfree(vi);
-- 
2.25.1

[PATCH v2 4/5] virtio-blk: Fix probe failed when modprobe virtio_blk

[Li Zetao]

When doing the following test steps, an error was found:
  step 1: modprobe virtio_blk succeeded
    # modprobe virtio_blk      <-- OK

  step 2: fault injection in __blk_mq_alloc_disk()
    # modprobe -r virtio_blk   <-- OK
    # ...
      CPU: 0 PID: 4578 Comm: modprobe Tainted: G        W
      6.1.0-rc6-00308-g644e9524388a-dirty
      Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
      Call Trace:
       <TASK>
       should_failslab+0xa/0x20
       ...
       blk_alloc_queue+0x3a4/0x780
       __blk_mq_alloc_disk+0x91/0x1f0
       virtblk_probe+0x6ff/0x1f20 [virtio_blk]
       ...
       </TASK>
      virtio_blk: probe of virtio1 failed with error -12

  step 3: modprobe virtio_blk failed
    # modprobe virtio_blk       <-- failed
      virtio_blk: probe of virtio1 failed with error -2

The root cause of the problem is that the virtqueues are not
stopped on the error handling path when __blk_mq_alloc_disk()
fails in virtblk_probe(), resulting in an error "-ENOENT"
returned in the next modprobe call in setup_vq().

virtio_pci_modern_device uses virtqueues to send or
receive message, and "queue_enable" records whether the
queues are available. In vp_modern_find_vqs(), all queues
will be selected and activated, but once queues are enabled
there is no way to go back except reset.

Fix it by reset virtio device on error handling path. After
init_vq() succeeded, all virtqueues should be stopped on error
handling path.

Signed-off-by: Li Zetao <lizetao1@huawei.com>
---
v1 -> v2: modify the description error.

 drivers/block/virtio_blk.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
index 19da5defd734..f401546d4e85 100644
--- a/drivers/block/virtio_blk.c
+++ b/drivers/block/virtio_blk.c
@@ -1157,6 +1157,7 @@ static int virtblk_probe(struct virtio_device *vdev)
 	put_disk(vblk->disk);
 out_free_tags:
 	blk_mq_free_tag_set(&vblk->tag_set);
+	virtio_reset_device(vdev);
 out_free_vq:
 	vdev->config->del_vqs(vdev);
 	kfree(vblk->vqs);
-- 
2.25.1

[PATCH v2 5/5] drm/virtio: Fix probe failed when modprobe virtio_gpu

[Li Zetao]

When doing the following test steps, an error was found:
  step 1: modprobe virtio_gpu succeeded
    # modprobe virtio_gpu      <-- OK

  step 2: fault injection in virtio_gpu_alloc_vbufs()
    # modprobe -r virtio_gpu   <-- OK
    # ...
      CPU: 0 PID: 1714 Comm: modprobe Not tainted 6.1.0-rc7-dirty
      Hardware name: QEMU Standard PC (i440FX + PIIX, 1996)
      Call Trace:
       <TASK>
       should_fail_ex.cold+0x1a/0x1f
       ...
       kmem_cache_create+0x12/0x20
       virtio_gpu_alloc_vbufs+0x2f/0x90 [virtio_gpu]
       virtio_gpu_init.cold+0x659/0xcad [virtio_gpu]
       virtio_gpu_probe+0x14f/0x260 [virtio_gpu]
       virtio_dev_probe+0x608/0xae0
       ?...
       </TASK>
      kmem_cache_create_usercopy(virtio-gpu-vbufs) failed with error -12

  step 3: modprobe virtio_gpu failed
    # modprobe virtio_gpu       <-- failed
      failed to find virt queues
      virtio_gpu: probe of virtio6 failed with error -2

The root cause of the problem is that the virtqueues are not
stopped on the error handling path when virtio_gpu_alloc_vbufs()
fails in virtio_gpu_init(), resulting in an error "-ENOENT"
returned in the next modprobe call in setup_vq().

virtio_pci_modern_device uses virtqueues to send or
receive message, and "queue_enable" records whether the
queues are available. In vp_modern_find_vqs(), all queues
will be selected and activated, but once queues are enabled
there is no way to go back except reset.

Fix it by reset virtio device on error handling path. After
virtio_find_vqs() succeeded, all virtqueues should be stopped
on error handling path.

Fixes: dc5698e80cf7 ("Add virtio gpu driver.")
Signed-off-by: Li Zetao <lizetao1@huawei.com>
---
v1 -> v2: patch is new.

 drivers/gpu/drm/virtio/virtgpu_kms.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/virtio/virtgpu_kms.c b/drivers/gpu/drm/virtio/virtgpu_kms.c
index 27b7f14dae89..1a03e8e13b5b 100644
--- a/drivers/gpu/drm/virtio/virtgpu_kms.c
+++ b/drivers/gpu/drm/virtio/virtgpu_kms.c
@@ -255,6 +255,7 @@ int virtio_gpu_init(struct virtio_device *vdev, struct drm_device *dev)
 err_scanouts:
 	virtio_gpu_free_vbufs(vgdev);
 err_vbufs:
+	virtio_reset_device(vgdev->vdev);
 	vgdev->vdev->config->del_vqs(vgdev->vdev);
 err_vqs:
 	dev->dev_private = NULL;
-- 
2.25.1

Re: [PATCH v2 0/5] Fix probe failed when modprobe modules

[Jens Axboe]

On 11/29/22 9:06 AM, Li Zetao wrote:
> This patchset fixes similar issue, the root cause of the
> problem is that the virtqueues are not stopped on error
> handling path.

Not related to just this patchset, but guys, Huawei really *REALLY* need
to get the email situation sorted. I'm digging whole/half patchsets out
of spam every morning.

This has been brought up in the past. And no, the cloud variant of
the email also doesn't work properly.

Talk to your IT department, get this sorted once and for all. You risk
your patches being dumped on the floor because people don't see them,
or only see small parts of a patchset. And it's really annoying to have
to deal with as a recipient.

-- 
Jens Axboe