9 years agofirewire: add CSR PRIORITY_BUDGET support
Clemens Ladisch [Thu, 10 Jun 2010 06:35:06 +0000]
firewire: add CSR PRIORITY_BUDGET support

If supported by the OHCI controller, implement the PRIORITY_BUDGET
register, which is required for nodes that can use asynchronous
priority arbitration.

To allow the core to determine what features the lowlevel device
supports, add a new card driver callback.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: add CSR BUSY_TIMEOUT support
Clemens Ladisch [Thu, 10 Jun 2010 06:34:13 +0000]
firewire: add CSR BUSY_TIMEOUT support

Implement the BUSY_TIMEOUT register, which is required for nodes that
support retries.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: add CSR BUS_TIME support
Clemens Ladisch [Thu, 10 Jun 2010 06:33:07 +0000]
firewire: add CSR BUS_TIME support

Implement the BUS_TIME register, which is required for cycle master
capable nodes and tested for by the Base 1393 Test Suite.  Even when
there is not yet bus master initialization support, this register allows
us to work together with other bus masters.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: add CSR CYCLE_TIME write support
Clemens Ladisch [Thu, 10 Jun 2010 06:26:48 +0000]
firewire: add CSR CYCLE_TIME write support

The specification requires that CYCLE_TIME is writable so that it can be
initialized, so we better implement it.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: core: add CSR SPLIT_TIMEOUT support
Clemens Ladisch [Thu, 10 Jun 2010 06:26:28 +0000]
firewire: core: add CSR SPLIT_TIMEOUT support

Implement the SPLIT_TIMEOUT registers.  Besides being required by the
spec, this is desirable for some IIDC devices and necessary for many
audio devices to be able to increase the timeout from userspace.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: core: add CSR RESET_START support
Clemens Ladisch [Thu, 10 Jun 2010 06:25:46 +0000]
firewire: core: add CSR RESET_START support

This implements the RESET_START register (as a dummy) to make the Base
1394 Test Suite happy.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: add CSR NODE_IDS support
Clemens Ladisch [Thu, 10 Jun 2010 06:25:19 +0000]
firewire: add CSR NODE_IDS support

The NODE_IDS register, and especially its bus_id field, is quite
useless because 1394.1 requires that the bus_id field always stays
0x3ff.  However, the 1394 specification requires this register on all
transaction capable nodes, and the Base 1394 Test Suite tests for it,
so we better implement it.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: add read_csr_reg driver callback
Clemens Ladisch [Thu, 10 Jun 2010 06:24:35 +0000]
firewire: add read_csr_reg driver callback

To prepare for the following additions of more OHCI-implemented CSR
registers, replace the get_cycle_time driver callback with a generic
CSR register callback.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: core: add CSR STATE_CLEAR/STATE_SET support
Clemens Ladisch [Thu, 10 Jun 2010 06:24:03 +0000]
firewire: core: add CSR STATE_CLEAR/STATE_SET support

The state registers are zero and read-only in this implementation, so
they are not of much use.  However, the specification requires that they
are present for transaction capable nodes, and the Base 1394 Test Suite
tests for them, so we better implement them.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: core: retry on local errors in bus manager election
Clemens Ladisch [Thu, 10 Jun 2010 06:23:28 +0000]
firewire: core: retry on local errors in bus manager election

When the candidate bus manager fails to do the lock request with which
it tries to become bus manager, it assumes that the current IRM is not
actually IRM capable and forces itself to become root.  However, if that
lock request failed because the local node itself was not able to send
it, then we cannot blame the current IRM and should not steal its

In this case, RCODE_SEND_ERROR is likely to indicate a temporary error
condition such as exhausted tlabels or low memory, so we better try
again later.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: ohci: speed up PHY register accesses
Clemens Ladisch [Thu, 10 Jun 2010 06:22:07 +0000]
firewire: ohci: speed up PHY register accesses

Most PHY chips, when idle, can complete a register access in the time
needed for two or three PCI read transactions; bigger delays occur only
when data is currently being moved over the link/PHY interface.  So if
we busy-wait a few times when waiting for the register access to finish,
it is likely that we can finish without having to sleep.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

9 years agofirewire: core: trivial fix for warning strings
Stefan Richter [Sat, 5 Jun 2010 18:32:50 +0000]
firewire: core: trivial fix for warning strings

WARN's format string argument should not carry a printk level prefix.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: check cdev response length
Clemens Ladisch [Wed, 19 May 2010 06:28:32 +0000]
firewire: check cdev response length

Add a check that the data length in the SEND_RESPONSE ioctl is correct.
Incidentally, this also fixes the previously wrong response length of
software-handled lock requests.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: add MSI support
Clemens Ladisch [Sat, 5 Jun 2010 10:31:25 +0000]
firewire: ohci: add MSI support

This patch adds support for message-signaled interrupts.

Any native PCI-Express OHCI controller should support MSI, but most are
just PCI cores behind a PCI-E/PCI bridge.  The only chips that are known
to claim to support MSI are the Lucent/Agere/LSI FW643 and the VIA
VT6315, none of which I have been able to test.

