linux-kernel - Re: linux-next: dock_link

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] [day] [month] [year] [list]

Message-ID: <20120221223025.GA12989@homac.suse.de>
Date:	Tue, 21 Feb 2012 23:30:25 +0100
From:	Holger Macht <holger@...ac.de>
To:	Jeff Garzik <jeff@...zik.org>
Cc:	Hugh Dickins <hughd@...gle.com>, Hillf Danton <dhillf@...il.com>,
	Matthew Garrett <mjg@...hat.com>,
	Jeff Garzik <jgarzik@...hat.com>,
	Stephen Rothwell <sfr@...b.auug.org.au>,
	linux-kernel@...r.kernel.org,
	Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: linux-next: dock_link_device is oopsy

On Tue 21. Feb - 17:24:24, Jeff Garzik wrote:
> On 02/18/2012 02:57 PM, Holger Macht wrote:
> >On Sa 18. Feb - 10:46:04, Hugh Dickins wrote:
> >>On Sat, 18 Feb 2012, Holger Macht wrote:
> >>>How about that one?
> >>
> >>It's more broken than that.  Here's my attempt.  It boots on the
> >>systems with dock_station_count 0, and it boots on my laptop with
> >>dock_station_count 2; but I don't actually have any docking station,
> >>so it still doesn't test very much (dock is 0 after the loop).
> >
> >Well, there doesn't have to actually exist a physical dock station (or
> >bay device) for dock_station_count to be>  0. It just tells that the
> >ACPI objects are present and thus the system is capable of it.
> >
> >So does this function actually also break on your laptop and you're
> >getting the oops there, too?
> >
> >>I have no idea if what goes on in the loop is correct, but it looks
> >>to me as if (as predicted) there's further breakage, that it would
> >>have been writing beyond the end of what it allocated if I did have
> >>a docking station.
> >>
> >>Hugh
> >>
> >>[PATCH] dock: fix bootup oops and other dock_link breakage
> >>
> >>dock_link_device() and dock_unlink_device() should bail out early
> >>to avoid oops on zero-length kmalloc() when dock_station_count is 0.
> >>
> >>But isn't there an off-by-one in that kmalloc() length anyway?
> >>An extra NULL appended at the end suggests so.
> >>
> >>Rework the ordering with gotos on failure to fix several issues.
> >>
> >>And presumably dock_unlink_device() should be presenting the same
> >>interface as dock_link_device(), with NULL returned when none found.
> >>
> >>Signed-off-by: Hugh Dickins<hughd@...gle.com>
> >
> >Fine with me.
> 
> So, just to be clear, the preferred patch is Hugh's, and I should
> drop your earlier proposed fix found in this thread?

Correct.

> And what about that warning?

You mean the fix for the compile error when compiling with
CONFIG_ACPI_DOCK=n? Here it is again:

[PATCH] acpi: Fix compiler error when setting CONFIG_ACPI_DOCK=n

When compiling with CONFIG_ACPI_DOCK=n,
is_registered_hotplug_dock_device() needs to be defined

Signed-off-by: Holger Macht <holger@...ac.de>
---
 include/acpi/acpi_drivers.h |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/include/acpi/acpi_drivers.h b/include/acpi/acpi_drivers.h
index 3c4e381..3319574 100644
--- a/include/acpi/acpi_drivers.h
+++ b/include/acpi/acpi_drivers.h
@@ -155,6 +155,10 @@ static inline int register_hotplug_dock_device(acpi_handle handle,
 static inline void unregister_hotplug_dock_device(acpi_handle handle)
 {
 }
+static inline int is_registered_hotplug_dock_device(const struct acpi_dock_ops *ops)
+{
+	return 0;
+}
 static inline struct device **dock_link_device(acpi_handle handle)
 {
 	return NULL;
-- 
1.7.7

--
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/