linux-kernel - Re: [PATCH 13/15] samsung-laptop: make the dmi check less strict

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]

Message-ID: <CAHR064igjTPmeeXrnO2BxNn_U0RXiw9fVJnERYj6f_Y1tChM0w@mail.gmail.com>
Date:	Wed, 23 Nov 2011 08:17:08 +0100
From:	Corentin Chary <corentin.chary@...il.com>
To:	Greg KH <gregkh@...e.de>
Cc:	platform-driver-x86@...r.kernel.org,
	Matthew Garrett <mjg@...hat.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 13/15] samsung-laptop: make the dmi check less strict

On Tue, Nov 22, 2011 at 11:27 PM, Greg KH <gregkh@...e.de> wrote:
> On Tue, Nov 22, 2011 at 11:02:44PM +0100, Corentin Chary wrote:
>> This enable the driver for everything that look like
>> a laptop and is from vendor "SAMSUNG ELECTRONICS CO., LTD.".
>> Note that laptop supported by samsung-q10 seem to have a different
>> vendor strict.
>>
>> Also remove every log output until we know that we have a SABI interface
>> (except if the driver is forced to load, or debug is enabled).
>>
>> Keeping a whitelist of laptop with a model granularity is something that can't
>> work without close vendor cooperation (and we don't have that).
>>
>> Signed-off-by: Corentin Chary <corentincj@...aif.net>
>
> Oh, this is bold.
>
> I would like to see this work, but I would need to see it tested on a
> bunch of machines before I trust it.
>
> I guess you now have the acpi video backlight check, which is good, but
> again, doesn't work for all devices.
>
> And yeah, a whitelist is tough, especially as the vendor isn't working
> with us at all, but I don't want to break machines that don't need this
> driver (and there are some, although I thought the newest ones would not
> need it, but that might be true.)

Here is what the driver does with the patch:
- automatically load the driver and each machine with vendor =
"SAMSUNG ELECTRONICS CO., LTD." and type looks like a laptop or
netbook
- map the f0000 segment (or exit gracefully)
- find SwSMI or SECLINUX in this segment (or exit gracefully)

At this point, if the driver is still loaded, but the laptop doesn't
support SABI, then it really looks like a model that should be
blacklisted.
I'm ok to use dmi for quirks, or blacklist, but honestly if a *samsung
laptop* advertise SABI support, the driver should be loaded.
We can't prevent samsung to built a laptop that will burn if we use
SABI, but if samsung does that, they can also add some fake DMI
strings too ! And has I said in the changelog, all models supported by
samsung-q10 are not affected because they don't use the same vendor
string, and even if they were, I doubt we would found a valid
signature.

But yes, it needs testing, so if you know a samsung laptop user,
please ask him to test this patch.

I set up a git repo that can help to build the driver more easilly:
https://github.com/iksaif/samsung-laptop-dkms (yeah, it's named dkms,
but there is no dkms.conf yet, that will come later).

> Hm, if you can figure out how to get the acpi-video thing working for
> some machines properly, I'll support this kind of change.

For those, boot with acpi_backlight=vendor  :)

Thanks,

-- 
Corentin Chary
http://xf.iksaif.net
--
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/