Due to the high level of trust I have in the competence of these and any
future chip makers, I thought it a good idea to add a disable-MSI quirk.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>

Tested Agere FW643 rev 07 [11c1:5901] and JMicron JMB381 [197b:2380].
Added a quirks list entry for JMB38X since it kept its count of MSI
events consistently at zero.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: do not enable interrupts without the handler
Stefan Richter [Sat, 5 Jun 2010 09:46:49 +0000]
firewire: ohci: do not enable interrupts without the handler

On 26 Apr 2010, Clemens Ladisch wrote:
> In theory, none of the interrupts should occur before the link is
> enabled.  In practice, I'd rather make sure to not set the master
> interrupt enable bit until we have installed the interrupt handler.

and proposed to move OHCI1394_masterIntEnable out of the present
reg_write() into a new one before the HCControl.linkEnable reg_write().

Why not defer setting /all/ of the bits until right before linkEnable?

Reviewed-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agoieee1394: video1394: Use memdup_user
Julia Lawall [Sat, 22 May 2010 08:22:00 +0000]
ieee1394: video1394: Use memdup_user

...when user data is immediately copied into the allocated region.

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (changelog)

9 years agoieee1394: schedule for removal
Stefan Richter [Mon, 24 May 2010 19:31:23 +0000]
ieee1394: schedule for removal

All application domains that are supported by the old ieee1394 driver
stack are supported by the newer firewire driver stack too.  There is
now good and extensive experience with the newer stack from deployment
in Fedora since F7 as well as by enthusiast users of other

The new drivers have consequently been recommended as the default ones
since 2.6.33, in order to fix some severe usability problems of FireWire
on Linux due to limitations of the old stack.  It is now high time to
announce when the obsolete drivers will be removed.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Acked-by: Jarod Wilson <jarod@redhat.com>

9 years agofirewire: core: use separate timeout for each transaction
Clemens Ladisch [Tue, 27 Apr 2010 07:07:00 +0000]
firewire: core: use separate timeout for each transaction

Using a single timeout for all transaction that need to be flushed does
not work if the submission of new transactions can defer the timeout
indefinitely into the future.  We need to have timeouts that do not
change due to other transactions; the simplest way to do this is with a
separate timer for each transaction.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (+ one lockdep annotation)

9 years agofirewire: core: Fix tlabel exhaustion problem
Peter Hurley [Fri, 23 Apr 2010 23:27:25 +0000]
firewire: core: Fix tlabel exhaustion problem

fw_core_handle_response() was not properly clearing tlabel_mask. This
was resulting in premature tlabel exhaustion.

Signed-off-by: Peter Hurley <phurley@charter.net>

This fixes an omission in 2.6.31-rc1 commit 1e626fdc "firewire: core:
use more outbound tlabels" which prevented to really use 64 instead of
32 transaction labels, as soon as split transactions occurred that had
their AR-resp tasklet run after the AT-req tasklet.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: make transaction label allocation more robust
Clemens Ladisch [Mon, 19 Apr 2010 15:29:14 +0000]
firewire: core: make transaction label allocation more robust

If one request is so long-lived that it does not get a response before
the following 63 requests, its bit in tlabel_mask is still set when the
next request tries to allocate a transaction label for that number.  In
this state, while the first request is not completed or timed out, no
new requests can be submitted.

To fix this, skip over any label still in use, and do not error out
unless we have entirely run out of labels.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: clean up config ROM related defined constants
Stefan Richter [Wed, 14 Apr 2010 20:30:18 +0000]
firewire: core: clean up config ROM related defined constants

Clemens Ladisch pointed out that
  - BIB_IMC is not named like the field is called in the standard,
  - readers of the code may get worried about the magic 0x0c0083c0,
  - a CSR_NODE_CAPABILITIES key is there in the header but not put to
    good use.

So let's rename BIB_IMC, add a defined constant for Node_Capabilities
and a comment which reassures people that somebody thought about it and
they don't have to (or if they still do, tell them where they have to
look for confirmation), and prune our incomplete and arbitrary set of
defined constants of CSR key IDs.  And there is a nother magic number,
that of Bus_Information_Block.Bus_Name, to be defined and commented.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agoieee1394: mark char device files as not seekable
Stefan Richter [Sat, 10 Apr 2010 15:47:18 +0000]
ieee1394: mark char device files as not seekable

  - raw1394   (/dev/raw1394),
  - video1394 (/dev/video1394/*),
  - dv1394    (/dev/dv1394/*)
character device file ABIs do not make any use of lseek(), pread(), or
pwrite().  Therefore use nonseekable_open() and, redundantly, set
file_operations.llseek to no_llseek to remove any doubt whether the BKL-
grabbing default_llseek handler is used.

Although all this is legacy code which should be left in peace until it
is eventually removed (as it is superseded by firewire-core's
<linux/firewire-cdev.h> ABI), this change seems still worth doing to
further minimize the presence of BKL usage in the kernel.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: cdev: mark char device files as not seekable
Stefan Richter [Sat, 10 Apr 2010 15:38:05 +0000]
firewire: cdev: mark char device files as not seekable

The <linux/firewire-cdev.h> character device file ABI (i.e. /dev/fw*
character device file interface) does not make any use of lseek(),
pread(), pwrite() (or any kind of write() at all).

Use nonseekable_open() and, redundantly, set file_operations.llseek to
no_llseek to remove any doubt whether the BKL-grabbing default_llseek
handler is used.  (Also shuffle file_operations initialization according
to the order of handler definitions.)

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: cleanups and fix for nonstandard build without debug facility
Stefan Richter [Fri, 2 Apr 2010 12:05:02 +0000]
firewire: ohci: cleanups and fix for nonstandard build without debug facility

1) Clean up two function names:  The ohci_ prefix is only used in names
of fw_card_driver hooks.  There were two unnecessary exceptions.

2) Replace empty macros by empty inline functions so that call parameter
type checking is available in #ifndef'd builds.

