[PATCH 1/5] Bluetooth: hci_bcm: Add irq_polarity module option

[PATCH 1/5] Bluetooth: hci_bcm: Add irq_polarity module option

[Hans de Goede]

Add irq_polarity module option for easier troubleshooting of irq-polarity
issues.

Cc: stable@vger.kernel.org
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 drivers/bluetooth/hci_bcm.c | 20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)

diff --git a/drivers/bluetooth/hci_bcm.c b/drivers/bluetooth/hci_bcm.c
index 5e3aba3c3185..bb6cd1623132 100644
--- a/drivers/bluetooth/hci_bcm.c
+++ b/drivers/bluetooth/hci_bcm.c
@@ -126,6 +126,10 @@ struct bcm_data {
 static DEFINE_MUTEX(bcm_device_lock);
 static LIST_HEAD(bcm_device_list);
 
+static int irq_polarity = -1;
+module_param(irq_polarity, int, 0444);
+MODULE_PARM_DESC(irq_polarity, "IRQ polarity 0: active-high 1: active-low");
+
 static inline void host_set_baudrate(struct hci_uart *hu, unsigned int speed)
 {
 	if (hu->serdev)
@@ -988,11 +992,17 @@ static int bcm_acpi_probe(struct bcm_device *dev)
 	}
 	acpi_dev_free_resource_list(&resources);
 
-	dmi_id = dmi_first_match(bcm_active_low_irq_dmi_table);
-	if (dmi_id) {
-		dev_warn(dev->dev, "%s: Overwriting IRQ polarity to active low",
-			    dmi_id->ident);
-		dev->irq_active_low = true;
+	if (irq_polarity != -1) {
+		dev->irq_active_low = irq_polarity;
+		dev_warn(dev->dev, "Overwriting IRQ polarity to active %s by module-param\n",
+			 dev->irq_active_low ? "low" : "high");
+	} else {
+		dmi_id = dmi_first_match(bcm_active_low_irq_dmi_table);
+		if (dmi_id) {
+			dev_warn(dev->dev, "%s: Overwriting IRQ polarity to active low",
+				 dmi_id->ident);
+			dev->irq_active_low = true;
+		}
 	}
 
 	return 0;
-- 
2.14.3

[PATCH 2/5] Bluetooth: hci_bcm: Treat Interrupt ACPI resources as always being active-low

[Hans de Goede]

Older devices with a serdev attached bcm bt hci, use an Interrupt ACPI
resource to describe the IRQ (rather then a GpioInt resource).

These device seem to all claim the IRQ is active-high and seem to all need
a DMI quirk to treat it as active-low. Instead simply always assume that
Interrupt resource specified IRQs are always active-low.

This fixes the bt device not being able to wake the host from runtime-
suspend on the: Asus T100TAM, Asus T200TA, Lenovo Yoga2 and the Toshiba
Encore, without the need to add 4 new DMI quirks for these models.

This also allows us to remove 2 DMI quirks for the Asus T100TA and Asus
T100CHI series. Likely the 2 remaining quirks can also be removed but I
could not find a DSDT of these devices to verify this.

Cc: stable@vger.kernel.org
Buglink: https://bugzilla.kernel.org/show_bug.cgi?id=198953
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1554835
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 drivers/bluetooth/hci_bcm.c | 20 +++-----------------
 1 file changed, 3 insertions(+), 17 deletions(-)

diff --git a/drivers/bluetooth/hci_bcm.c b/drivers/bluetooth/hci_bcm.c
index bb6cd1623132..7acd77a53b08 100644
--- a/drivers/bluetooth/hci_bcm.c
+++ b/drivers/bluetooth/hci_bcm.c
@@ -799,22 +799,6 @@ static const struct acpi_gpio_mapping acpi_bcm_int_first_gpios[] = {
 #ifdef CONFIG_ACPI
 /* IRQ polarity of some chipsets are not defined correctly in ACPI table. */
 static const struct dmi_system_id bcm_active_low_irq_dmi_table[] = {
-	{
-		.ident = "Asus T100TA",
-		.matches = {
-			DMI_EXACT_MATCH(DMI_SYS_VENDOR,
-					"ASUSTeK COMPUTER INC."),
-			DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "T100TA"),
-		},
-	},
-	{
-		.ident = "Asus T100CHI",
-		.matches = {
-			DMI_EXACT_MATCH(DMI_SYS_VENDOR,
-					"ASUSTeK COMPUTER INC."),
-			DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "T100CHI"),
-		},
-	},
 	{	/* Handle ThinkPad 8 tablets with BCM2E55 chipset ACPI ID */
 		.ident = "Lenovo ThinkPad 8",
 		.matches = {
@@ -842,7 +826,9 @@ static int bcm_resource(struct acpi_resource *ares, void *data)
 	switch (ares->type) {
 	case ACPI_RESOURCE_TYPE_EXTENDED_IRQ:
 		irq = &ares->data.extended_irq;
-		dev->irq_active_low = irq->polarity == ACPI_ACTIVE_LOW;
+		if (irq->polarity != ACPI_ACTIVE_LOW)
+			dev_info(dev->dev, "ACPI Interrupt resource is active-high, this is usually wrong, treating the IRQ as active-low\n");
+		dev->irq_active_low = true;
 		break;
 
 	case ACPI_RESOURCE_TYPE_GPIO:
-- 
2.14.3

[PATCH 3/5] Bluetooth: hci_bcm: Add 6 new ACPI HIDs

[Hans de Goede]

