linux-kernel - Re: [PATCH] extcon-intel-cht-wc: Don't reset USB data connection at probe

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: <492098ce-ec4c-a860-4625-27c410d7deb3@redhat.com>
Date:   Mon, 12 Aug 2019 19:18:45 +0200
From:   Hans de Goede <hdegoede@...hat.com>
To:     Yauhen Kharuzhy <jekhor@...il.com>
Cc:     linux-kernel@...r.kernel.org, Chanwoo Choi <cw00.choi@...sung.com>,
        MyungJoo Ham <myungjoo.ham@...sung.com>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Subject: Re: [PATCH] extcon-intel-cht-wc: Don't reset USB data connection at
 probe

Hi,

On 09-08-19 17:11, Yauhen Kharuzhy wrote:
> On Fri, Aug 09, 2019 at 01:06:01PM +0200, Hans de Goede wrote:
>> Hi,
>>
>> On 8/9/19 12:01 AM, Yauhen Kharuzhy wrote:
>>> Intel Cherry Trail Whiskey Cove extcon driver connect USB data lines to
>>> PMIC at driver probing for further charger detection. This causes reset of
>>> USB data sessions and removing all devices from bus. If system was
>>> booted from Live CD or USB dongle, this makes system unusable.
>>>
>>> Check if USB ID pin is floating and re-route data lines in this case
>>> only, don't touch otherwise.
>>>
>>> Signed-off-by: Yauhen Kharuzhy <jekhor@...il.com>
>>> ---
>>>    drivers/extcon/extcon-intel-cht-wc.c | 16 ++++++++++++++--
>>>    1 file changed, 14 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/extcon/extcon-intel-cht-wc.c b/drivers/extcon/extcon-intel-cht-wc.c
>>> index 9d32150e68db..3ae573e93e6e 100644
>>> --- a/drivers/extcon/extcon-intel-cht-wc.c
>>> +++ b/drivers/extcon/extcon-intel-cht-wc.c
>>> @@ -338,6 +338,7 @@ static int cht_wc_extcon_probe(struct platform_device *pdev)
>>>    	struct intel_soc_pmic *pmic = dev_get_drvdata(pdev->dev.parent);
>>>    	struct cht_wc_extcon_data *ext;
>>>    	unsigned long mask = ~(CHT_WC_PWRSRC_VBUS | CHT_WC_PWRSRC_USBID_MASK);
>>> +	int pwrsrc_sts, id;
>>>    	int irq, ret;
>>>    	irq = platform_get_irq(pdev, 0);
>>> @@ -387,8 +388,19 @@ static int cht_wc_extcon_probe(struct platform_device *pdev)
>>>    		goto disable_sw_control;
>>>    	}
>>> -	/* Route D+ and D- to PMIC for initial charger detection */
>>> -	cht_wc_extcon_set_phymux(ext, MUX_SEL_PMIC);
>>> +	ret = regmap_read(ext->regmap, CHT_WC_PWRSRC_STS, &pwrsrc_sts);
>>> +	if (ret) {
>>> +		dev_err(ext->dev, "Error reading pwrsrc status: %d\n", ret);
>>> +		goto disable_sw_control;
>>> +	}
>>> +
>>> +	id = cht_wc_extcon_get_id(ext, pwrsrc_sts);
>>> +
>>> +	/* If no USB host or device connected, route D+ and D- to PMIC for
>>> +	 * initial charger detection
>>> +	 */
>>> +	if (id == INTEL_USB_ID_FLOAT)
>>> +		cht_wc_extcon_set_phymux(ext, MUX_SEL_PMIC);
>>
>> The check here should be != INTEL_USB_ID_GND, when we are connected as
>> device we are charging from the host we are connected to and the port
>> we are connected to may be a CDP (charging downstream port) instead of
>> a SDP (standard downstream port) allowing us to charge at 1.5A instead
>> of 0.5A, also != INTEL_USB_ID_GND matches the condition used in
>> cht_wc_extcon_pwrsrc_event to determine if we should continue with
>> charger detection there.
> 
> Actually it should be checked as ((id != INTEL_USB_ID_GND) && (id !=
> INTEL_USB_RID_A)), to not interrupt connection if ACA used and our host
> is OTG A device (for example, if LiveCD boots from USB flash inserted
> into an ACA USB hub).

Right, except that cht_wc_extcon_get_id never returns INTEL_USB_RID_A,
so checking for that would be superfluous and also would be inconsistent
with what cht_wc_extcon_pwrsrc_event is doing, so for now please just test
for != INTEL_USB_ID_GND, if cht_wc_extcon_get_id ever starts supporting
RID_A then we should fix both cht_wc_extcon_pwrsrc_event and you new check
at that time.

Hmm, I now see that cht_wc_extcon_get_id currently only reports one of:
INTEL_USB_ID_GND or INTEL_USB_ID_FLOAT so == INTEL_USB_ID_FLOAT is
equivalent to != INTEL_USB_ID_GND still please check for 1= INTEL_USB_ID_GND
as that is what cht_wc_extcon_pwrsrc_event is doing.

>> Like your other patch I will try to give this one a  test-run tomorrow.

I've given this patch a test-run today and I can confirm that BC1.2 charger
detection still works fine on my device using this driver.

Regards,

Hans