3) CONFIG_FIREWIRE_OHCI_DEBUG is currently a hidden kconfig variable,
hence is not going to be switched off by anybody.  Still, it can be
switched off but then compilation will fail in ohci_enable() at the
expression param_debug & OHCI_PARAM_DEBUG_BUSRESETS.  Add the necessary
definitions in the nonstandard case.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: wait for PHY register accesses to complete
Stefan Richter [Sat, 10 Apr 2010 14:04:56 +0000]
firewire: ohci: wait for PHY register accesses to complete

Rather than having the arbitrary msleep(2) pause, let read_phy_reg()
loop until the link--phy access was finished.

Factor write_phy_reg() out of ohci_update_phy_reg() and of
read_paged_phy_reg() and let it loop too until the link--phy access was

Like in the older ohci1394 driver, a timeout of 100 milliseconds is
chosen.  Unlike the old driver, we sleep instead of busy-wait in each
waiting loop iteration.  Instead of a loop, the waiting could probably
also be implemented interrupt driven, but why bother.  It would require
up and running interrupt handling before the link was fully configured
and enabled.

Also modify functions a bit:  Error return and value return can be
combined in read_phy_reg() since the domain of values is only u8.
Likewise in read_paged_phy_reg().

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: fix up configuration of TI chips
Clemens Ladisch [Thu, 1 Apr 2010 14:43:59 +0000]
firewire: ohci: fix up configuration of TI chips

On TI chips (OHCI-Lynx and later), enable link enhancements features
that TI recommends to be used.  None of these are required for proper
operation, but they are safe and nice to have.

In theory, these bits should have been set by default, but in practice,
some BIOS/EEPROM writers apparently do not read the datasheet, or get
spooked by names like "unfair".

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: enable 1394a enhancements
Clemens Ladisch [Sun, 4 Apr 2010 13:19:54 +0000]
firewire: ohci: enable 1394a enhancements

The OHCI spec says that, if the programPhyEnable bit is set, the driver
is responsible for configuring the IEEE1394a enhancements within the PHY
and the link consistently.  So do this.

Also add a quirk to allow disabling these enhancements; this is needed
for the TSB12LV22 where ack accelerations are buggy (erratum b).

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: do not clear PHY interrupt status inadvertently
Clemens Ladisch [Thu, 1 Apr 2010 14:40:18 +0000]
firewire: ohci: do not clear PHY interrupt status inadvertently

The interrupt status bits in PHY register 5 are cleared by writing a one
bit.  To avoid clearing them unadvertently, do not write them back when
they were read as set, but only when they have been explicitly requested
to be set.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: add a function for reading PHY registers
Clemens Ladisch [Sun, 4 Apr 2010 13:19:52 +0000]
firewire: ohci: add a function for reading PHY registers

Move the register reading code from ohci_update_phy_reg() into
a function which can be used separately.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: cdev: comment fixlet
Stefan Richter [Sat, 10 Apr 2010 10:23:09 +0000]
firewire: cdev: comment fixlet

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: cdev: iso packet documentation
Clemens Ladisch [Wed, 31 Mar 2010 14:26:52 +0000]
firewire: cdev: iso packet documentation

Add the missing documentation for iso packets.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: cdev: fix information leak
Stefan Richter [Wed, 7 Apr 2010 06:30:50 +0000]
firewire: cdev: fix information leak

A userspace client got to see uninitialized stack-allocated memory if it
specified an _IOC_READ type of ioctl and an argument size larger than
expected by firewire-core's ioctl handlers (but not larger than the
core's union ioctl_arg).

Fix this by clearing the requested buffer size to zero, but only at _IOR
ioctls.  This way, there is almost no runtime penalty to legitimate
ioctls.  The only legitimate _IOR is FW_CDEV_IOC_GET_CYCLE_TIMER with 12
or 16 bytes to memset.

[Another way to fix this would be strict checking of argument size (and
possibly direction) vs. command number.  However, we then need a lookup
table, and we need to allow for slight size deviations in case of 32bit
userland on 64bit kernel.]

Reported-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: cdev: require quadlet-aligned headers for transmit packets
Clemens Ladisch [Wed, 31 Mar 2010 14:26:46 +0000]
firewire: cdev: require quadlet-aligned headers for transmit packets

The definition of struct fw_cdev_iso_packet seems to imply that the
header_length must be quadlet-aligned, and in fact, specifying an
unaligned header has never really worked when using multiple packet
structures, because the position of the next control word is computed by
rounding the header_length _down_, so the last one to three bytes of the
header would overlap the next control word.

To avoid this problem, check that the header length is properly aligned.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: cdev: disallow receive packets without header
Clemens Ladisch [Wed, 31 Mar 2010 14:26:39 +0000]
firewire: cdev: disallow receive packets without header

In receive contexts, reject packets with header_length==0.  This would
be an instruction to queue zero packets which would not make sense.

This prevents a division by zero in the OHCI driver.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: align driver match with modalias
Stefan Richter [Thu, 18 Mar 2010 23:39:07 +0000]
firewire: core: align driver match with modalias

The driver match strategy was:
  - Match vendor/model/specifier/version of the unit directory.
  - If that was a miss, match vendor from the root directory and
    model/specifier/version of the unit directory.

