linux-kernel - Re: [PATCH] usb: dwc3: meson-g12a: Add support for IRQ based OTG switching

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: <8950adde-0942-5f04-0ce6-922a9886c440@baylibre.com>
Date:   Wed, 12 Jun 2019 17:13:10 +0200
From:   Neil Armstrong <narmstrong@...libre.com>
To:     Martin Blumenstingl <martin.blumenstingl@...glemail.com>
Cc:     balbi@...nel.org, linux-amlogic@...ts.infradead.org,
        linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH] usb: dwc3: meson-g12a: Add support for IRQ based OTG
 switching

On 11/06/2019 20:08, Martin Blumenstingl wrote:
> Hi Neil,
> 
> On Tue, Jun 11, 2019 at 3:58 PM Neil Armstrong <narmstrong@...libre.com> wrote:
>>
>> Add support for the OTG ID change interrupt to switch between Host
>> and Device mode.
>>
>> Tested on the Hardkernel Odroid-N2 board.
>>
>> Signed-off-by: Neil Armstrong <narmstrong@...libre.com>
> with the three questions/comments below answered/addressed:
> Reviewed-by: Martin Blumenstingl <martin.blumenstingl@...glemail.com>
> 
>> ---
>>  drivers/usb/dwc3/dwc3-meson-g12a.c | 32 ++++++++++++++++++++++++++++--
>>  1 file changed, 30 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/usb/dwc3/dwc3-meson-g12a.c b/drivers/usb/dwc3/dwc3-meson-g12a.c
>> index 2aec31a2eacb..e5c5ad0d529e 100644
>> --- a/drivers/usb/dwc3/dwc3-meson-g12a.c
>> +++ b/drivers/usb/dwc3/dwc3-meson-g12a.c
> the comment block at the start of the driver file also contains a
> "TOFIX" which points to the missing IRQ handling
> can you please also drop that TOFIX comment in lines 15/16?

Indeed, thanks for pointing that !

> 
>> @@ -348,6 +348,22 @@ static enum usb_role dwc3_meson_g12a_role_get(struct device *dev)
>>                 USB_ROLE_HOST : USB_ROLE_DEVICE;
>>  }
>>
>> +static irqreturn_t dwc3_meson_g12a_irq_thread(int irq, void *data)
>> +{
>> +       struct dwc3_meson_g12a *priv = data;
>> +       enum phy_mode otg_id;
>> +
>> +       otg_id = dwc3_meson_g12a_get_id(priv);
>> +       if (otg_id != priv->otg_phy_mode) {
>> +               if (dwc3_meson_g12a_otg_mode_set(priv, otg_id))
>> +                       dev_warn(priv->dev, "Failed to switch OTG mode\n");
>> +       }
>> +
>> +       regmap_update_bits(priv->regmap, USB_R5, USB_R5_ID_DIG_IRQ, 0);
>> +
>> +       return IRQ_HANDLED;
>> +}
>> +
>>  static struct device *dwc3_meson_g12_find_child(struct device *dev,
>>                                                 const char *compatible)
>>  {
>> @@ -374,7 +390,7 @@ static int dwc3_meson_g12a_probe(struct platform_device *pdev)
>>         void __iomem *base;
>>         struct resource *res;
>>         enum phy_mode otg_id;
>> -       int ret, i;
>> +       int ret, i, irq;
>>
>>         priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
>>         if (!priv)
>> @@ -436,6 +452,19 @@ static int dwc3_meson_g12a_probe(struct platform_device *pdev)
>>         /* Get dr_mode */
>>         priv->otg_mode = usb_get_dr_mode(dev);
>>
>> +       if (priv->otg_mode == USB_DR_MODE_OTG) {
>> +               /* Ack irq before registering */
>> +               regmap_update_bits(priv->regmap, USB_R5,
>> +                                  USB_R5_ID_DIG_IRQ, 0);
> I assume that either the IRQ line is:
> - always enabled
> - enabled when (USB_R5_ID_DIG_EN_0 | USB_R5_ID_DIG_EN_1 |
> USB_R5_ID_DIG_TH_MASK) are set (which we already do in
> dwc3_meson_g12a_usb_init)

Can't say... I suspect the (USB_R5_ID_DIG_EN_0 | USB_R5_ID_DIG_EN_1 |
> USB_R5_ID_DIG_TH_MASK) enables the detection.
The regmap_update_bits(USB_R5_ID_DIG_IRQ) is only here to make sure the "current"
irq event is masked, whatever the previous init.

Or I misunderstood question ?

> 
>> +               irq = platform_get_irq(pdev, 0);
> do we need to check the IRQ before trying to request it?
> drivers/gpu/drm/meson/meson_dw_hdmi.c and drivers/usb/dwc3/host.c for
> example error out if irq number is lower than 0

No, devm_request_threaded_irq() will fail if -1 is given, I've using this scheme
for a while now !

Neil

> 
> (it's great to see that this only required a small patch to make it work :))

yeah, I was impressed when I wrote it, I expected much more work

> 
> 
> Martin


Thanks,
Neil
>