linux-kernel - Re: 2.6.24: Serial disabled in BIOS but serial modules still loaded (probably PnP related)

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:	Sun, 25 Nov 2007 23:01:33 +0300
From:	Andrey Borzenkov <arvidjaar@...l.ru>
To:	Robert Hancock <hancockr@...w.ca>
Cc:	linux-serial@...r.kernel.org, linux-kernel@...r.kernel.org,
	linux-acpi@...r.kernel.org
Subject: Re: 2.6.24: Serial disabled in BIOS but serial modules still loaded (probably PnP related)

On Sunday 25 November 2007, Robert Hancock wrote:
> Andrey Borzenkov wrote:
> > I have no COM port on notebook (without port replicator which I do not have)
> > so COM is disabled in BIOS. No ttyS* is detected during boot (and no device
> > created) but I just noticed that serial modules are still loaded. Well, this
> > partially defeats the purpose of disabling COM port - the intention was to
> > free resources by *not* loading unneeded modules ...
> > 
> > This may have something to do with (ACPI) PnP which apparently believes COM is alive.
> > Notebook is Toshiba Portege 4000.
> 
> Probably a BIOS bug. It still lists the port in PnP data even though the 
> hardware is disabled, so the kernel still tries to load the serial 
> driver for it, which finds there's no port there.
> 

Here is what we get from _STA:

[ 1689.988245] HID=PNP0501 flags=0d

If I read specs correctly, if bit 1 is cleared it means, device is not able
to access its hardware resources, so it makes no sense to even try to
load driver for it.

And of course it makes PnP output completely bogus:

00:09 PNP0501 16550A-compatible serial port
    state = active
    allocated resources:
        io 0x3f8-0x3ff
        irq 5
 
So two questions:

1. why device got enabled when ACPI explicitly stated it cannot be?
2. can we get information about _STA result in user space so we do not try
to autoload driver for device that can't work?

Download attachment "signature.asc " of type "application/pgp-signature" (190 bytes)