From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jean-Christian de Rivaz Subject: Re: [PATCH 1/1] Add poll method to mkiss let notify hangup to the user process. Date: Sat, 03 Oct 2015 03:02:32 +0200 Message-ID: <560F2928.3080102@eclis.ch> References: <560EBE07.70801@eclis.ch> <1443822373-18395-1-git-send-email-jc@eclis.ch> <560F217B.6020106@trinnet.net> Mime-Version: 1.0 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <560F217B.6020106@trinnet.net> Sender: linux-hams-owner@vger.kernel.org List-ID: Content-Type: text/plain; charset="iso-8859-1"; format="flowed" To: David Ranch , Thomas Osterried , Ralf Baechle DL5RB , Greg Kroah-Hartman , Jiri Slaby , Peter Hurley Cc: linux-hams@vger.kernel.org Le 03. 10. 15 02:29, David Ranch a =E9crit : > > Hello Jean-Christian, Everyone, > > Thanks for working on this as I'm pretty sure I've bit hit by this=20 > panic as well though I wasn't able to reproduce it more readily.=20 > Anyway, if this is the proper fix moving forward, will the kernel not= =20 > panic even if kissattach is not updated? Hello David, The kernel will still panic and crash the system. The patch purpose is=20 only to allow hangup notification to the user space. It has the side=20 effect to reduce the race window of the panic&crash only if the user=20 application get the hangup notification and close the file descriptor.=20 The window is reduced in that case, but the window still exists. Even with this patch and the user space application patch, the kernel i= s=20 still unsafe, as a packet can still reach the parasitic AX25 interface=20 in the time between the kiss_open() in the kernel and the close() in th= e=20 user application. More important, as I stated before, safety of the=20 kernel cannot depend on a user space application closing a file=20 descriptor. A fix to prevent mkiss_open() call on a removed serial=20 device is still needed. I proposed patches and options, but to date I=20 only get two critics and no hint on how to solve the problem. The tty=20 maintainers stay silent. > Can you also include the needed patch for the kissattach program to=20 > complete this solution? Yes, it was posted a few minutes before in the linux-hams mailing list=20 only because it don't concern the kernel mailing list: http://www.spinics.net/lists/linux-hams/msg03520.html Jean-Christian -- To unsubscribe from this list: send the line "unsubscribe linux-hams" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html