linux-kernel - Re: [PATCH -next] xhci: fix two places when dealing with return value of function xhci_check

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <c7f6a8bb-76b6-cd2d-7551-b599a8276f5c@kylinos.cn>
Date:   Wed, 26 Jan 2022 18:22:45 +0800
From:   谢泓宇 <xiehongyu1@...inos.cn>
To:     Greg KH <gregkh@...uxfoundation.org>,
        Hongyu Xie <xy521521@...il.com>
Cc:     mathias.nyman@...el.com, linux-kernel@...r.kernel.org,
        linux-usb@...r.kernel.org, 125707942@...com, stable@...r.kernel.org
Subject: Re: [PATCH -next] xhci: fix two places when dealing with return value
 of function xhci_check_args

1."What problem?
r8152_submit_rx needs to detach netdev if -ENODEV happened, but -ENODEV 
will never happen
because xhci_urb_enqueue only returns -EINVAL if the return value of 
xhci_check_args <= 0. So
r8152_submit_rx will will call napi_schedule to re-submit that urb, and 
this will cause infinite urb
submission.
The whole point is, if xhci_check_args returns value A, 
xhci_urb_enqueque shouldn't return any
other value, because that will change some driver's behavior(like r8152.c).

2."So if 0 is returned, you will now return that here, is that ok?
That is a change in functionality.
But this can only ever be the case for a root hub, is that ok?"

It's the same logic, but now xhci_urb_enqueue can return -ENODEV if xHC 
is halted.
If it happens on a root hub,  xhci_urb_enqueue won't be called.

3."Again, this means all is good?  Why is this being called for a root hub?"

It is the same logic with the old one, but now 
xhci_check_streams_endpoint can return -ENODEV if xHC is halted.


thanks

Hongyu Xie


On Tue, 25 Jan 2022 at 22:02, Greg Kroah-Hartman
<gregkh@...uxfoundation.org>  wrote:

> On Wed, Jan 26, 2022 at 05:41:26PM +0800, Hongyu Xie wrote:
>> From: Hongyu Xie <xiehongyu1@...inos.cn>
>>
>> xhci_check_args returns 4 types of value, -ENODEV, -EINVAL, 1 and 0.
>> xhci_urb_enqueue and xhci_check_streams_endpoint return -EINVAL if
>> the return value of xhci_check_args <= 0.
>> This will cause a problem.
> What problem?
>
>> For example, r8152_submit_rx calling usb_submit_urb in
>> drivers/net/usb/r8152.c.
>> r8152_submit_rx will never get -ENODEV after submiting an urb
>> when xHC is halted,
>> because xhci_urb_enqueue returns -EINVAL in the very beginning.
>>
>> Fixes: 203a86613fb3 ("xhci: Avoid NULL pointer deref when host dies.")
>> Cc: stable@...r.kernel.org
>> Signed-off-by: Hongyu Xie <xiehongyu1@...inos.cn>
>> ---
>>   drivers/usb/host/xhci.c | 9 ++++++---
>>   1 file changed, 6 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/usb/host/xhci.c b/drivers/usb/host/xhci.c
>> index dc357cabb265..a7a55dd206fe 100644
>> --- a/drivers/usb/host/xhci.c
>> +++ b/drivers/usb/host/xhci.c
>> @@ -1604,9 +1604,12 @@ static int xhci_urb_enqueue(struct usb_hcd *hcd, struct urb *urb, gfp_t mem_flag
>>   	struct urb_priv	*urb_priv;
>>   	int num_tds;
>>   
>> -	if (!urb || xhci_check_args(hcd, urb->dev, urb->ep,
>> -					true, true, __func__) <= 0)
>> +	if (!urb)
>>   		return -EINVAL;
>> +	ret = xhci_check_args(hcd, urb->dev, urb->ep,
>> +					true, true, __func__);
>> +	if (ret <= 0)
>> +		return ret;
> So if 0 is returned, you will now return that here, is that ok?
> That is a change in functionality.
>
> But this can only ever be the case for a root hub, is that ok?
>
>>   
>>   	slot_id = urb->dev->slot_id;
>>   	ep_index = xhci_get_endpoint_index(&urb->ep->desc);
>> @@ -3323,7 +3326,7 @@ static int xhci_check_streams_endpoint(struct xhci_hcd *xhci,
>>   		return -EINVAL;
>>   	ret = xhci_check_args(xhci_to_hcd(xhci), udev, ep, 1, true, __func__);
>>   	if (ret <= 0)
>> -		return -EINVAL;
>> +		return ret;
> Again, this means all is good?  Why is this being called for a root hub?
>
> thanks,
>
> greg k-h