If the TPM has already been sent a SaveState command before the driver
is loaded it may have problems sending that same command again later.

This issue is seen with the Chromebook Pixel due to a firmware bug in
the legacy mode boot path which is sending the SaveState command
before booting the kernel.  More information is available at
http://crbug.com/203524

This change introduces a retry of the SaveState command in the suspend
path in order to work around this issue.  A future firmware update
should fix this but this is also a trivial workaround in the driver
that has no effect on systems that do not show this problem.

When this does happen the TPM responds with a non-fatal TPM_RETRY code
that is defined in the specification:

  The TPM is too busy to respond to the command immediately, but the
  command could be resubmitted at a later time.  The TPM MAY return
  TPM_RETRY for any command at any time.

It can take several seconds before the TPM will respond again.  I
measured a typical time between 3 and 4 seconds and the timeout is set
at a safe 5 seconds.

It is also possible to reproduce this with commands via /dev/tpm0.
The bug linked above has a python script attached which can be used to
test for this problem.  I tested a variety of TPMs from Infineon,
Nuvoton, Atmel, and STMicro but was only able to reproduce this with
LPC and I2C TPMs from Infineon.

The TPM specification only loosely defines this behavior:

  TPM Main Level 2 Part 3 v1.2 r116, section 3.3. TPM_SaveState:
  The TPM MAY declare all preserved values invalid in response to any
  command other than TPM_Init.

  TCG PC Client BIOS Spec 1.21 section 8.3.1.
  After issuing a TPM_SaveState command, the OS SHOULD NOT issue TPM
  commands before transitioning to S3 without issuing another
  TPM_SaveState command.

  TCG PC Client TIS 1.21, section 4. Power Management:
  The TPM_SaveState command allows a Static OS to indicate to the TPM
  that the platform may enter a low power state where the TPM will be
  required to enter into the D3 power state.  The use of the term "may"
  is significant in that there is no requirement for the platform to
  actually enter the low power state after sending the TPM_SaveState
  command.  The software may, in fact, send subsequent commands after
  sending the TPM_SaveState command.

Change-Id: I52b41e826412688e5b6c8ddd3bb16409939704e9
Signed-off-by: NDuncan Laurie <dlaurie@chromium.org>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

Kent Yoder indicated that the code might be a bit clearer with a comment
here, so this patch adds a small explanation of the code.
Signed-off-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

As the subject says.
It's probably a good idea to have these fields populated.
Signed-off-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

module.h and sched.h were included twice.
Signed-off-by: NSachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

This driver adds support for Infineon's new SLB 9645 TT 1.2 I2C TPMs,
which supports clockstretching, combined reads and a bus speed of
up to 400khz. The device also has a new device id.

The driver works now also fine with device trees, so you can
instantiate your device by adding:
 +       tpm {
 +               compatible = "infineon,slb9645tt";
 +               reg = <0x20>;
 +       };
 for SLB 9645 devices or

 +       tpm {
 +               compatible = "infineon,slb9635tt";
 +               reg = <0x20>;
 +       };
 for SLB 9635 devices

to your device tree.
tpm_i2c_infineon is also retained as a compatible id as a fallback to
slb9635 protocol.

The driver was tested on Beaglebone.
Signed-off-by: NPeter Huewe <peter.huewe@infineon.com>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

Convert the struct i2c_msg initialization to C99 format. This makes
maintaining and editing the code simpler. Also helps once other fields
like transferred are added in future.

Thanks to Julia Lawall for automating the conversion.
Signed-off-by: NShubhrajyoti D <shubhrajyoti@ti.com>
Acked-by: NPeter Huewe <peter.huewe@infineon.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

 Ensure that the 'version' string includes a NULL terminator after its
copied out of the acpi table.
Signed-off-by: NChen Gang <gang.chen@asianux.com>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

Signed-off-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

virtio_rng feeds the randomness buffer handed by the core directly
into the scatterlist, since commit bb347d98.

