[PATCH] [v2,1/1] This aims to support Foxconn SDX55

[PATCH] [v2,1/1] This aims to support Foxconn SDX55

[Slark Xiao]

Foxconn SDX55 T77W175 device is working in PCIe mode normally.
You can find it in drivers/bus/mhi/pci_geneirc.c file.
But in some scenario, we need to capture the memory dump once it crashed.
So a diag port driver is needed.

Signed-off-by: Slark Xiao <slark_xiao@163.com>
---
 drivers/usb/serial/qcserial.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/usb/serial/qcserial.c b/drivers/usb/serial/qcserial.c
index 83da8236e3c8..d8b58aea3c60 100644
--- a/drivers/usb/serial/qcserial.c
+++ b/drivers/usb/serial/qcserial.c
@@ -111,6 +111,7 @@ static const struct usb_device_id id_table[] = {
 	{USB_DEVICE(0x16d8, 0x8002)},	/* CMDTech Gobi 2000 Modem device (VU922) */
 	{USB_DEVICE(0x05c6, 0x9204)},	/* Gobi 2000 QDL device */
 	{USB_DEVICE(0x05c6, 0x9205)},	/* Gobi 2000 Modem device */
+	{USB_DEVICE(0x05c6, 0x901d)},	/* Foxconn SDX55 QDL */
 
 	/* Gobi 3000 devices */
 	{USB_DEVICE(0x03f0, 0x371d)},	/* HP un2430 Gobi 3000 QDL */
-- 
2.25.1

Re: [PATCH] [v2,1/1] This aims to support Foxconn SDX55

[Sergei Shtylyov]

Hello!

On 11.08.2021 11:56, Slark Xiao wrote:

> Foxconn SDX55 T77W175 device is working in PCIe mode normally.
> You can find it in drivers/bus/mhi/pci_geneirc.c file.

    pci_generic? :-)

> But in some scenario, we need to capture the memory dump once it crashed.
> So a diag port driver is needed.
> 
> Signed-off-by: Slark Xiao <slark_xiao@163.com>
[...]

MBR, Sergei

Re: [PATCH] [v2,1/1] This aims to support Foxconn SDX55

[Greg KH]

On Wed, Aug 11, 2021 at 04:56:35PM +0800, Slark Xiao wrote:
> Foxconn SDX55 T77W175 device is working in PCIe mode normally.
> You can find it in drivers/bus/mhi/pci_geneirc.c file.
> But in some scenario, we need to capture the memory dump once it crashed.
> So a diag port driver is needed.
> 
> Signed-off-by: Slark Xiao <slark_xiao@163.com>
> ---
>  drivers/usb/serial/qcserial.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/usb/serial/qcserial.c b/drivers/usb/serial/qcserial.c
> index 83da8236e3c8..d8b58aea3c60 100644
> --- a/drivers/usb/serial/qcserial.c
> +++ b/drivers/usb/serial/qcserial.c
> @@ -111,6 +111,7 @@ static const struct usb_device_id id_table[] = {
>  	{USB_DEVICE(0x16d8, 0x8002)},	/* CMDTech Gobi 2000 Modem device (VU922) */
>  	{USB_DEVICE(0x05c6, 0x9204)},	/* Gobi 2000 QDL device */
>  	{USB_DEVICE(0x05c6, 0x9205)},	/* Gobi 2000 Modem device */
> +	{USB_DEVICE(0x05c6, 0x901d)},	/* Foxconn SDX55 QDL */

Why is this not sorted?

And the subject needs some work still...

thanks,

greg k-h

Re: [PATCH] [v2,1/1] This aims to support Foxconn SDX55

[Bjørn Mork]

Slark Xiao <slark_xiao@163.com> writes:

> diff --git a/drivers/usb/serial/qcserial.c b/drivers/usb/serial/qcserial.c
> index 83da8236e3c8..d8b58aea3c60 100644
> --- a/drivers/usb/serial/qcserial.c
> +++ b/drivers/usb/serial/qcserial.c
> @@ -111,6 +111,7 @@ static const struct usb_device_id id_table[] = {
>  	{USB_DEVICE(0x16d8, 0x8002)},	/* CMDTech Gobi 2000 Modem device (VU922) */
>  	{USB_DEVICE(0x05c6, 0x9204)},	/* Gobi 2000 QDL device */
>  	{USB_DEVICE(0x05c6, 0x9205)},	/* Gobi 2000 Modem device */
> +	{USB_DEVICE(0x05c6, 0x901d)},	/* Foxconn SDX55 QDL */

I assume this device will expose other serial functions when booted in
application USB mode?  But probably not with a Gobi 2k layout... Maybe
add the application device ID to some USB serial driver too, and include
the QDL device ID there as well to reduce confusion?

