linux-kernel - Re: [3.18.3 BISECTED REGRESSION] scx200_acb / cs5535-smb / geodewdt / cs5535-clockevt torpedoed

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, 3 Feb 2015 13:33:18 -0700
From:	Myron Stowe <myron.stowe@...il.com>
To:	Nix <nix@...eri.org.uk>
Cc:	Bjorn Helgaas <bhelgaas@...gle.com>,
	Myron Stowe <mstowe@...hat.com>,
	"linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Myron Stowe <myron.stowe@...hat.com>,
	Bill Unruh <unruh@...sics.ubc.ca>, martin@...ina.net,
	Matthew Wilcox <willy@...ux.intel.com>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Subject: Re: [3.18.3 BISECTED REGRESSION] scx200_acb / cs5535-smb / geodewdt /
 cs5535-clockevt torpedoed

On Tue, Feb 3, 2015 at 1:01 PM, Nix <nix@...eri.org.uk> wrote:
> On 3 Feb 2015, Myron Stowe uttered the following:
>> As expressed above, I believe this device is non-conformant.  This could be
>> validated by instrumenting the kernel's sizing code in '__pci_read_base()';
>> specifically the initial 'pci_read_config_dword()'s of 'l' and 'sz'.
>>
>> Whereas we have been able to ignore read-only BARs in past occurrances, this
>> time they are needed.  As such, I think we can solve this issue by expanding
>> the existing quirk for this device.  I'll work that up and post.  If you
>> would, please apply and test what is posted and report back.  In the mean
>> time, I would be interested in obtaining confirmation as to my belief that
>> this device's BARs are read-only (i.e. the instrumentation mentioned).  If
>> you have time and are willing I would appreaciate that.  If not, I'm
>> confident that is what is occurring and you can just stick with applying and
>> testing the expanded quirk patch I intend to post.
>
> Here, and the device behind it, just in case it's useful:
>
> [    1.050352] pci 0000:00:14.0: [1022:2090] type 00 class 0x060100
> [    1.060152] name: 0000:00:14.0; l: 6001; sz: 6001
> [    1.070037] pci 0000:00:14.0: reg 0x10: [io  0x6000-0x7fff]
> [    1.080114] name: 0000:00:14.0; l: 6101; sz: 6101
> [    1.090037] pci 0000:00:14.0: reg 0x14: [io  0x6100-0x61ff]
> [    1.100113] name: 0000:00:14.0; l: 6201; sz: 6201
> [    1.110036] pci 0000:00:14.0: reg 0x18: [io  0x6200-0x63ff]

Bingo - all three BARs above are not sizing, they are behaving as
read-only as suspected.

> [    1.120113] name: 0000:00:14.0; l: 0; sz: 0
> [    1.130130] name: 0000:00:14.0; l: 0; sz: 0
> [    1.140129] name: 0000:00:14.0; l: 0; sz: 0
> [    1.150129] name: 0000:00:14.0; l: 0; sz: 0
> [    1.160075] pci 0000:00:14.0: CS5536 ISA bridge bug detected (incorrect header); workaround applied
> [    1.180018] pci 0000:00:14.2: [1022:209a] type 00 class 0x010180
> [    1.190155] name: 0000:00:14.2; l: 0; sz: 0
> [    1.200133] name: 0000:00:14.2; l: 0; sz: 0
> [    1.210133] name: 0000:00:14.2; l: 0; sz: 0
> [    1.220133] name: 0000:00:14.2; l: 0; sz: 0
> [    1.230150] name: 0000:00:14.2; l: e001; sz: fffffff1

Above is from proper sizing behavior and thus the next line shows a proper,
size based region based upon such.

> [    1.240037] pci 0000:00:14.2: reg 0x20: [io  0xe000-0xe00f]
> [    1.250115] name: 0000:00:14.2; l: 0; sz: 0
> [    1.260133] name: 0000:00:14.2; l: 0; sz: 0
> [    1.270091] pci 0000:00:14.2: legacy IDE quirk: reg 0x10: [io  0x01f0-0x01f7]
> [    1.280021] pci 0000:00:14.2: legacy IDE quirk: reg 0x14: [io  0x03f6]
> [    1.290019] pci 0000:00:14.2: legacy IDE quirk: reg 0x18: [io  0x0170-0x0177]
> [    1.300018] pci 0000:00:14.2: legacy IDE quirk: reg 0x1c: [io  0x0376]

Don't be alarmed by the 'l' and 'sz' values of 0 for device 00:14.2.
Its a "legacy
IDE" device and thus is handled by special code (see:
drivers/pci/probe.c::pci_setup_device()).

The existing quirk I'm expanding - quirk_cs5536_vsa - will be similar to what
'pci_setup_device()' is doing for IDE.  I have it coded up, just need
to write up a
commit log which is often harder to come up with something good than writing the
patch.

Thanks again, stay tuned,
 Myron
>
> --
> NULL && (void)
--
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/