[PATCH][Xenial SRU] UBUNTU: SAUCE: xr-usb-serial: re-initialise baudrate after resume from S3/S4

[Colin Ian King]

On 17/05/17 15:33, Seth Forshee wrote:
> On Fri, May 12, 2017 at 08:30:52PM +0800, Shrirang Bagul wrote:
>> BugLink: https://bugs.launchpad.net/bugs/1690362
>>
>> The XR21V1412 ports were using the hardware power-on default of 9600bps
>> after resume from S3/S4. This patch re-initialises the baud rate and
>> fixes the re-connection issues after S3/S4.
>>
>> Changelog:
>> Version 1B, 11/6/2015
>> Fixed Bug: The conditional logic to support kernel 3.9 was incorrect(line 396 in xr_usb_serial_common.c).
>>
>> Version 1A, 1/9/2015
>> This driver will work with any USB UART function in these Exar devices:
>>         XR21V1410/1412/1414
>>         XR21B1411
>>         XR21B1420/1422/1424
>>         XR22801/802/804
>>
>> The source code has been tested on various Linux kernels from 3.6.x to 3.17.x.
>> This may also work with newer kernels as well.
>>
>> Signed-off-by: Shrirang Bagul <shrirang.bagul at canonical.com>
>> ---
>>  ubuntu/xr-usb-serial/Makefile               |  3 ++-
>>  ubuntu/xr-usb-serial/xr_usb_serial_common.c | 14 ++++++++++++--
>>  2 files changed, 14 insertions(+), 3 deletions(-)
>>
>> diff --git a/ubuntu/xr-usb-serial/Makefile b/ubuntu/xr-usb-serial/Makefile
>> index 1a323a62b348..533988cede12 100644
>> --- a/ubuntu/xr-usb-serial/Makefile
>> +++ b/ubuntu/xr-usb-serial/Makefile
>> @@ -1,6 +1,7 @@
>>  obj-m := xr_usb_serial_common.o
>>  
>> -KERNELDIR ?= /lib/modules/$(shell uname -r)/build
>> +#KERNELDIR ?= /lib/modules/$(shell uname -r)/build
>> +KERNELDIR ?=$(shell pwd)/../parts/kernel/build
>>  PWD       := $(shell pwd)
>>  
>>  EXTRA_CFLAGS	:= -DDEBUG=0
>> diff --git a/ubuntu/xr-usb-serial/xr_usb_serial_common.c b/ubuntu/xr-usb-serial/xr_usb_serial_common.c
>> index 5d049855fb1a..09d21e63606e 100644
>> --- a/ubuntu/xr-usb-serial/xr_usb_serial_common.c
>> +++ b/ubuntu/xr-usb-serial/xr_usb_serial_common.c
>> @@ -1146,7 +1146,7 @@ static void xr_usb_serial_tty_set_termios(struct tty_struct *tty,
>>      }
>>  
>>  		
>> -	if (memcmp(&xr_usb_serial->line, &newline, sizeof newline))
>> +	//if (memcmp(&xr_usb_serial->line, &newline, sizeof newline))
>>  	{
>>  		memcpy(&xr_usb_serial->line, &newline, sizeof newline);
>>  		/*
>> @@ -1819,11 +1819,21 @@ static int xr_usb_serial_resume(struct usb_interface *intf)
>>  {
>>  	struct xr_usb_serial *xr_usb_serial = usb_get_intfdata(intf);
>>  	struct xr_usb_serial_wb *wb;
>> +#if 0	
>> +	struct tty_struct *tty = xr_usb_serial->port.tty;
>> +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 7, 0)	
>> +	struct ktermios *termios = tty->termios;
>> +#else
>> +    struct ktermios *termios = &tty->termios;
>> +#endif
>> +#endif
>>  	int rv = 0;
>>  	int cnt;
>>  	     
>>  	xr_usb_serial_pre_setup(xr_usb_serial);
>> -	
>> +	xr_usb_serial_disable(xr_usb_serial);
>> +	xr_usb_serial_set_line(xr_usb_serial, &xr_usb_serial->line);
>> +	xr_usb_serial_enable(xr_usb_serial);	
>>  	spin_lock_irq(&xr_usb_serial->read_lock);
>>  	xr_usb_serial->susp_count -= 1;
>>  	cnt = xr_usb_serial->susp_count;
> 
> This patch is sloppy and has some obviously unrelated changes (those to
> the makefile and adding in code that is commented out). So it's not very
> nice, but for this particular driver I guess it isn't that important so
> long as the changes work. There's no comments about testing on the bug;
> I assume it has been tested?
> 

personally speaking, I'd prefer it to be cleaned up first before we
apply it to our repo.

Colin