Could you provide a view of the default USB descriptors in both QDL and
application mode?  E.g from lsusb -v or  /sys/kernel/debug/usb/devices?


Bjørn

Re:Re: [PATCH] [v2,1/1] This aims to support Foxconn SDX55

[Slark Xiao]

At 2021-08-11 18:58:09, "Sergei Shtylyov" <sergei.shtylyov@gmail.com> wrote:
>Hello!
>
>On 11.08.2021 11:56, Slark Xiao wrote:
>
>> Foxconn SDX55 T77W175 device is working in PCIe mode normally.
>> You can find it in drivers/bus/mhi/pci_geneirc.c file.
>
>    pci_generic? :-)
>
>> But in some scenario, we need to capture the memory dump once it crashed.
>> So a diag port driver is needed.
>> 
>> Signed-off-by: Slark Xiao <slark_xiao@163.com>
>[...]
>
>MBR, Sergei

Hi Sergei,
   Yes, I didn't mix them up. Our product supports 3 modes: PCIE only, USB only, and PCIE+USB.
   Normally, it would be PCIE mode. When we want to reproduce some crash issue, we will switch it from PCIE mode to PCIE+USB mode.
  In this scenario, the USB port is used for capturing memory dump as our tool is designed based on Qualcomm USB Sahara protocol.

Thanks.

Re:Re: [PATCH] [v2,1/1] This aims to support Foxconn SDX55

[Slark Xiao]

At 2021-08-11 19:42:20, "Greg KH" <gregkh@linuxfoundation.org> wrote:
>On Wed, Aug 11, 2021 at 04:56:35PM +0800, Slark Xiao wrote:
>> Foxconn SDX55 T77W175 device is working in PCIe mode normally.
>> You can find it in drivers/bus/mhi/pci_geneirc.c file.
>> But in some scenario, we need to capture the memory dump once it crashed.
>> So a diag port driver is needed.
>> 
>> Signed-off-by: Slark Xiao <slark_xiao@163.com>
>> ---
>>  drivers/usb/serial/qcserial.c | 1 +
>>  1 file changed, 1 insertion(+)
>> 
>> diff --git a/drivers/usb/serial/qcserial.c b/drivers/usb/serial/qcserial.c
>> index 83da8236e3c8..d8b58aea3c60 100644
>> --- a/drivers/usb/serial/qcserial.c
>> +++ b/drivers/usb/serial/qcserial.c
>> @@ -111,6 +111,7 @@ static const struct usb_device_id id_table[] = {
>>  	{USB_DEVICE(0x16d8, 0x8002)},	/* CMDTech Gobi 2000 Modem device (VU922) */
>>  	{USB_DEVICE(0x05c6, 0x9204)},	/* Gobi 2000 QDL device */
>>  	{USB_DEVICE(0x05c6, 0x9205)},	/* Gobi 2000 Modem device */
>> +	{USB_DEVICE(0x05c6, 0x901d)},	/* Foxconn SDX55 QDL */
>
>Why is this not sorted?
>
>And the subject needs some work still...
>
>thanks,
>
>greg k-h

Hi Greg,
  Sorry,  can you show me where should I put it at?
  According to Bjørn comment, I think I need to create a new layout for Foxconn SDX55 device.

Thanks.

