On 20-09-23 12:22:37, Oliver Neukum wrote:
> Am Mittwoch, den 23.09.2020, 14:35 +0530 schrieb Himadri Pandya:
>
> Hi,
>
> > Many usage of usb_control_msg() do not have proper error check on return
> > value leaving scope for bugs on short reads. New usb_control_msg_recv()
> > and usb_control_msg_send() nicely wraps usb_control_msg() with proper
> > error check. Hence use the wrappers instead of calling usb_control_msg()
> > directly.
> >
> > Signed-off-by: Himadri Pandya <himadrispandya@xxxxxxxxx>
> Nacked-by: Oliver Neukum <oneukum@xxxxxxxx>
>
> > ---
> > drivers/net/usb/rtl8150.c | 32 ++++++--------------------------
> > 1 file changed, 6 insertions(+), 26 deletions(-)
> >
> > diff --git a/drivers/net/usb/rtl8150.c b/drivers/net/usb/rtl8150.c
> > index 733f120c852b..e3002b675921 100644
> > --- a/drivers/net/usb/rtl8150.c
> > +++ b/drivers/net/usb/rtl8150.c
> > @@ -152,36 +152,16 @@ static const char driver_name [] = "rtl8150";
> > */
> > static int get_registers(rtl8150_t * dev, u16 indx, u16 size, void *data)
> > {
> > - void *buf;
> > - int ret;
> > -
> > - buf = kmalloc(size, GFP_NOIO);
>
> GFP_NOIO is used here for a reason. You need to use this helper
> while in contexts of error recovery and runtime PM.
>
> > - if (!buf)
> > - return -ENOMEM;
> > -
> > - ret = usb_control_msg(dev->udev, usb_rcvctrlpipe(dev->udev, 0),
> > - RTL8150_REQ_GET_REGS, RTL8150_REQT_READ,
> > - indx, 0, buf, size, 500);
> > - if (ret > 0 && ret <= size)
> > - memcpy(data, buf, ret);
> > - kfree(buf);
> > - return ret;
> > + return usb_control_msg_recv(dev->udev, 0, RTL8150_REQ_GET_REGS,
> > + RTL8150_REQT_READ, indx, 0, data,
> > + size, 500);
>
> This internally uses kmemdup() with GFP_KERNEL.
> You cannot make this change. The API does not support it.
> I am afraid we will have to change the API first, before more
> such changes are done.
One possible fix is to add yet another argument to usb_control_msg_recv(), which
would be the GFP_XYZ flag to pass on to kmemdup(). Up to Greg, of course.
cheers,
Petko
