Thread (12 messages) flat view 12 messages, 4 authors, 2021-07-05

Re: [PATCH 3/3] vDPA/ifcvf: set_status() should get a adapter from the mgmt dev

From: Zhu, Lingshan <hidden>
Date: 2021-07-05 09:46:54
Also in: kvm


On 7/5/2021 1:09 PM, Jason Wang wrote:
在 2021/6/30 下午4:21, Zhu Lingshan 写道:
quoted
ifcvf_vdpa_set_status() should get a adapter from the
management device

Signed-off-by: Zhu Lingshan <redacted>
---
  drivers/vdpa/ifcvf/ifcvf_main.c | 4 +++-
  1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/vdpa/ifcvf/ifcvf_main.c 
b/drivers/vdpa/ifcvf/ifcvf_main.c
index 7c2f64ca2163..28c71eef1d2b 100644
--- a/drivers/vdpa/ifcvf/ifcvf_main.c
+++ b/drivers/vdpa/ifcvf/ifcvf_main.c
@@ -212,13 +212,15 @@ static u8 ifcvf_vdpa_get_status(struct 
vdpa_device *vdpa_dev)
    static void ifcvf_vdpa_set_status(struct vdpa_device *vdpa_dev, 
u8 status)
  {
+    struct ifcvf_vdpa_mgmt_dev *ifcvf_mgmt_dev;
      struct ifcvf_adapter *adapter;
      struct ifcvf_hw *vf;
      u8 status_old;
      int ret;
        vf  = vdpa_to_vf(vdpa_dev);
-    adapter = dev_get_drvdata(vdpa_dev->dev.parent);

If this is a fix for patch 2, you need to squash this into that one.
sure will squash it to patch 2
Any reason that vdpa_to_adapter() can't work?
will use it in V2.
And I see:

+struct ifcvf_vdpa_mgmt_dev {
+    struct vdpa_mgmt_dev mdev;
+    struct ifcvf_adapter *adapter;
+    struct pci_dev *pdev;
+};

What's the reason for having a adapter pointer here?
because in ifcvf_remove(), we need to get the management device from 
pdev struct, so need to set the management device pointor
to the pdev drvdata, then need this *adapter pointor to address the adapter.

Thanks

Thanks

quoted
+    ifcvf_mgmt_dev = dev_get_drvdata(vdpa_dev->dev.parent);
+    adapter = ifcvf_mgmt_dev->adapter;
      status_old = ifcvf_get_status(vf);
        if (status_old == status)
  
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help