Re:Re: [PATCH] [v2,1/1] This aims to support Foxconn SDX55

[Slark Xiao]

At 2021-08-11 20:01:57, "Bjørn Mork" <bjorn@mork.no> wrote:
>Slark Xiao <slark_xiao@163.com> writes:
>
>> diff --git a/drivers/usb/serial/qcserial.c b/drivers/usb/serial/qcserial.c
>> index 83da8236e3c8..d8b58aea3c60 100644
>> --- a/drivers/usb/serial/qcserial.c
>> +++ b/drivers/usb/serial/qcserial.c
>> @@ -111,6 +111,7 @@ static const struct usb_device_id id_table[] = {
>>  	{USB_DEVICE(0x16d8, 0x8002)},	/* CMDTech Gobi 2000 Modem device (VU922) */
>>  	{USB_DEVICE(0x05c6, 0x9204)},	/* Gobi 2000 QDL device */
>>  	{USB_DEVICE(0x05c6, 0x9205)},	/* Gobi 2000 Modem device */
>> +	{USB_DEVICE(0x05c6, 0x901d)},	/* Foxconn SDX55 QDL */
>
>I assume this device will expose other serial functions when booted in
>application USB mode?  But probably not with a Gobi 2k layout... Maybe
>add the application device ID to some USB serial driver too, and include
>the QDL device ID there as well to reduce confusion?
>
>Could you provide a view of the default USB descriptors in both QDL and
>application mode?  E.g from lsusb -v or  /sys/kernel/debug/usb/devices?
>
>
>Bjørn

Hi Bjørn,
  Thank you for your good advice. Actually it's my typo for 'QDL'. It should be 'Diag' port.
  Please see my local setting for PCIE+USB mode:

 jbd@jbd-ThinkPad-P1-Gen-4:~/git/net-next$ lsusb | grep 05c6
Bus 003 Device 003: ID 05c6:901d Qualcomm, Inc. Qualcomm HS-USB Android Diag 901D
jbd@jbd-ThinkPad-P1-Gen-4:~$ lspci | grep Foxconn
0000:08:00.0 Wireless controller [0d40]: Foxconn International, Inc. Device e0ab

  One more, should I add a new layout for Qualcomm SDX55 platform? As there are G1K,G2K, SIERRA and  HUAWEI layout, none of them includes Qualcomm SDX55.

Thanks

Re: Re: [PATCH] [v2,1/1] This aims to support Foxconn SDX55

[Greg KH]

On Thu, Aug 12, 2021 at 09:52:56AM +0800, Slark Xiao wrote:
> At 2021-08-11 19:42:20, "Greg KH" <gregkh@linuxfoundation.org> wrote:
> >On Wed, Aug 11, 2021 at 04:56:35PM +0800, Slark Xiao wrote:
> >> Foxconn SDX55 T77W175 device is working in PCIe mode normally.
> >> You can find it in drivers/bus/mhi/pci_geneirc.c file.
> >> But in some scenario, we need to capture the memory dump once it crashed.
> >> So a diag port driver is needed.
> >> 
> >> Signed-off-by: Slark Xiao <slark_xiao@163.com>
> >> ---
> >>  drivers/usb/serial/qcserial.c | 1 +
> >>  1 file changed, 1 insertion(+)
> >> 
> >> diff --git a/drivers/usb/serial/qcserial.c b/drivers/usb/serial/qcserial.c
> >> index 83da8236e3c8..d8b58aea3c60 100644
> >> --- a/drivers/usb/serial/qcserial.c
> >> +++ b/drivers/usb/serial/qcserial.c
> >> @@ -111,6 +111,7 @@ static const struct usb_device_id id_table[] = {
> >>  	{USB_DEVICE(0x16d8, 0x8002)},	/* CMDTech Gobi 2000 Modem device (VU922) */
> >>  	{USB_DEVICE(0x05c6, 0x9204)},	/* Gobi 2000 QDL device */
> >>  	{USB_DEVICE(0x05c6, 0x9205)},	/* Gobi 2000 Modem device */
> >> +	{USB_DEVICE(0x05c6, 0x901d)},	/* Foxconn SDX55 QDL */
> >
> >Why is this not sorted?
> >
> >And the subject needs some work still...
> >
> >thanks,
> >
> >greg k-h
> 
> Hi Greg,
>   Sorry,  can you show me where should I put it at?

