linux-kernel - Re: sharing interrupt between PCI device

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]

Date:	Tue, 11 Nov 2008 12:49:03 -0600
From:	Robert Hancock <hancockr@...w.ca>
To:	"Altobelli, David" <david.altobelli@...com>
CC:	Nobin Mathew <nobin.mathew@...il.com>,
	Jiri Slaby <jirislaby@...il.com>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Ingo Molnar <mingo@...e.hu>,
	Grant Grundler <grundler@...isc-linux.org>
Subject: Re: sharing interrupt between PCI device

Altobelli, David wrote:
>>> On Mon, Nov 10, 2008 at 4:01 PM, Nobin Mathew <nobin.mathew@...il.com> wrote:
>>>> Yes USB Virtual Controller has input devices like keyboard and mouse
>>>> (routed through hpilo card (remote console)), so those things stops
>>>> working when we remove hpilo driver.
>>>>
> 
> That is not quite correct.  Nothing is routed through hpilo.
> 
> The problem is that there is a shared hardware resource in iLO function 2 (hpilo),
> and it is used by iLO function 4 (uhci_hcd kdb/mouse).  When hpilo is unloaded,
> the pci_device_disable() call removes the ability for this function to act as a
> busmaster, which stops the uhci_hcd from functioning correctly.
> 
> Reloading hpilo will enable the remote console again.
> 
> Removing the pci_disable_device() "fixes" the problem, but doing that worries me.

I suspect that removing the pci_disable_device in that driver is 
probably the best solution. It's not really mandatory to disable the 
device in all cases, we have other devices that have similar issues and 
so can't be disabled. As long as it can't generate interrupts or 
otherwise affect the system after the module is unloaded (but it sounds 
like it doesn't use interrupts anyway).

It's not really a good hardware design (mucking with one PCI function on 
the device shouldn't disable functionality on other functions), but what 
can you do..
--
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/