Re: [PATCH 1/2] USB: array-index-out-of-bounds in ehci_brcm_hub_control

From: Greg KH
Date: Fri Nov 05 2021 - 04:49:59 EST


On Fri, Nov 05, 2021 at 04:32:50PM +0800, Haimin Zhang wrote:
> There isn't enough check parameter `wIndex` in the function
> `ehci_brcm_hub_control`;due to the size of array `port_status`
> is 15, so it may lead to out of bounds.

Odd use of ';'

And have you seen this get out of bounds? If so, how?


>
> Signed-off-by: Haimin Zhang <tcs_kernel@xxxxxxxxxxx>
> Reported-by: TCS Robot <tcs_robot@xxxxxxxxxxx>

Signed-off-by goes below the reported-by lines.

And why are these 2 patches not "threaded" properly in email. How did
you send them?

Also, your "From" line does not match your signed-off-by line, so I
could not take these at all even if this was ok.

> ---
> drivers/usb/host/ehci-brcm.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/usb/host/ehci-brcm.c b/drivers/usb/host/ehci-brcm.c
> index d3626bfa966b..a1e3290e5459 100644
> --- a/drivers/usb/host/ehci-brcm.c
> +++ b/drivers/usb/host/ehci-brcm.c
> @@ -62,8 +62,11 @@ static int ehci_brcm_hub_control(
> u32 __iomem *status_reg;
> unsigned long flags;
> int retval, irq_disabled = 0;
> + u32 temp;
>
> - status_reg = &ehci->regs->port_status[(wIndex & 0xff) - 1];
> + temp = (wIndex & 0xff) - 1;
> + if (temp < ports)
> + status_reg = &ehci->regs->port_status[temp];

What if the test fails? Should you do something about that? status_reg
is now uninitialized, doesn't the code fail now?

How did you test this?

thanks,

greg k-h