Message ID | 20241025064331.1049538-1-chenridong@huaweicloud.com (mailing list archive) |
---|---|
State | New |
Headers | show |
Series | usb: cdns3: avoid possible null_ptr_deref in cdns3_gadget_ep_queue | expand |
On Fri, Oct 25, 2024 at 06:43:31AM +0000, Chen Ridong wrote: > From: Chen Ridong <chenridong@huawei.com> > > The cdns3_gadget_ep_alloc_request functions may return NULL > when memory is out of use. To void possible null_ptr_deref, > return ENOMEM when cdns3_gadget_ep_alloc_request returns NULL. > > Signed-off-by: Chen Ridong <chenridong@huawei.com> > --- > drivers/usb/cdns3/cdns3-gadget.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/usb/cdns3/cdns3-gadget.c b/drivers/usb/cdns3/cdns3-gadget.c > index fd1beb10bba7..982e8e0759e8 100644 > --- a/drivers/usb/cdns3/cdns3-gadget.c > +++ b/drivers/usb/cdns3/cdns3-gadget.c > @@ -2657,6 +2657,10 @@ static int cdns3_gadget_ep_queue(struct usb_ep *ep, struct usb_request *request, > struct cdns3_request *priv_req; > > zlp_request = cdns3_gadget_ep_alloc_request(ep, GFP_ATOMIC); > + if (!zlp_request) { > + spin_unlock_irqrestore(&priv_dev->lock, flags); > + return -ENOMEM; Are you sure this is ok? What about the resources that have been allocated before this return would interrupt them? How was this tested? thanks, greg k-h
diff --git a/drivers/usb/cdns3/cdns3-gadget.c b/drivers/usb/cdns3/cdns3-gadget.c index fd1beb10bba7..982e8e0759e8 100644 --- a/drivers/usb/cdns3/cdns3-gadget.c +++ b/drivers/usb/cdns3/cdns3-gadget.c @@ -2657,6 +2657,10 @@ static int cdns3_gadget_ep_queue(struct usb_ep *ep, struct usb_request *request, struct cdns3_request *priv_req; zlp_request = cdns3_gadget_ep_alloc_request(ep, GFP_ATOMIC); + if (!zlp_request) { + spin_unlock_irqrestore(&priv_dev->lock, flags); + return -ENOMEM; + } zlp_request->buf = priv_dev->zlp_buf; zlp_request->length = 0;