If it belongs in this list, then it should be in sorted order, which you
did not do here, so that would be a good idea, right?

>   According to Bjørn comment, I think I need to create a new layout for Foxconn SDX55 device.

That is up to you and Bjørn to work out :)

thanks,

greg k-h

Re: [PATCH] [v2,1/1] This aims to support Foxconn SDX55

[Sergei Shtylyov]

Hello!

On 12.08.2021 4:08, Slark Xiao wrote:

> HI Sergei,
>    Yes. Our produt has 3 mode: PCIe mode only, USB mode only, and PCIE+USB.
>    When we need to reproduce some crash issue, we will let it enter into 
> PCIE+USB mode.
>    In this scenario, USB would have a diag port to capture memory dump.

    That's all fine, I was just pointing out the typo. :-)
    Couldn't open your another reply...

> At 2021-08-11 18:58:09, "Sergei Shtylyov" <sergei.shtylyov@gmail.com> wrote:
>>Hello!
>>
>>On 11.08.2021 11:56, Slark Xiao wrote:
>>
>>> Foxconn SDX55 T77W175 device is working in PCIe mode normally.
>>> You can find it in drivers/bus/mhi/pci_geneirc.c file.
>>
>>    pci_generic? :-)
>>
>>> But in some scenario, we need to capture the memory dump once it crashed.
>>> So a diag port driver is needed.
>>> 
>>> Signed-off-by: Slark Xiao <slark_xiao@163.com>

[...]

MBR, Sergei

Re:Re: Re: [PATCH] [v2,1/1] This aims to support Foxconn SDX55

[Slark Xiao]

At 2021-08-12 14:46:54, "Greg KH" <gregkh@linuxfoundation.org> wrote:
>On Thu, Aug 12, 2021 at 09:52:56AM +0800, Slark Xiao wrote:
>> At 2021-08-11 19:42:20, "Greg KH" <gregkh@linuxfoundation.org> wrote:
>> >On Wed, Aug 11, 2021 at 04:56:35PM +0800, Slark Xiao wrote:
>> >> Foxconn SDX55 T77W175 device is working in PCIe mode normally.
>> >> You can find it in drivers/bus/mhi/pci_geneirc.c file.
>> >> But in some scenario, we need to capture the memory dump once it crashed.
>> >> So a diag port driver is needed.
>> >> 
>> >> Signed-off-by: Slark Xiao <slark_xiao@163.com>
>> >> ---
>> >>  drivers/usb/serial/qcserial.c | 1 +
>> >>  1 file changed, 1 insertion(+)
>> >> 
>> >> diff --git a/drivers/usb/serial/qcserial.c b/drivers/usb/serial/qcserial.c
>> >> index 83da8236e3c8..d8b58aea3c60 100644
>> >> --- a/drivers/usb/serial/qcserial.c
>> >> +++ b/drivers/usb/serial/qcserial.c
>> >> @@ -111,6 +111,7 @@ static const struct usb_device_id id_table[] = {
>> >>  	{USB_DEVICE(0x16d8, 0x8002)},	/* CMDTech Gobi 2000 Modem device (VU922) */
>> >>  	{USB_DEVICE(0x05c6, 0x9204)},	/* Gobi 2000 QDL device */
>> >>  	{USB_DEVICE(0x05c6, 0x9205)},	/* Gobi 2000 Modem device */
>> >> +	{USB_DEVICE(0x05c6, 0x901d)},	/* Foxconn SDX55 QDL */
>> >
>> >Why is this not sorted?
>> >
>> >And the subject needs some work still...
>> >
>> >thanks,
>> >
>> >greg k-h
>> 
>> Hi Greg,
>>   Sorry,  can you show me where should I put it at?
>
>If it belongs in this list, then it should be in sorted order, which you
>did not do here, so that would be a good idea, right?
>
>>   According to Bjørn comment, I think I need to create a new layout for Foxconn SDX55 device.
>
>That is up to you and Bjørn to work out :)
>
>thanks,
>
>greg k-h

Hi Greg, Bjørn,
   How about move this change to /drivers/usb/serial/option.c ? I just test it and everything is ok for my device.
   And then we can ignore the layout issue.

Thanks