This was inconsistent with how the modalias string was constructed
until recently (take vendor/model from root directory and specifier/
version from unit directory).  It was also inconsistent with how it is
done since the parent commit:
  - Use vendor/model/specifier/version of the unit directory if possible,
  - fall back to one or more of vendor/model/specifier/version from the
    root directory depending on which ones are not present at the unit

Fix this inconsistency by sharing the ROM scanner function between
modalias printer function and driver match function.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: fix Model_ID in modalias
Stefan Richter [Thu, 18 Mar 2010 23:38:29 +0000]
firewire: core: fix Model_ID in modalias

The modalias string of devices that represent units on a FireWire node
did not show Module_ID entries within unit directories.  This was
because firewire-core searched only the root directory of the
configuration ROM for a Model_ID entry.

We now search first the root directory, then the unit directory.  IOW
honor a unit directory's Model_ID if present, otherwise fall back to the
root directory's model ID (if present).

Furthermore, apply the same change to Vendor_ID.  This had the same
issue but it was less apparent because most devices provide Vendor_ID
only in the root directory.

And finally, also use this strategy for the remaining two IDs in the
modalias, Specifier_ID and Version.  It does not actually make sense to
look for them elsewhere than in the unit directory because they are
mandatory there.  However, a uniform search order simplifies the
implementation and has no adverse affect in practice.

Side notes:
  - The older counterpart of this, nodemgr.c of ieee1394, looked for
    Vendor_ID first in the root directory, then in the unit directory,
    and for Model_ID only in the unit directory.
  - There is a single mainline driver which requires Vendor_ID and
    Model_ID --- the firedtv driver.  This one worked because FireDTVs
    provide Vendor_ID in the root directory and Model_ID identically in
    root directory and unit directory.
  - Apart from firedtv, there are currently no drivers known to me
    (including userspace drivers) that look at the Vendor_ID or Model_ID
    of the modalias.

Reported-by: Maciej ┼╗enczykowski <zenczykowski@gmail.com>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: add cycle timer quirk for the TI TSB12LV22
Clemens Ladisch [Wed, 17 Mar 2010 10:07:55 +0000]
firewire: ohci: add cycle timer quirk for the TI TSB12LV22

Among the many entries in the TSB12LV22 errata list (TI literature
number SLLS312) is the following:

  PCI Slave reads of the Cycle Timer register may occasionally get an
  incorrect value.
  Software may be able to validate value by reading the register
  multiple times rapidly and evaluating for a reasonable difference.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de> (untested)
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (added #define)

9 years agofirewire: core: fw_iso_resource_manage: fix error handling
Clemens Ladisch [Mon, 15 Mar 2010 12:20:32 +0000]
firewire: core: fw_iso_resource_manage: fix error handling

If the bandwidth allocation fails, the error must be returned in
*channel regardless of whether the channel allocation succeeded.
Checking for c >= 0 is not correct if no channel allocation was
requested, in which case this part of the code is reached with
c == -EINVAL.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: extend initialization log message
Stefan Richter [Sun, 21 Feb 2010 16:59:14 +0000]
firewire: ohci: extend initialization log message

by the number of available isochronous DMA contexts and active quirks
which is occasionally useful information.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: fix IR/IT context mask mixup
Stefan Richter [Sun, 21 Feb 2010 16:58:52 +0000]
firewire: ohci: fix IR/IT context mask mixup

This bug was present in firewire-ohci since day one:  The number of
available isochronous receive DMA contexts was mixed up with that of
available isochronous transmit DMA contexts.

This is harmless on a few chips which offer the same number of contexts
in both directions, but most chips nowadays implement only the standard
minimum of 4 IR contexts, but 8 IT contexts.  If a user attempted to run
a lot of IR contexts at once, results with more than four were therefore
unpredictable.  I suppose the controller would simply refuse to start
DMA of any unimplemented context.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: add module parameter to activate quirk fixes
Stefan Richter [Sun, 21 Feb 2010 16:58:29 +0000]
firewire: ohci: add module parameter to activate quirk fixes

This way, we can advise users of precompiled kernel packages to test
existing quirk fixes on chips which have not been listed yet, without
them having to build a kernel from source.

Note, to use this feature on a machine with more than one controller,
steps like these are necessary:
# lspci | grep 1394
# ls /sys/bus/pci/drivers/firewire_ohci/
# echo -n "0000:03:02.0" > /sys/bus/pci/drivers/firewire_ohci/unbind
# echo 2 > /sys/module/firewire_ohci/parameters/quirks
# echo -n "0000:03:02.0" > /sys/bus/pci/drivers/firewire_ohci/bind
# echo 0 > /sys/module/firewire_ohci/parameters/quirks

The parameter can also be used to switch off quirk flags that were
hardwired into firewire-ohci's quirks table.  Simply specify a non-zero
quirks value but without any known flags, e.g. 0x100.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: use an ID table for quirks detection
Stefan Richter [Sun, 21 Feb 2010 16:58:01 +0000]
firewire: ohci: use an ID table for quirks detection

We don't have a lot of quirks to take into account (especially since
dual-buffer IR is out of the picture), but still, a table-based approach
is more organized than a series of if () clauses.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: reorder struct fw_ohci for better cache efficiency
Stefan Richter [Sun, 21 Feb 2010 16:57:32 +0000]
firewire: ohci: reorder struct fw_ohci for better cache efficiency

The config_rom struct members are only accessed during relatively
infrequent self-ID-complete interrupts and only if the local config ROM
was changed, while the ar_, at_, ir_, it_ members are used very
frequently during I/O.  Hence move the config_rom members further down.