Add 6 new ACPI HIDs to enable bluetooth on devices using these HIDs,
I've tested the following HIDs / devices:

BCM2E74: Jumper ezPad mini 3
BCM2E83: Acer Iconia Tab8 w1-810
BCM2E90: Meegopad T08
BCM2EAA: Chuwi Vi8 plus (CWI519)

The reporter of Red Hat bugzilla 1554835 has tested:
BCM2E84: Lenovo Yoga2

The reporter of kernel bugzilla 274481 has tested:
BCM2E38: Toshiba Encore

Note the Lenovo Yoga2 and Toshiba Encore also needs the earlier patch to
treat all Interrupt ACPI resources as active low.

Cc: stable@vger.kernel.org
Buglink: https://bugzilla.kernel.org/attachment.cgi?id=274481
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1554835
Reported-and-tested-by: Robert R. Howell <rhowell@uwyo.edu>
Reported-and-tested-by: Christian Herzog <daduke@daduke.org>
Tested-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 drivers/bluetooth/hci_bcm.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/bluetooth/hci_bcm.c b/drivers/bluetooth/hci_bcm.c
index 7acd77a53b08..18416c435b20 100644
--- a/drivers/bluetooth/hci_bcm.c
+++ b/drivers/bluetooth/hci_bcm.c
@@ -1075,6 +1075,7 @@ static const struct hci_uart_proto bcm_proto = {
 #ifdef CONFIG_ACPI
 static const struct acpi_device_id bcm_acpi_match[] = {
 	{ "BCM2E1A", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
+	{ "BCM2E38", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
 	{ "BCM2E39", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
 	{ "BCM2E3A", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
 	{ "BCM2E3D", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
@@ -1087,12 +1088,17 @@ static const struct acpi_device_id bcm_acpi_match[] = {
 	{ "BCM2E67", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
 	{ "BCM2E71", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
 	{ "BCM2E72", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
+	{ "BCM2E74", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
 	{ "BCM2E7B", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
 	{ "BCM2E7C", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
 	{ "BCM2E7E", (kernel_ulong_t)&acpi_bcm_int_first_gpios },
+	{ "BCM2E83", (kernel_ulong_t)&acpi_bcm_int_first_gpios },
+	{ "BCM2E84", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
+	{ "BCM2E90", (kernel_ulong_t)&acpi_bcm_int_last_gpios },
 	{ "BCM2E95", (kernel_ulong_t)&acpi_bcm_int_first_gpios },
 	{ "BCM2E96", (kernel_ulong_t)&acpi_bcm_int_first_gpios },
 	{ "BCM2EA4", (kernel_ulong_t)&acpi_bcm_int_first_gpios },
+	{ "BCM2EAA", (kernel_ulong_t)&acpi_bcm_int_first_gpios },
 	{ },
 };
 MODULE_DEVICE_TABLE(acpi, bcm_acpi_match);
-- 
2.14.3

Re: [PATCH 1/5] Bluetooth: hci_bcm: Add irq_polarity module option

[Hans de Goede]

Hi Marcel, et al.,

On 16-03-18 21:28, Hans de Goede wrote:
> Add irq_polarity module option for easier troubleshooting of irq-polarity
> issues.
> 
> Cc: stable@vger.kernel.org
> Signed-off-by: Hans de Goede <hdegoede@redhat.com>

Marcel, is there any reason why this (bug-fix) series is not merged
yet? Any review remarks I missed?

Regards,

Hans



> ---
>   drivers/bluetooth/hci_bcm.c | 20 +++++++++++++++-----
>   1 file changed, 15 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/bluetooth/hci_bcm.c b/drivers/bluetooth/hci_bcm.c
> index 5e3aba3c3185..bb6cd1623132 100644
> --- a/drivers/bluetooth/hci_bcm.c
> +++ b/drivers/bluetooth/hci_bcm.c
> @@ -126,6 +126,10 @@ struct bcm_data {
>   static DEFINE_MUTEX(bcm_device_lock);
>   static LIST_HEAD(bcm_device_list);
>   
> +static int irq_polarity = -1;
> +module_param(irq_polarity, int, 0444);
> +MODULE_PARM_DESC(irq_polarity, "IRQ polarity 0: active-high 1: active-low");
> +
>   static inline void host_set_baudrate(struct hci_uart *hu, unsigned int speed)
>   {
>   	if (hu->serdev)
> @@ -988,11 +992,17 @@ static int bcm_acpi_probe(struct bcm_device *dev)
>   	}
>   	acpi_dev_free_resource_list(&resources);
>   
> -	dmi_id = dmi_first_match(bcm_active_low_irq_dmi_table);
> -	if (dmi_id) {
> -		dev_warn(dev->dev, "%s: Overwriting IRQ polarity to active low",
> -			    dmi_id->ident);
> -		dev->irq_active_low = true;
> +	if (irq_polarity != -1) {
> +		dev->irq_active_low = irq_polarity;
> +		dev_warn(dev->dev, "Overwriting IRQ polarity to active %s by module-param\n",
> +			 dev->irq_active_low ? "low" : "high");
> +	} else {
> +		dmi_id = dmi_first_match(bcm_active_low_irq_dmi_table);
> +		if (dmi_id) {
> +			dev_warn(dev->dev, "%s: Overwriting IRQ polarity to active low",
> +				 dmi_id->ident);
> +			dev->irq_active_low = true;
> +		}
>   	}
>   
>   	return 0;
> 

