linux-kernel - Re: PROBLEM: Data corruption when pasting large data to terminal

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAGWcZkJ+uvg=GwFJ+AizCAope2rWw2wujJYUZMvz8GBkKr34sA@mail.gmail.com>
Date:	Thu, 16 Feb 2012 12:42:07 +0100
From:	Egmont Koblinger <egmont@...il.com>
To:	Greg KH <gregkh@...uxfoundation.org>
Cc:	Parag Warudkar <parag.lkml@...il.com>, linux-kernel@...r.kernel.org
Subject: Re: PROBLEM: Data corruption when pasting large data to terminal

On Thu, Feb 16, 2012 at 01:10, Greg KH <gregkh@...uxfoundation.org> wrote:
>
> Odds are emacs also does this correctly, anyone care to verify that?

Actually, thanks for the hint, emacs is the first editor where I could
easily trigger the bug.  I made the input data a bit longer (200
lines), and it's being pasted incorrectly all the time.

Stracing xterm contains exactly this, with the "= 68" at the end, 200 times:
write(5, "a=(12345678912345678912345678912"..., 68) = 68

Contrary to xterm, gnome-terminal doesn't split the buffer at
newlines, it tries to write to /dev/ptmx in one single step. In this
example I extended the file to 1000 lines long (68000 bytes), here's
what strace gnome-terminal says:
open("/dev/ptmx", O_RDWR) = 17
[...]
write(17, "a=(12345678912345678912345678912"..., 68000) = 65280
[...]
write(17, "a=(12345678912345678912345678912"..., 2720) = 2720

So copy-paste starts misbehaving after 4kB, which suggests a buffer
size of around that size; on the other hand, writes to /dev/ptmx can
return up to almost 64kB, which suggests a much larger terminal buffer
which makes it strange that it starts misbehaving as early as at 4kB.

On a side note, apparently gnome-terminal handles short writes
correctly. (It's a mere accident that the size of the first write,
65280, is dividable by the line length of my example, 68.  A different
input shows that the second write indeed continues sending the buffer
from the correct offset, from the middle of a line. The return value
of the first write is the same, 65280 in that case too.)

thx,
egmont
--
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/