More importantly, make the huge self_id_buffer member the last one; this
is only accessed in self-ID-complete interrupts.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: ohci: remove unused dualbuffer IR code
Stefan Richter [Sun, 21 Feb 2010 16:57:05 +0000]
firewire: ohci: remove unused dualbuffer IR code

This code was no longer used since 2.6.33, "firewire: ohci: always use
packet-per-buffer mode for isochronous reception" commit 090699c0.  If
anybody needs this code in the future for special purposes, it can be
brought back in.  But it must not be re-enabled by default; drivers
(kernelspace or userspace drivers) should only get this mode if they
explicitly request it.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: combine a bit of repeated code
Stefan Richter [Sun, 21 Feb 2010 16:56:42 +0000]
firewire: core: combine a bit of repeated code

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: change type of a data buffer
Stefan Richter [Sun, 21 Feb 2010 16:56:21 +0000]
firewire: core: change type of a data buffer

from array of char to union of structs.  I already used a union to size
the buffer which holds ioctl arguments; more consequent is to define it
as an instance of this union in the first place.

Also rename several local variables from "request" to "a"(rgument) since
the term request can be mistaken to mean a transaction subaction, e.g.
an instance of struct fw_request.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: cdev: increment ABI version number
Stefan Richter [Sun, 21 Feb 2010 11:48:57 +0000]
firewire: cdev: increment ABI version number

so that clients can detect whether the FW_CDEV_IOC_GET_CYCLE_TIMER ioctl
is reliable (on all tested controllers, especially the widely used VIA
controllers, also NEC controllers, see commits b677532b and 1c1517ef).

Also add a comment on the 2.6.32 iso xmit enhancement and on dual-buffer
IR having been disabled in 2.6.33.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: cdev: add more flexible cycle timer ioctl
Stefan Richter [Sat, 20 Feb 2010 11:13:49 +0000]
firewire: cdev: add more flexible cycle timer ioctl

The system time from CLOCK_REALTIME is not monotonic, hence problematic
for the main user of the FW_CDEV_IOC_GET_CYCLE_TIMER ioctl.  This issue
exists in its successor ABI, i.e. raw1394, too.

We now offer an alternative ioctl which lets the caller choose between
the local time, very similar to the clock_gettime libc function.  The
format of the local time return value matches that of clock_gettime
(seconds and nanoseconds, instead of a single microseconds value from
the existing ioctl).

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: rename an internal function
Stefan Richter [Fri, 19 Feb 2010 20:00:31 +0000]
firewire: core: rename an internal function

according to what it really does.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: fix an information leak
Stefan Richter [Fri, 19 Feb 2010 20:00:02 +0000]
firewire: core: fix an information leak

If a device exposes a sparsely populated configuration ROM,
firewire-core's sysfs interface and character device file interface
showed random data in the gaps between config ROM blocks.  Fix this by
zero-initialization of the config ROM reader's scratch buffer.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: increase stack size of config ROM reader
Stefan Richter [Thu, 18 Feb 2010 00:54:00 +0000]
firewire: core: increase stack size of config ROM reader

The stack size of 16 was artificially chosen and may be too small in
extreme cases.  A device won't be accessible then.

Since it doesn't really matter to the slab allocator whether we ask for
1088 bytes or 2048 bytes of scratch memory, just allocate 2048 bytes for
the sum of temporary config ROM image and stack, and we will never ever
overflow the stack (because there simply can't be more stack items than
ROM entries).

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: don't fail device creation in case of too large config ROM blocks
Stefan Richter [Thu, 18 Feb 2010 00:52:45 +0000]
firewire: core: don't fail device creation in case of too large config ROM blocks

It never happened yet, but better safe than sorry:  If a device's config
ROM contains a block which overlaps the boundary at 0xfffff00007ff, just
ignore that one block instead of refusing to add the device
representation.  That way, upper layers (kernelspace or userspace
drivers) might still be able to use the device to some degree.

That's better than total inaccessibility of the device.  Worse, the core
would have logged only a generic "giving up on config rom" message which
could only be debugged by feeding a firewire-ohci debug logging session
through a config ROM interpreter, IOW would likely remain undiagnosed.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agofirewire: core: fix "giving up on config rom" with Panasonic AG-DV2500
Stefan Richter [Thu, 18 Feb 2010 00:50:31 +0000]
firewire: core: fix "giving up on config rom" with Panasonic AG-DV2500

The Panasonic AG-DV2500 tape deck contains an invalid entry in its
configuration ROM root directory:  A leaf pointer with the undefined key
ID 0 and an offset that points way out of the standard config ROM area.
This caused firewire-core to dismiss the device with the generic log
message "giving up on config rom for node id...", after which it was of
course impossible to access the tape deck with dvgrab or any other
program.  https://bugzilla.redhat.com/show_bug.cgi?id=449252#c29

The fix is to simply ignore this invalid ROM entry and proceed to read
the valid rest of the ROM.  There is a catch though:  When the kernel
later iterates over the ROM, it would be nasty having to check again for
such too large ROM offsets.  Therefore we manipulate the defective or
unsupported ROM entry to become a harmless immediate entry that won't
have any side effects later (an entry with the value 0x00000000).

Reported-by: George Chriss
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agoMerge tag 'v2.6.33' for its firewire changes since last branch point
Stefan Richter [Wed, 24 Feb 2010 19:31:04 +0000]
Merge tag 'v2.6.33' for its firewire changes since last branch point

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

9 years agoLinux 2.6.33
Linus Torvalds [Wed, 24 Feb 2010 18:52:17 +0000]
Linux 2.6.33

9 years agoMerge branch 'urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/kyle/parisc-2.6
Linus Torvalds [Wed, 24 Feb 2010 18:51:21 +0000]
Merge branch 'urgent' of git://git./linux/kernel/git/kyle/parisc-2.6

* 'urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/kyle/parisc-2.6:
  parisc: Set PCI CLS early in boot.

