| Message ID | af2c8c27124f1988fc439fe74f8cfe3cd9f47b48.1551259482.git.mirq-linux@rere.qmqm.pl (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | usb: gadget: u_serial: console for multiple ports | expand |
On Wed, Feb 27, 2019 at 10:29:38AM +0100, Michał Mirosław wrote: > /* management of individual TTY ports */ > +int gserial_alloc_line_raw(unsigned char *port_line); > int gserial_alloc_line(unsigned char *port_line); What is the difference between "raw" and "not raw"? I can't even answer that question, and I maintain the tty layer, so what hope is there for someone else? :) nameing is hard, I know... thanks, greg k-h
On Wed, Feb 27, 2019 at 10:36:25AM +0100, Greg Kroah-Hartman wrote: > On Wed, Feb 27, 2019 at 10:29:38AM +0100, Michał Mirosław wrote: > > /* management of individual TTY ports */ > > +int gserial_alloc_line_raw(unsigned char *port_line); > > int gserial_alloc_line(unsigned char *port_line); > What is the difference between "raw" and "not raw"? I can't even answer > that question, and I maintain the tty layer, so what hope is there for > someone else? :) > > nameing is hard, I know... Yes it is. gserial_alloc_line_no_console() then? Best Regards, Michał Mirosław
On Wed, Feb 27, 2019 at 11:14:27AM +0100, Michał Mirosław wrote: > On Wed, Feb 27, 2019 at 10:36:25AM +0100, Greg Kroah-Hartman wrote: > > On Wed, Feb 27, 2019 at 10:29:38AM +0100, Michał Mirosław wrote: > > > /* management of individual TTY ports */ > > > +int gserial_alloc_line_raw(unsigned char *port_line); > > > int gserial_alloc_line(unsigned char *port_line); > > What is the difference between "raw" and "not raw"? I can't even answer > > that question, and I maintain the tty layer, so what hope is there for > > someone else? :) > > > > nameing is hard, I know... > > Yes it is. gserial_alloc_line_no_console() then? That makes a lot more sense to me, thank you. greg k-h
diff --git a/drivers/usb/gadget/function/f_obex.c b/drivers/usb/gadget/function/f_obex.c index 55b7f57d2dc7..3f11f41619f9 100644 --- a/drivers/usb/gadget/function/f_obex.c +++ b/drivers/usb/gadget/function/f_obex.c @@ -432,7 +432,7 @@ static struct usb_function_instance *obex_alloc_inst(void) return ERR_PTR(-ENOMEM); opts->func_inst.free_func_inst = obex_free_inst; - ret = gserial_alloc_line(&opts->port_num); + ret = gserial_alloc_line_raw(&opts->port_num); if (ret) { kfree(opts); return ERR_PTR(ret); diff --git a/drivers/usb/gadget/function/u_serial.c b/drivers/usb/gadget/function/u_serial.c index 8d2d861e1543..3466d94f1441 100644 --- a/drivers/usb/gadget/function/u_serial.c +++ b/drivers/usb/gadget/function/u_serial.c @@ -1179,7 +1179,7 @@ void gserial_free_line(unsigned char port_num) } EXPORT_SYMBOL_GPL(gserial_free_line); -int gserial_alloc_line(unsigned char *line_num) +int gserial_alloc_line_raw(unsigned char *line_num) { struct usb_cdc_line_coding coding; struct gs_port *port; @@ -1220,12 +1220,20 @@ int gserial_alloc_line(unsigned char *line_num) goto err; } *line_num = port_num; - - if (!port_num) - gs_console_init(port); err: return ret; } +EXPORT_SYMBOL_GPL(gserial_alloc_line_raw); + +int gserial_alloc_line(unsigned char *line_num) +{ + int ret = gserial_alloc_line_raw(line_num); + + if (!ret && !*line_num) + gs_console_init(ports[*line_num].port); + + return ret; +} EXPORT_SYMBOL_GPL(gserial_alloc_line); /** diff --git a/drivers/usb/gadget/function/u_serial.h b/drivers/usb/gadget/function/u_serial.h index 9acaac1cbb75..a86bd3ce4781 100644 --- a/drivers/usb/gadget/function/u_serial.h +++ b/drivers/usb/gadget/function/u_serial.h @@ -54,6 +54,7 @@ struct usb_request *gs_alloc_req(struct usb_ep *ep, unsigned len, gfp_t flags); void gs_free_req(struct usb_ep *, struct usb_request *req); /* management of individual TTY ports */ +int gserial_alloc_line_raw(unsigned char *port_line); int gserial_alloc_line(unsigned char *port_line); void gserial_free_line(unsigned char port_line);
Prevent OBEX serial port from ever becoming a console. Console messages will definitely break the protocol, and since you have to instantiate the port making it explicitly for OBEX, there is no point in allowing console to break it by mistake. Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl> --- v2: change of API + commit message massage --- drivers/usb/gadget/function/f_obex.c | 2 +- drivers/usb/gadget/function/u_serial.c | 16 ++++++++++++---- drivers/usb/gadget/function/u_serial.h | 1 + 3 files changed, 14 insertions(+), 5 deletions(-)