However, if CONFIG_HW_RANDOM=m, the static buffer isn't a linear address
(at least on most archs).  We could fix this in virtio_rng, but it's actually
far easier to just do it in the core as virtio_rng would have to allocate
a buffer every time (it doesn't know how much the core will want to read).
Reported-by: NAurelien Jarno <aurelien@aurel32.net>
Tested-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Cc: stable@kernel.org

Commit 61337054 introduced a circular lock dependency because
posix_cpu_timers_exit() is called by release_task(), which is holding
a writer lock on tasklist_lock, and this can cause a deadlock since
kill_fasync() gets called with nonblocking_pool.lock taken.

There's no reason why kill_fasync() needs to be taken while the random
pool is locked, so move it out to fix this locking dependency.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Reported-by: NRuss Dill <Russ.Dill@gmail.com>
Cc: stable@kernel.org

If the minor number is assigned dynamically, there is no need to search
for misc->minor in misc_list, since misc->minor == MISC_DYNAMIC_MINOR.

[akpm@linux-foundation.org: reduce scope of local `c']
Signed-off-by: NDae S. Kim <dae@velatum.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add try...  parameters to disable pci and platform (openfirmware) device
scanning for IPMI.  Also add docs for all the try...  parameters.
Signed-off-by: NCorey Minyard <cminyard@mvista.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The configuration change building ipmi_si into the kernel precludes the
use of a custom driver that can utilize more than one KCS interface,
multiple IPMBs, and more than one BMC.  This capability is important for
fault-tolerant systems.

Even if the kernel option ipmi_si.trydefaults=0 is specified, ipmi_si
discovers and claims one of the KCS interfaces on a Stratus server.  The
inability to now prevent the kernel from managing this device is a
regression from previous kernels.  The regression breaks a capability
fault-tolerant vendors have relied upon.

To support both ACPI opregion access and the need to avoid activation of
ipmi_si on some platforms, we've added two new kernel options,
ipmi_si.tryacpi and ipmi_si.trydmi be added to prevent ipmi_si from
initializing when these options are set to 0 on the kernel command line.
With these options at the default value of 1, ipmi_si init proceeds
according to the kernel default.
Tested-by: NJim Paradis <jparadis@redhat.com>
Signed-off-by: NRobert Evans <Robert.Evans@stratus.com>
Signed-off-by: NJim Paradis <jparadis@redhat.com>
Signed-off-by: NTony Camuso <tcamuso@redhat.com>
Signed-off-by: NCorey Minyard <cminyard@mvista.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Whilst IOMMU is enabled for the Intel GPU on Ironlake, it appears that
using WC writes to update the PTE on the GPU fails miserably. The
result looks like the majority of the writes do not land leading to
lots of screen corruption and a hard system hang.

v2: s/</<=/ to preserve the current exclusion of Sandybridge
Reported-by: NNathan Myers <ncm@cantrip.org>
Bugzilla: https://bugzilla.freedesktop.org/show_bug.cgi?id=60391Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Tested-by: NNathan Myers <ncm@cantrip.org>
[danvet: Remove cc: stable and add tested-by.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The static lock initializers want to be fed the proper name of the
lock and not some random string. In mainline random strings are
obfuscating the readability of debug output, but for RT they prevent
the spinlock substitution. Fix it up.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

When I refactored the code initially, I forgot that gen2 uses a
different bar for the CPU mappable aperture. The agp-less code knows
nothing of generations less than 5, so we have to expand the gtt_probe
function to include the mappable base and end.

It was originally broken by me:
commit baa09f5f
Author: Ben Widawsky <ben@bwidawsk.net>
Date:   Thu Jan 24 13:49:57 2013 -0800

    drm/i915: Add probe and remove to the gtt ops
Reported-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

When rproc_serial is initialized, guest_connected should be set
to true. We can then revert the extra checks introduced in
commit: "virtio_console: Let unconnected rproc device receive data."
Signed-off-by: NSjur Brændeland <sjur.brandeland@stericsson.com>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

Use virtio device index for creating unique device port names.
Current index allocation in virtio is based on a monotonically
increasing variable "index". A better handling of this is to
use device index which is allocated by ida.
Signed-off-by: NSjur Brændeland <sjur.brandeland@stericsson.com>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

ERROR: open brace '{' following struct go on the same line
ERROR: space required after that ','
ERROR: space prohibited after that open parenthesis '('
WARNING: please, no spaces at the start of a line
WARNING: please, no space before tabs
Signed-off-by: NAlexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

mgslpc_probe() ignores errors in mgslpc_add_device() and
does not release all resource if mgslpc_config() failed.