9 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Wed, 24 Feb 2010 18:51:04 +0000]
Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] Fix broken sn2 build

9 years agoparisc: Set PCI CLS early in boot.
Carlos O'Donell [Mon, 22 Feb 2010 23:25:59 +0000]
parisc: Set PCI CLS early in boot.

Set the PCI CLS early in the boot process to prevent
device failures. In pcibios_set_master use the new
pci_cache_line_size instead of a hard-coded value.

Signed-off-by: Carlos O'Donell <carlos@codesourcery.com>
Reviewed-by: Grant Grundler <grundler@google.com>
Signed-off-by: Kyle McMartin <kyle@redhat.com>

9 years agoMerge branch 'for-linus' of git://git.monstr.eu/linux-2.6-microblaze
Linus Torvalds [Wed, 24 Feb 2010 15:43:02 +0000]
Merge branch 'for-linus' of git://git.monstr.eu/linux-2.6-microblaze

* 'for-linus' of git://git.monstr.eu/linux-2.6-microblaze:
  microblaze: Fix out_le32() macro
  microblaze: Fix cache loop function for cache range

9 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block
Linus Torvalds [Wed, 24 Feb 2010 15:42:42 +0000]
Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  Revert "block: improve queue_should_plug() by looking at IO depths"

9 years agomicroblaze: Fix out_le32() macro
Steven J. Magnani [Mon, 22 Feb 2010 15:25:42 +0000]
microblaze: Fix out_le32() macro

Trailing semicolon causes compilation involving out_le32() to fail.

Signed-off-by: Steven J. Magnani <steve@digidescorp.com>
Signed-off-by: Michal Simek <monstr@monstr.eu>

9 years agomicroblaze: Fix cache loop function for cache range
Michal Simek [Mon, 15 Feb 2010 09:50:42 +0000]
microblaze: Fix cache loop function for cache range

I create wrong asm code but none test shows that this part of code is wrong.
I am not convinces that were good idea to create asm optimized macros
for caches. The reason is that there is not optimization with previous code
that's why make sense to add old code and do some benchmarking which
functions are faster.

Signed-off-by: Michal Simek <monstr@monstr.eu>

9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 24 Feb 2010 03:44:07 +0000]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
  net: bug fix for vlan + gro issue
  tc35815: Remove a wrong netif_wake_queue() call which triggers BUG_ON
  cdc_ether: new PID for Ericsson C3607w to the whitelist (resubmit)
  IPv6: better document max_addresses parameter
  MAINTAINERS: update mv643xx_eth maintenance status
  e1000: Fix DMA mapping error handling on RX
  iwlwifi: sanity check before counting number of tfds can be free
  iwlwifi: error checking for number of tfds in queue
  iwlwifi: set HT flags after channel in rxon

9 years agonet: bug fix for vlan + gro issue
Ajit Khaparde [Tue, 16 Feb 2010 20:25:43 +0000]
net: bug fix for vlan + gro issue

Traffic (tcp) doesnot start on a vlan interface when gro is enabled.
Even the tcp handshake was not taking place.
This is because, the eth_type_trans call before the netif_receive_skb
in napi_gro_finish() resets the skb->dev to napi->dev from the previously
set vlan netdev interface. This causes the ip_route_input to drop the
incoming packet considering it as a packet coming from a martian source.

I could repro this on (stable) and 2.6.33-rc7.
With this fix, the traffic starts and the test runs fine on both vlan
and non-vlan interfaces.

CC: Herbert Xu <herbert@gondor.apana.org.au>
CC: Patrick McHardy <kaber@trash.net>
Signed-off-by: Ajit Khaparde <ajitk@serverengines.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

9 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux...
Linus Torvalds [Wed, 24 Feb 2010 02:15:05 +0000]
Merge branch 'release' of git://git./linux/kernel/git/lenb/linux-acpi-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6:
  ACPI: Be in TS_POLLING state during mwait based C-state entry
  ACPI: Fix regression where _PPC is not read at boot even when ignore_ppc=0
  acer-wmi: Respect current backlight level when loading

9 years agoMerge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Wed, 24 Feb 2010 02:13:34 +0000]
Merge branch 'drm-linus' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/vmwgfx: Fix queries if no dma buffer thrashing is occuring.
  drm/nv50: fix vram ptes on IGPs to point at stolen system memory
  drm/nv50: fix instmem binding on IGPs to point at stolen system memory
  drm/nv50: improve vram page table construction
  drm/nv50: more efficient clearing of gpu page table entries
  drm/nv50: make nv50_mem_vm_{bind,unbind} operate only on vram
  drm/nouveau: Fix up pre-nv17 analog load detection.

9 years ago[IA64] Fix broken sn2 build
Hedi Berriche [Tue, 23 Feb 2010 23:58:49 +0000]
[IA64] Fix broken sn2 build

Revert the change made to arch/ia64/sn/kernel/setup.c by commit
204fba4aa303ea4a7bb726a539bf4a5b9e3203d0 as it breaks the build.

Fixing the build the b94b08081fcecf83fa690d6c5664f6316fe72208 way
breaks xpc because genksyms then fails to generate an CRC for
per_cpu____sn_cnodeid_to_nasid because of limitations in the
generic genksyms code.

