Re: [PATCH 1/1] pty: BREAK for pseudoterminals

[Petr Tesarik <ptesarik@suse.cz>]

On Tue, 17 Feb 2015 14:03:58 -0500
Peter Hurley <peter@hurleysoftware.com> wrote:

> On 02/16/2015 08:04 AM, Peter Hurley wrote:
> > On 02/05/2015 02:11 PM, Nan Li wrote:
> >> This will greatly enhance the usefulness of QEMU virtual serial ports, because the Linux kernel interprets a break on the serial console as a SysRq, but there is currently no way to pass this signal over a pseudo-terminal. This patch will work for transmitting BREAK from master to slave through pseudo-terminal.
> > 
> > pty is an IPC mechanism, not a virtualization driver.
> > 
> > Those programs that know they're consuming from a slave pty may not
> > have bothered to set termios for BREAKs, and may not be prepared
> > for NULLs in the input stream.
> 
> I'm not even seeing how this works as advertised.
> 
> In QEMU, the *master* end is the source/sink for the guest VM, while the
> *slave* pty is sink/source for the host. Like this:

Yes, you're totally right. This patch implements break handling on the
"wrong" side of the pty pair. I think you have already pointed this
out in one of the previous mails, and we're waiting for a patch that
implements it on the slave side.

Regards,
Petr Tesarik

> Host <--> slave pty <===> master pty <--> qemu <==> emulated serial
> port
> 
> So this patch enables the guest VM to transmit BREAK to the host.
> Why is that useful?
> 
> Regards,
> Peter Hurley
> 
> >> Signed-off-by: Nan Li <nli@suse.com>
> >> ---
> >>  drivers/tty/pty.c | 19 +++++++++++++++++++
> >>  1 file changed, 19 insertions(+)
> >>
> >> diff --git a/drivers/tty/pty.c b/drivers/tty/pty.c
> >> index e72ee62..ac8893a 100644
> >> --- a/drivers/tty/pty.c
> >> +++ b/drivers/tty/pty.c
> >> @@ -31,6 +31,7 @@
> >>  static struct tty_driver *ptm_driver;
> >>  static struct tty_driver *pts_driver;
> >>  static DEFINE_MUTEX(devpts_mutex);
> >> +#define BREAK_STRING '\0'
> >>  #endif
> >>  
> >>  static void pty_close(struct tty_struct *tty, struct file *filp)
> >> @@ -674,6 +675,23 @@ static void pty_unix98_shutdown(struct
> >> tty_struct *tty) devpts_kill_index(tty->driver_data, tty->index);
> >>  }
> >>  
> >> +static int pty_unix98_break_ctl(struct tty_struct *tty, int
> >> break_state) +{
> >> +	int c;
> >> +	struct tty_struct *to = tty->link;
> >> +
> >> +	if (break_state) {
> >> +		/* Stuff the break data into the input queue of
> >> the other end */
> >> +		c = tty_insert_flip_char(to->port, BREAK_STRING,
> >> TTY_BREAK);
> >> +		/* And shovel */
> >> +		if (c)
> >> +			tty_flip_buffer_push(to->port);
> >> +		else
> >> +			return -ENOMEM;
> >> +	}
> >> +	return 0;
> >> +}
> >> +
> >>  static const struct tty_operations ptm_unix98_ops = {
> >>  	.lookup = ptm_unix98_lookup,
> >>  	.install = pty_unix98_install,
> >> @@ -686,6 +704,7 @@ static const struct tty_operations
> >> ptm_unix98_ops = { .chars_in_buffer = pty_chars_in_buffer,
> >>  	.unthrottle = pty_unthrottle,
> >>  	.ioctl = pty_unix98_ioctl,
> >> +	.break_ctl = pty_unix98_break_ctl,
> >>  	.resize = pty_resize,
> >>  	.shutdown = pty_unix98_shutdown,
> >>  	.cleanup = pty_cleanup
> >>
> > 
>