Re: [RFT PATCH] xhci: Fix null pointer dereference in resume if xhci has only one roothub

[Mathias Nyman <mathias.nyman@linux.intel.com>]

*On 9.6.2022 18.41, Matthias Kaehlcke wrote:
> On Thu, Jun 09, 2022 at 03:03:36PM +0300, Mathias Nyman wrote:
>> In the re-init path xhci_resume() passes 'hcd->primary_hcd' to hci_init(),
>> however this field isn't initialized by __usb_create_hcd() for a HCD
>> without secondary controller.
>>
>> xhci_resume() is called once per xHC device, not per hcd, so the extra
>> checking for primary hcd can be removed.
>>
>> Fixes: e0fe986972f5 ("usb: host: xhci-plat: prepare operation w/o shared hcd")
>> Reported-by: Matthias Kaehlcke <mka@chromium.org>
>> Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
>> ---
>>   drivers/usb/host/xhci.c | 15 +++++----------
>>   1 file changed, 5 insertions(+), 10 deletions(-)
>>
>> diff --git a/drivers/usb/host/xhci.c b/drivers/usb/host/xhci.c
>> index f0ab63138016..9ac56e9ffc64 100644
>> --- a/drivers/usb/host/xhci.c
>> +++ b/drivers/usb/host/xhci.c
>> @@ -1107,7 +1107,6 @@ int xhci_resume(struct xhci_hcd *xhci, bool hibernated)
>>   {
>>   	u32			command, temp = 0;
>>   	struct usb_hcd		*hcd = xhci_to_hcd(xhci);
>> -	struct usb_hcd		*secondary_hcd;
>>   	int			retval = 0;
>>   	bool			comp_timer_running = false;
>>   	bool			pending_portevent = false;
>> @@ -1214,23 +1213,19 @@ int xhci_resume(struct xhci_hcd *xhci, bool hibernated)
>>   		 * first with the primary HCD, and then with the secondary HCD.
>>   		 * If we don't do the same, the host will never be started.
>>   		 */
>> -		if (!usb_hcd_is_primary_hcd(hcd))
>> -			secondary_hcd = hcd;
>> -		else
>> -			secondary_hcd = xhci->shared_hcd;
>> -
>>   		xhci_dbg(xhci, "Initialize the xhci_hcd\n");
>> -		retval = xhci_init(hcd->primary_hcd);
>> +		retval = xhci_init(hcd);
>>   		if (retval)
>>   			return retval;
>>   		comp_timer_running = true;
>>   
>>   		xhci_dbg(xhci, "Start the primary HCD\n");
> 
> Is the log still correct? IIUC this now isn't necessarily the primary HCD.

It's still correct as this is always the xhci->main_hcd, the one that is created first.
There could be a better word than "primary", but my brain is accustomed to seeing this
line while debugging.


> 
>> -		retval = xhci_run(hcd->primary_hcd);
>> -		if (!retval && secondary_hcd) {
>> +		retval = xhci_run(hcd);
>> +		if (!retval && xhci->shared_hcd) {
>>   			xhci_dbg(xhci, "Start the secondary HCD\n");
> 
> ditto

same, always xhci->shared_hcd, the one that is created second.

> 
>> -			retval = xhci_run(secondary_hcd);
>> +			retval = xhci_run(xhci->shared_hcd);
>>   		}
>> +
>>   		hcd->state = HC_STATE_SUSPENDED;
>>   		if (xhci->shared_hcd)
>>   			xhci->shared_hcd->state = HC_STATE_SUSPENDED;
> 
> Tested-by: Matthias Kaehlcke <mka@chromium.org>

Thanks for testing
-Mathias