Signed-off-by: Hedi Berriche <hedi@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

9 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
David S. Miller [Tue, 23 Feb 2010 09:27:05 +0000]
Merge branch 'master' of git://git./linux/kernel/git/linville/wireless-2.6

9 years agotc35815: Remove a wrong netif_wake_queue() call which triggers BUG_ON
Atsushi Nemoto [Fri, 19 Feb 2010 05:13:58 +0000]
tc35815: Remove a wrong netif_wake_queue() call which triggers BUG_ON

The netif_wake_queue() is called correctly (i.e. only on !txfull
condition) from txdone routine.  So Unconditional call to the
netif_wake_queue() here is wrong.  This might cause calling of
start_xmit routine on txfull state and trigger BUG_ON.

This bug does not happen when NAPI disabled.  After txdone there
must be at least one free tx slot.  But with NAPI, this is not
true anymore and the BUG_ON can hits on heavy load.

In this driver NAPI was enabled on 2.6.33-rc1 so this is
regression from 2.6.32 kernel.

Reported-by: Ralf Roesch <ralf.roesch@rw-gmbh.de>
Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>

9 years agocdc_ether: new PID for Ericsson C3607w to the whitelist (resubmit)
Torgny Johansson [Fri, 19 Feb 2010 01:59:15 +0000]
cdc_ether: new PID for Ericsson C3607w to the whitelist (resubmit)

This patch adds a new vid/pid to the cdc_ether whitelist.

Device added:
- Ericsson Mobile Broadband variant C3607w

Signed-off-by: Torgny Johansson <torgny.johansson@gmail.com>

To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Signed-off-by: David S. Miller <davem@davemloft.net>

9 years agoIPv6: better document max_addresses parameter
Brian Haley [Mon, 22 Feb 2010 12:27:21 +0000]
IPv6: better document max_addresses parameter

Andrew Morton wrote:
>> >From ip-sysctl.txt file in kernel documentation I can see following description
>> for max_addresses:
>> max_addresses - INTEGER
>>         Number of maximum addresses per interface.  0 disables limitation.
>>         It is recommended not set too large value (or 0) because it would
>>         be too easy way to crash kernel to allow to create too much of
>>         autoconfigured addresses.

>> If this parameter applies only for auto-configured IP addressed, please state
>> it more clearly in docs or rename the parameter to show that it refers to
>> auto-configuration.

It did mention autoconfigured in the text, but the below makes it more obvious.

More clearly document IPv6 max_addresses parameter.

Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

9 years agoMAINTAINERS: update mv643xx_eth maintenance status
Lennert Buytenhek [Mon, 22 Feb 2010 22:34:54 +0000]
MAINTAINERS: update mv643xx_eth maintenance status

I am no longer with Marvell.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

9 years agoe1000: Fix DMA mapping error handling on RX
Anton Blanchard [Fri, 19 Feb 2010 17:54:53 +0000]
e1000: Fix DMA mapping error handling on RX

Check for error return from pci_map_single/pci_map_page and clean up.

With this and the previous patch the driver was able to handle a significant
percentage of errors (I set the fault injection rate to 10% and could still
download large files at a reasonable speed).

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

9 years agoRevert "block: improve queue_should_plug() by looking at IO depths"
Jens Axboe [Tue, 23 Feb 2010 07:40:43 +0000]
Revert "block: improve queue_should_plug() by looking at IO depths"

This reverts commit fb1e75389bd06fd5987e9cda1b4e0305c782f854.

"Benjamin S." <sbenni@gmx.de> reports that the patch in question
causes a big drop in sequential throughput for him, dropping from
200MB/sec down to only 70MB/sec.

Needs to be investigated more fully, for now lets just revert the
offending commit.



Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

9 years agodrm/vmwgfx: Fix queries if no dma buffer thrashing is occuring.
Thomas Hellstrom [Sun, 21 Feb 2010 14:54:55 +0000]
drm/vmwgfx: Fix queries if no dma buffer thrashing is occuring.

Intercept query commands and apply relocations to their guest pointers.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

9 years agoMerge remote branch 'nouveau/for-airlied' of ../drm-nouveau-next into drm-linus
Dave Airlie [Tue, 23 Feb 2010 05:42:18 +0000]
Merge remote branch 'nouveau/for-airlied' of ../drm-nouveau-next into drm-linus

* 'nouveau/for-airlied' of ../drm-nouveau-next:
  drm/nv50: fix vram ptes on IGPs to point at stolen system memory
  drm/nv50: fix instmem binding on IGPs to point at stolen system memory
  drm/nv50: improve vram page table construction
  drm/nv50: more efficient clearing of gpu page table entries
  drm/nv50: make nv50_mem_vm_{bind,unbind} operate only on vram
  drm/nouveau: Fix up pre-nv17 analog load detection.