Re: [PATCH 1/5] Bluetooth: hci_bcm: Add irq_polarity module option

[Marcel Holtmann]

Hi Hans,

>> Add irq_polarity module option for easier troubleshooting of irq-polarity
>> issues.
>> Cc: stable@vger.kernel.org
>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> 
> Marcel, is there any reason why this (bug-fix) series is not merged
> yet? Any review remarks I missed?

I was waiting for net to be merged back into net-next so that I don’t have to deal with conflicts.

Regards

Marcel

Re: [PATCH 1/5] Bluetooth: hci_bcm: Add irq_polarity module option

[Hans de Goede]

Hi,

On 23-03-18 12:16, Marcel Holtmann wrote:
> Hi Hans,
> 
>>> Add irq_polarity module option for easier troubleshooting of irq-polarity
>>> issues.
>>> Cc: stable@vger.kernel.org
>>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
>>
>> Marcel, is there any reason why this (bug-fix) series is not merged
>> yet? Any review remarks I missed?
> 
> I was waiting for net to be merged back into net-next so that I don’t have to deal with conflicts.

Ah ok, no problem. I was just wondering if there was anything
I needed to do.

Regards,

Hans

Re: [PATCH 1/5] Bluetooth: hci_bcm: Add irq_polarity module option

[Marcel Holtmann]

Hi Hans,

>>>> Add irq_polarity module option for easier troubleshooting of irq-polarity
>>>> issues.
>>>> Cc: stable@vger.kernel.org
>>>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
>>> 
>>> Marcel, is there any reason why this (bug-fix) series is not merged
>>> yet? Any review remarks I missed?
>> I was waiting for net to be merged back into net-next so that I don’t have to deal with conflicts.
> 
> Ah ok, no problem. I was just wondering if there was anything
> I needed to do.

I would really like to move over to using btuart.c that I posted a few days ago, but that will be obviously more work and maybe it is good to stabilize and simplify this inside hci_bcm.c first.

Regards

Marcel

Re: [PATCH 1/5] Bluetooth: hci_bcm: Add irq_polarity module option

[Hans de Goede]

Hi Marcel,

On 23-03-18 12:23, Marcel Holtmann wrote:
> Hi Hans,
> 
>>>>> Add irq_polarity module option for easier troubleshooting of irq-polarity
>>>>> issues.
>>>>> Cc: stable@vger.kernel.org
>>>>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
>>>>
>>>> Marcel, is there any reason why this (bug-fix) series is not merged
>>>> yet? Any review remarks I missed?
>>> I was waiting for net to be merged back into net-next so that I don’t have to deal with conflicts.
>>
>> Ah ok, no problem. I was just wondering if there was anything
>> I needed to do.
> 
> I would really like to move over to using btuart.c that I posted a few days ago, but that will be obviously more work and maybe it is good to stabilize and simplify this inside hci_bcm.c first.

Yes I already saw your remark about this a few days ago. I do think that
moving to btuart.c eventually is a good idea, but your RFC seemed somewhat
feature incomplete. Mainly lacking the runtime-pm stuff, which is not
only needed for powersaving, but on some machines the shutdown gpio
defaults to "off", so the bluetooth won't work until we at a minimum
have added gpio support to btuart.c.

So for now I would like to get things fixed in hci_bcm.c first, as
you suggest. This will also allow easily getting the IRQ polarity and
missing IDs (at least those IDs which are confirmed to be actually
used) fixed in current stable kernels.

I'm afraid I don't have time to help much with btuart.c my hci_bcm.c
work is a small part of a project I'm working on to better support
Bay and Cherry Trail devices and that project itself is a personal
side-project. But when btuart.c is more feature complete I would be
happy to test it on various devices I own which have an uart attach
broadcom bt device.

Regards,

Hans

Re: [PATCH 1/5] Bluetooth: hci_bcm: Add irq_polarity module option

[Marcel Holtmann]

Hi Hans,

> Add irq_polarity module option for easier troubleshooting of irq-polarity
> issues.
> 
> Cc: stable@vger.kernel.org
> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> ---
> drivers/bluetooth/hci_bcm.c | 20 +++++++++++++++-----
> 1 file changed, 15 insertions(+), 5 deletions(-)

all 5 patches have been applied to bluetooth-next tree.

Regards

Marcel