The patch adds returned code to mgslpc_add_device()
and fixes the both issues.

Found by Linux Driver Verification project (linuxtesting.org).
Signed-off-by: NAlexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Don't access uninitialized work-queue when removing device.
The work queue is initialized only if the device multi-queue.
So don't call cancel_work unless this is a multi-queue device.

This fixes the following panic:

Kernel panic - not syncing: BUG!
Call Trace:
62031b28:  [<6026085d>] panic+0x16b/0x2d3
62031b30:  [<6004ef5e>] flush_work+0x0/0x1d7
62031b60:  [<602606f2>] panic+0x0/0x2d3
62031b68:  [<600333b0>] memcpy+0x0/0x140
62031b80:  [<6002d58a>] unblock_signals+0x0/0x84
62031ba0:  [<602609c5>] printk+0x0/0xa0
62031bd8:  [<60264e51>] __mutex_unlock_slowpath+0x13d/0x148
62031c10:  [<6004ef5e>] flush_work+0x0/0x1d7
62031c18:  [<60050234>] try_to_grab_pending+0x0/0x17e
62031c38:  [<6004e984>] get_work_gcwq+0x71/0x8f
62031c48:  [<60050539>] __cancel_work_timer+0x5b/0x115
62031c78:  [<628acc85>] unplug_port+0x0/0x191 [virtio_console]
62031c98:  [<6005061c>] cancel_work_sync+0x12/0x14
62031ca8:  [<628ace96>] virtcons_remove+0x80/0x15c [virtio_console]
62031ce8:  [<628191de>] virtio_dev_remove+0x1e/0x7e [virtio]
62031d08:  [<601cf242>] __device_release_driver+0x75/0xe4
62031d28:  [<601cf2dd>] device_release_driver+0x2c/0x40
62031d48:  [<601ce0dd>] driver_unbind+0x7d/0xc6
62031d88:  [<601cd5d9>] drv_attr_store+0x27/0x29
62031d98:  [<60115f61>] sysfs_write_file+0x100/0x14d
62031df8:  [<600b737d>] vfs_write+0xcb/0x184
62031e08:  [<600b58b8>] filp_close+0x88/0x94
62031e38:  [<600b7686>] sys_write+0x59/0x88
62031e88:  [<6001ced1>] handle_syscall+0x5d/0x80
62031ea8:  [<60030a74>] userspace+0x405/0x531
62031f08:  [<600d32cc>] sys_dup+0x0/0x5e
62031f28:  [<601b11d6>] strcpy+0x0/0x18
62031f38:  [<600be46c>] do_execve+0x10/0x12
62031f48:  [<600184c7>] run_init_process+0x43/0x45
62031fd8:  [<60019a91>] new_thread_handler+0xba/0xbc
Signed-off-by: NSjur Brændeland <sjur.brandeland@stericsson.com>
Cc: stable@kernel.org
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

This patch fixes four foo * bar errors, and one trailing whitespace
complaint from checkpatch.pl
Signed-off-by: NHans Grob <H.Grob@physik.uni-muenchen.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Reported-by: NPeter Hüwe <peterhuewe@gmx.de>
Reviewed-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

We don't need a temporary variable just to store the return value which
gets return in the next statement.
Signed-off-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

Err is never read before it is assigned again -> remove the dead
assigment.
Found with clang static analyzer
Signed-off-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

With the HOTPLUG changes 3.8 this attribute is going away.
Signed-off-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

We don't need to call memcpy for one byte, but assign it directly.
And to make the offset clearer we use the array syntax on the subsequent
call to memset to make the relationship clearer.
Signed-off-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

Reported-by: NPeter Hüwe <PeterHuewe@gmx.de>
Signed-off-by: NPeter Hüwe <PeterHuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

The TIS specification (pg 47) says the valid bit must be set, but
the TPM will not set it until it has completed its internal startup.

The driver checks that the valid bit is set during request_locality,
but it issues a TPM_ACCESS_REQUEST_USE without validating the
valid bit is set.

Some TPMs will ignore the TPM_ACCESS_REQUEST_USE, until valid is
set which causes the request_locality to timeout, which breaks the
driver attach.

