[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20110401084218.d803b38d.rdunlap@xenotime.net>
Date: Fri, 1 Apr 2011 08:42:18 -0700
From: Randy Dunlap <rdunlap@...otime.net>
To: Jeremy Jackson <jerj@...lanar.net>
Cc: Dmitry Torokhov <dmitry.torokhov@...il.com>,
linux-kernel@...r.kernel.org
Subject: Re: signal to PID 1 from serial console like KeyboardRequest or CAD
On Fri, 01 Apr 2011 11:39:28 -0400 Jeremy Jackson wrote:
> On 04/01/2011 01:37 AM, Dmitry Torokhov wrote:
> > On Thu, Mar 31, 2011 at 12:57:06PM -0700, Randy Dunlap wrote:
> >> On Fri, 25 Mar 2011 13:41:40 -0400 Jeremy Jackson wrote:
> >>
> >>> Hi,
> >>>
> >>> Sent to linux-console with no reply for over a week...
> >>>
> >>> I ran into a problem recently on a Linux server with a remote serial
> >>> console. /sbin/init (actually Upstart) was in a state where nothing
> >>> (useful) was running. I had to power cycle the machine to get back to
> >>> the boot prompt. This has the side-effect of triggering raid rebuilds.
> >>>
> >>> It would be nice to have what Ctrl-Alt-Del does at the VGA console with
> >>> PS/2 or USB keyboard. This sends a SIGINT to PID 1, and most distros
> >>> have this set to do "shutdown -r now" which attempts a clean reboot.
> >>
> >> so C_A_D works at console via PS/2 or USB keyboard, but not on a remote
> >> serial console? Could it be an issue with whatever remote/serial software
> >> you are using?
> >
> > I believe it is simply a matter of console keymap and assigning "boot"
> > value to the desired combination.
>
> The VGA console/keyboard isn't an issue here, C_A_D works great. The
> problem is, that functionality isn't available when using *serial* console.
That's why I asked whether it could be an issue with the serial interface
software. E.g., minicom handles/processes certain keystrokes, other apps do things
differently, etc...
> >>> Once upon a time there was KeyboardRequest, which would send a SIGWINCH
> >>> to PID 1 (when pressing Alt-UP_arrow?). I don't see any sign of this
> >>> feature in Linux 2.6.32 though.
> >>
> >> I couldn't find it in 2.4.37 either.
> >>
> >> Can 'loadkeys' help you any?
> >>
> >>> Looking at the Sysrq handlers in the kernel I don't see any keystrokes
> >>> that send SIGINT to PID 1 either.
> >>>
> >>> So, I'm wondering if it makes sense to add a Sysrq key to send a signal
> >>> to PID1 ?
> >>
> >> Dmitry, any ideas?
> >
> > We have SysRq bindings to unmount filesystems and reboot, isn't that
> > enough?
>
> Reboot isn't clean... it doesn't signal MD-raid to "unmount" does it?
> SysRq powerOff does cleanly shutdown md-raid, but who will drive to the
> datacenter to press the power button to turn it back on?
>
> I guess there are two ways to achieve clean (md-raid) reboot when using
> serial console... a new SysRq to reboot "clean", or a new SysRq to
> signal init. Looking at ctrlaltdel(8), perhaps it could just one SysRq,
> and let userspace (/sbin/init) configure it as needed.
>
>
> --
> Coplanar Networks
> http://www.coplanar.net
> (519)489-4903
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@...r.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
---
~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists