[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <YFT5woqIyNtypwXX@smile.fi.intel.com>
Date: Fri, 19 Mar 2021 21:21:38 +0200
From: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To: "Rafael J. Wysocki" <rafael@...nel.org>
Cc: "Rafael J. Wysocki" <rafael.j.wysocki@...el.com>,
ACPI Devel Maling List <linux-acpi@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
"open list:ACPI COMPONENT ARCHITECTURE (ACPICA)" <devel@...ica.org>,
"Rafael J. Wysocki" <rjw@...ysocki.net>,
Len Brown <lenb@...nel.org>,
Robert Moore <robert.moore@...el.com>,
Erik Kaneda <erik.kaneda@...el.com>
Subject: Re: [PATCH v2 1/1] ACPI: scan: Use unique number for instance_no
On Fri, Mar 19, 2021 at 07:51:00PM +0100, Rafael J. Wysocki wrote:
> On Fri, Mar 19, 2021 at 7:33 PM Andy Shevchenko
> <andriy.shevchenko@...ux.intel.com> wrote:
> >
> > Current mechanism of incrementing and decrementing plain integer
> > to get a next free instance_no when creating an ACPI device is buggy.
> >
> > The simple integer and operations line increment and decrement
> > on top of it can't cover the possible gaps during run time. The
> > arbitrary instantiation and elimination of the devices is racy
>
> But it isn't racy AFAICS. It always happens under acpi_device_lock().
Hmm.. indeed. I sent a v3 with the commit message based on your proposal.
> > and after a couple of iterations with unequal amount of devices
> > being added and removed we may reproduce a bug:
--
With Best Regards,
Andy Shevchenko
Powered by blists - more mailing lists