linux-kernel - Re: UART/TTY console deadlock

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 for Android: free password hash cracker in your pocket

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

Date:   Sat, 4 Jul 2020 14:37:14 +0300
From:   Andy Shevchenko <andy.shevchenko@...il.com>
To:     Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Cc:     Tony Lindgren <tony@...mide.com>, Petr Mladek <pmladek@...e.com>,
        Raul Rangel <rrangel@...gle.com>,
        Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>,
        linux-kernel <linux-kernel@...r.kernel.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        kurt@...utronix.de, "S, Shirish" <Shirish.S@....com>,
        Peter Zijlstra <peterz@...radead.org>,
        John Ogness <john.ogness@...utronix.de>,
        Steven Rostedt <rostedt@...dmis.org>
Subject: Re: UART/TTY console deadlock

On Fri, Jul 3, 2020 at 1:53 PM Sergey Senozhatsky
<sergey.senozhatsky@...il.com> wrote:
>
> On (20/07/02 11:20), Andy Shevchenko wrote:
> >
> > I didn't look into this deeply, but my understanding that this is something for
> > special case when you have several UART ports sharing the IRQ (multi-port card)
> > and IRQ even maybe undesirable b/c it will confuse real IRQ handler. I don't
> > remember details, but AFAIR IRQ handler does a busyloop to service as much as
> > possible and in between it may release the lock (again, multi-port UART cards),
> > that's why we better avoid IRQ event in the first place.
> > But it's my pure speculation.
>
> Hmm. Would different IRQ handlers use different UART port structures? ->irq
> might be the only thing they share. Even if they use the same port, we
> keep IRQs disabled on local CPU, and UART port locked. To unlock the UART
> port during IRQ, the handler (on another CPU) first needs to acquire the
> port->lock, which is locked by serial8250_do_startup().

It appears to be a RT kernel related. Don't know the state of affairs
in RT nowadays, but maybe still relevant.
See the commit 768aec0b5bcc ("serial: 8250: fix shared interrupts
issues with SMP and RT kernels").


-- 
With Best Regards,
Andy Shevchenko