linux-kernel - RE: [PATCH] usb: cdnsp: Fix issue with Clear Feature Halt Endpoint

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]

Date:   Mon, 7 Nov 2022 06:01:10 +0000
From:   Pawel Laszczak <pawell@...ence.com>
To:     Peter Chen <peter.chen@...nel.org>
CC:     "gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>,
        "linux-usb@...r.kernel.org" <linux-usb@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "stable@...r.kernel.org" <stable@...r.kernel.org>
Subject: RE: [PATCH] usb: cdnsp: Fix issue with Clear Feature Halt Endpoint

>
>On 22-10-26 05:37:10, Pawel Laszczak wrote:
>> During handling Clear Halt Endpoint Feature request driver invokes
>> Reset Endpoint command. Because this command has some issue with
>
>What are issues? Would you please explain more?

I don't know the internal behavior of controller. I know that there is the issue with resetting endpoint
in running state, resulting in the  controller after rearming start handling incorrect TRB.

>
>> transition endpoint from Running to Idle state the driver must stop
>> the endpoint by using Stop Endpoint command.
>>
>> cc: <stable@...r.kernel.org>
>> Fixes: 3d82904559f4 ("usb: cdnsp: cdns3 Add main part of Cadence
>> USBSSP DRD Driver")
>> Signed-off-by: Pawel Laszczak <pawell@...ence.com>
>> ---
>>  drivers/usb/cdns3/cdnsp-gadget.c | 12 ++++--------
>>  drivers/usb/cdns3/cdnsp-ring.c   |  3 ++-
>>  2 files changed, 6 insertions(+), 9 deletions(-)
>>
>> diff --git a/drivers/usb/cdns3/cdnsp-gadget.c
>> b/drivers/usb/cdns3/cdnsp-gadget.c
>> index e2e7d16f43f4..0576f9b0e4aa 100644
>> --- a/drivers/usb/cdns3/cdnsp-gadget.c
>> +++ b/drivers/usb/cdns3/cdnsp-gadget.c
>> @@ -600,11 +600,11 @@ int cdnsp_halt_endpoint(struct cdnsp_device
>> *pdev,
>>
>>  	trace_cdnsp_ep_halt(value ? "Set" : "Clear");
>>
>> -	if (value) {
>> -		ret = cdnsp_cmd_stop_ep(pdev, pep);
>> -		if (ret)
>> -			return ret;
>> +	ret = cdnsp_cmd_stop_ep(pdev, pep);
>> +	if (ret)
>> +		return ret;
>>
>
>In your change ,it call cdnsp_cmd_stop_ep unconditionally, no matter set or
>clear halt? Is it your expectation? If it is, why?

No exactly unconditionally.

The below condition:
	if (ep_state == EP_STATE_STOPPED || ep_state == EP_STATE_DISABLED 
		||   ep_state == EP_STATE_HALTED) {
			goto ep_stopped;

Will decide whether command will be called or not.
It will be called only when endpoint is in RUNNING state.

Pawel

>
>> +	if (value) {
>>  		if (GET_EP_CTX_STATE(pep->out_ctx) ==
>EP_STATE_STOPPED) {
>>  			cdnsp_queue_halt_endpoint(pdev, pep->idx);
>>  			cdnsp_ring_cmd_db(pdev);
>> @@ -613,10 +613,6 @@ int cdnsp_halt_endpoint(struct cdnsp_device
>> *pdev,
>>
>>  		pep->ep_state |= EP_HALTED;
>>  	} else {
>> -		/*
>> -		 * In device mode driver can call reset endpoint command
>> -		 * from any endpoint state.
>> -		 */
>>  		cdnsp_queue_reset_ep(pdev, pep->idx);
>>  		cdnsp_ring_cmd_db(pdev);
>>  		ret = cdnsp_wait_for_cmd_compl(pdev); diff --git
>> a/drivers/usb/cdns3/cdnsp-ring.c b/drivers/usb/cdns3/cdnsp-ring.c
>> index 25e5e51cf5a2..aa79bce89d8a 100644
>> --- a/drivers/usb/cdns3/cdnsp-ring.c
>> +++ b/drivers/usb/cdns3/cdnsp-ring.c
>> @@ -2081,7 +2081,8 @@ int cdnsp_cmd_stop_ep(struct cdnsp_device
>*pdev, struct cdnsp_ep *pep)
>>  	u32 ep_state = GET_EP_CTX_STATE(pep->out_ctx);
>>  	int ret = 0;
>>
>> -	if (ep_state == EP_STATE_STOPPED || ep_state ==
>EP_STATE_DISABLED) {
>> +	if (ep_state == EP_STATE_STOPPED || ep_state ==
>EP_STATE_DISABLED ||
>> +	    ep_state == EP_STATE_HALTED) {
>>  		trace_cdnsp_ep_stopped_or_disabled(pep->out_ctx);
>>  		goto ep_stopped;
>>  	}
>> --
>> 2.25.1
>>
>
>--
>
>Thanks,
>Peter Chen

Regards,
Pawel Laszczak