linux-kernel - Re: [PATCH] ACPI: Do not fail acpi_bind_one() if device is already bound correctly

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]

Date:	Mon, 05 Aug 2013 16:20:19 -0600
From:	Toshi Kani <toshi.kani@...com>
To:	"Rafael J. Wysocki" <rjw@...k.pl>
Cc:	ACPI Devel Maling List <linux-acpi@...r.kernel.org>,
	LKML <linux-kernel@...r.kernel.org>,
	Bjorn Helgaas <bhelgaas@...gle.com>
Subject: Re: [PATCH] ACPI: Do not fail acpi_bind_one() if device is already
 bound correctly

On Sun, 2013-08-04 at 16:03 +0200, Rafael J. Wysocki wrote:
> On Saturday, August 03, 2013 06:32:02 PM Toshi Kani wrote:
> > On Sat, 2013-08-03 at 02:47 +0200, Rafael J. Wysocki wrote:
> > > On Friday, August 02, 2013 04:38:38 PM Toshi Kani wrote:
> > > > On Fri, 2013-08-02 at 00:33 +0200, Rafael J. Wysocki wrote:
> > > > > From: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
> > > > > 
> > > > > Modify acpi_bind_one() so that it doesn't fail if the device
> > > > > represented by its first argument has already been bound to the
> > > > > given ACPI handle (second argument), because that is not a good
> > > > > enough reason for returning an error code.
> > > > 
> > > > While it seems reasonable to allow such case, I do not think we will hit
> > > > this case under the normal scenarios.  So, I do not think we need to
> > > > make this change now unless it actually solves Yasuaki's issue (which I
> > > > am guessing not).
> > > 
> > > In theory it should be possible to call acpi_bind_one() twice in a row
> > > for the same dev and the same handle without failure, that simply is
> > > logical.  The patch may not fix any problems visible now, but returning an
> > > error code in such a case is simply incorrect.
> > 
> > We changed acpi_bus_device_attach() to not call the handler or driver
> > again if it is already bound.  So, I was under impression that we
> > prevent from attaching a same device twice.  But I may be missing
> > something...
> 
> acpi_bind_one() may be called in code paths that don't start from
> acpi_bus_device_attach(), like acpi_platform_notify(), where the result
> depends on how .find_device() is implemented by the the given bus type,
> for example.

acpi_bind_one() always returns with 0 when it sets a handle to the
device.  So, acpi_platform_notify() should not call .find_device() in
this case.  acpi_bus_check_add() is also protected from adding a same
device twice.  But I see your point that the callers of acpi_bind_one()
could be changed / implementation dependent.  So, I agree that it would
be prudent to have this change. 

Acked-by: Toshi Kani <toshi.kani@...com>

Thanks,
-Toshi

--
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/