Re: [PATCH v5 2/2] console: Add persistent scrollback buffers for all VGA consoles

From: "Manuel Schölling" <manuel.schoelling@gmx.de>
To: Adam Borowski <kilobyte@angband.pl>
Cc: plagnioj@jcrosoft.com, tomi.valkeinen@ti.com, jslaby@suse.cz,
	gregkh@linuxfoundation.org, linux-fbdev@vger.kernel.org,
	linux-kernel@vger.kernel.org, andrey_utkin@fastmail.com
Subject: Re: [PATCH v5 2/2] console: Add persistent scrollback buffers for all VGA consoles
Date: Tue, 22 Nov 2016 17:56:42 +0100	[thread overview]
Message-ID: <1479833802.19820.8.camel@gmx.de> (raw)
In-Reply-To: <20161121201704.GA29655@angband.pl>

Hi Adam,

On Mo, 2016-11-21 at 21:17 +0100, Adam Borowski wrote:
> On Sun, Nov 20, 2016 at 10:58:08PM +0100, Manuel Schölling wrote:
> > Add a scrollback buffers for each VGA console. The benefit is that
> > the scrollback history is not flushed when switching between consoles
> > but is persistent.
> > The buffers are allocated on demand when a new console is opened.
> > 
> > This breaks tools like clear_console that rely on flushing the
> > scrollback history by switching back and forth between consoles
> > which is why this feature is disabled by default.
> > Use the escape sequence \e[3J instead for flushing the buffer.

> First, big thanks for this fix, it's something that greatly annoyed me
> since forever!
Yeah, me too! ;)

> The thing about clear_console is unfortunate: they abused the bug you're
> fixing.  I've asked to use \e[3J (https://bugs.debian.org/845177) so there's
> hope it'll be applied in stretch; with Debian configuring its glibc to
> support only kernels from two releases before (in jessie that's 2.6.32, in
> stretch 3.2)[1] there's hope we can flip the default in several years.
> 
> Do you suspect any other program relies on VT switch to clear the
> scrollback?
Not, AFAIK. Although I do not have a complete list of programs that are
suppose to do that.

> But alas, this commit breaks that very \e[3J.  It does only a \e[2J, leaving
> the scrollback uncleared.  For comparison, both mainline and with just your
> preparatory commit, \e[3J works as expected.
Really? All my tests worked fine: I compiled the kernel with the latest patches, started the kernel in QEMU and then did

  $ openvt /bin/sh
  $ echo -e '\e[3J' # scrollback buffer was flushed correctly
  $ chvt 2
  $ echo -e '\e[3J' # scrollback buffer was flushed correctly

Can you tell me how you tested it? Maybe I can reproduce the bug.

Thanks for spending the time to test it!

Bye,

Manuel