Wait one timeout unit for valid to assert. If valid does not assert
then assume -ENODEV.

Seen on embedded with a:
1.2 TPM (device-id 0x3204, rev-id 64)
Signed-off-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

Support cancellation of TPM commands when driver is used in interrupt
mode.
Signed-off-by: NStefan Berger <stefanb@linux.vnet.ibm.com>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

On one of my machines the cancellation of TPM commands does not work.
The reason is that by writing into sysfs 'cancel' the tpm_tis_ready
call causes the status flag TPM_STS_VALID to be set in the statusregister.
However, the TIS driver seems to wait for TPM_STS_COMMAND_READY.
Once a 2nd time sysfs 'cancel' is written to, the TPM_STS_COMMAND_READY flag
also gets set, resulting in TPM_STS_VALID|TPM_STS_COMMAND_READY to be
read from the status register.

This patch now converts req_canceled into a function to enable more complex
comparisons against possible cancellation status codes.
Signed-off-by: NStefan Berger <stefanb@linux.vnet.ibm.com>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

Store the TPM vendor ID for later use.
Signed-off-by: NStefan Berger <stefanb@linux.vnet.ibm.com>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

We've been testing an alternative TPM for our embedded products and
found random kernel boot failures due to time outs after the continue
self test command.

This was happening randomly, and has been *very* hard to track down, but
it
looks like with this chip there is some kind of race with the
tpm_tis_status()
check of TPM_STS_COMMAND_READY. If things get there 'too fast' then
it sees the chip is ready, or tpm_tis_ready() works. Otherwise it takes
somewhere over 400ms before the chip will return TPM_STS_COMMAND_READY.

Adding some delay after tpm_continue_selftest() makes things reliably
hit the failure path, otherwise it is a crapshot.

The spec says it should be returning TPM_WARN_DOING_SELFTEST, not
holding
off on ready..

Boot log during this event looks like this:

tpm_tis 70030000.tpm_tis: 1.2 TPM (device-id 0x3204, rev-id 64)
tpm_tis 70030000.tpm_tis: Issuing TPM_STARTUP
tpm_tis 70030000.tpm_tis: tpm_transmit: tpm_send: error -62
tpm_tis 70030000.tpm_tis: [Hardware Error]: TPM command timed out during
continue self test
tpm_tis 70030000.tpm_tis: tpm_transmit: tpm_send: error -62
tpm_tis 70030000.tpm_tis: [Hardware Error]: TPM command timed out during
continue self test
tpm_tis 70030000.tpm_tis: tpm_transmit: tpm_send: error -62
tpm_tis 70030000.tpm_tis: [Hardware Error]: TPM command timed out during
continue self test
tpm_tis 70030000.tpm_tis: tpm_transmit: tpm_send: error -62
tpm_tis 70030000.tpm_tis: [Hardware Error]: TPM command timed out during
continue self test

The other TPM vendor we use doesn't show this wonky behaviour:
tpm_tis 70030000.tpm_tis: 1.2 TPM (device-id 0xFE, rev-id 70)
Signed-off-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

When no i2c bus exists, user-space can cause an oops by triggering a
device probe through a message sent to an i2c "new_device" sysfs entry.
Adding a check for a NULL i2c client structure in the probe function
closes the hole.

This patch also fixes accessing the NULL client struct in the print
function call reporting the error.
Reported-by: NPeter Hüwe <PeterHuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

This patch converts the suspend and resume functions for
tpm_i2c_stm_st33 to the new dev_pm_ops.
Signed-off-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

 * STMicroelectronics version 1.2.0, Copyright (C) 2010
 * STMicroelectronics comes with ABSOLUTELY NO WARRANTY.
 * This is free software, and you are welcome to redistribute it
 * under certain conditions.

This is the driver for TPM chip from ST Microelectronics.

If you have a TPM security chip from STMicroelectronics working with
an I2C, in menuconfig or .config choose the tpm driver on
device --> tpm and activate the protocol of your choice before compiling
the kernel.
The driver will be accessible from within Linux.

Tested on linux x86/x64, beagleboard REV B & XM REV C and CHROMIUM OS
Signed-off-by: NMathias Leblanc <mathias.leblanc@st.com>
Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>

Signed-off-by: NKent Yoder <key@linux.vnet.ibm.com>