9 years agoMerge branch 'pcc' into release
Len Brown [Tue, 23 Feb 2010 05:39:00 +0000]
Merge branch 'pcc' into release

9 years agoMerge branches 'bugzilla-14207' and 'idle' into release
Len Brown [Tue, 23 Feb 2010 05:19:48 +0000]
Merge branches 'bugzilla-14207' and 'idle' into release

9 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Tue, 23 Feb 2010 03:51:39 +0000]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  MIPS: BCM47xx: Fix 128MB RAM support
  MIPS: Highmem: Fix build error

9 years agoMerge branch 'parisc/tracehook' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Tue, 23 Feb 2010 03:51:13 +0000]
Merge branch 'parisc/tracehook' of git://git./linux/kernel/git/frob/linux-2.6-roland

* 'parisc/tracehook' of git://git.kernel.org/pub/scm/linux/kernel/git/frob/linux-2.6-roland:
  Revert "parisc: HAVE_ARCH_TRACEHOOK"

9 years agofs/exec.c: fix initial stack reservation
Michael Neuling [Mon, 22 Feb 2010 20:44:24 +0000]
fs/exec.c: fix initial stack reservation

803bf5ec259941936262d10ecc84511b76a20921 ("fs/exec.c: restrict initial
stack space expansion to rlimit") attempts to limit the initial stack to
20*PAGE_SIZE.  Unfortunately, in attempting ensure the stack is not
reduced in size, we ended up not changing the stack at all.

This size reduction check is not necessary as the expand_stack call does
this already.

This caused a regression in UML resulting in most guest processes being

Signed-off-by: Michael Neuling <mikey@neuling.org>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: WANG Cong <xiyou.wangcong@gmail.com>
Cc: Anton Blanchard <anton@samba.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: James Morris <jmorris@namei.org>
Cc: Serge Hallyn <serue@us.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Jouni Malinen <j@w1.fi>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agoefifb: fix framebuffer handoff
Marcin Slusarz [Mon, 22 Feb 2010 20:44:22 +0000]
efifb: fix framebuffer handoff

Commit 4410f3910947dcea8672280b3adecd53cec4e85e ("fbdev: add support for
handoff from firmware to hw framebuffers") didn't add fb_destroy
operation to efifb.  Fix it and change aperture_size to match size
passed to request_mem_region.

Addresses http://bugzilla.kernel.org/show_bug.cgi?id=15151

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Reported-by: Alex Zhavnerchik <alex.vizor@gmail.com>
Tested-by: Alex Zhavnerchik <alex.vizor@gmail.com>
Acked-by: Peter Jones <pjones@redhat.com>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agogeode-mfgpt: restore previous behavior for selecting IRQ
Jens Rottmann [Mon, 22 Feb 2010 20:44:20 +0000]
geode-mfgpt: restore previous behavior for selecting IRQ

geode-mfgpt: restore previous behavior for selecting IRQ

The MFGPT IRQ used to be, in order of decreasing priority,
 * IRQ supplied by the user as a boot-time parameter,
 * IRQ previously set by the BIOS or another driver,
 * default IRQ given at compile time.

Return to this behavior, which got broken when splitting the
MFGPT/clocksource driver for 2.6.33-rc1.

Signed-off-by: Jens Rottmann <JRottmann@LiPPERTEmbedded.de>
Acked-by: Andres Salomon <dilinger@collabora.co.uk>
Cc: Jordan Crouse <jordan.crouse@amd.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: john stultz <johnstul@us.ibm.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agoidr: fix a critical misallocation bug, take#2
Tejun Heo [Mon, 22 Feb 2010 20:44:19 +0000]
idr: fix a critical misallocation bug, take#2

This is retry of reverted 859ddf09743a8cc680af33f7259ccd0fd36bfe9d
("idr: fix a critical misallocation bug") which contained two bugs.

* pa[idp->layers] should be cleared even if it's not used by
  sub_alloc() because it's used by mark idr_mark_full().

* The original condition check also assigned pa[l] to p which the new
  code didn't do thus leaving p pointing at the wrong layer.

Both problems have been fixed and the idr code has received good amount
testing using userland testing setup where simple bitmap allocator is
run parallel to verify the result of idr allocation.

The bug this patch fixes is caused by sub_alloc() optimization path
bypassing out-of-room condition check and restarting allocation loop
with starting value higher than maximum allowed value.  For detailed
description, please read commit message of 859ddf09.

Signed-off-by: Tejun Heo <tj@kernel.org>
Based-on-patch-from: Eric Paris <eparis@redhat.com>
Reported-by: Eric Paris <eparis@redhat.com>
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
Tested-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agokernel/sys.c: fix missing rcu protection for sys_getpriority()
Tetsuo Handa [Mon, 22 Feb 2010 20:44:16 +0000]
kernel/sys.c: fix missing rcu protection for sys_getpriority()

find_task_by_vpid() is not safe without rcu_read_lock().  2.6.33-rc7 got
RCU protection for sys_setpriority() but missed it for sys_getpriority().

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agomemcg: fix oom killing a child process in an other cgroup
KAMEZAWA Hiroyuki [Mon, 22 Feb 2010 20:44:14 +0000]
memcg: fix oom killing a child process in an other cgroup

Presently the oom-killer is memcg aware and it finds the worst process
from processes under memcg(s) in oom.  Then, it kills victim's child

It may kill a child in another cgroup and may not be any help for
recovery.  And it will break the assumption users have.

This patch fixes it.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Acked-by: David Rientjes <rientjes@google.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agodrm/nv50: fix vram ptes on IGPs to point at stolen system memory
Ben Skeggs [Fri, 19 Feb 2010 22:10:11 +0000]
drm/nv50: fix vram ptes on IGPs to point at stolen system memory

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

9 years agodrm/nv50: fix instmem binding on IGPs to point at stolen system memory
Ben Skeggs [Fri, 19 Feb 2010 22:06:36 +0000]
drm/nv50: fix instmem binding on IGPs to point at stolen system memory

This also modifies the unused PRAMIN PT entries to be all zeroes, can't
really recall why I used 9/0 initially, just that it didn't work for
some reason.  It was likely masking a bug elsewhere that's since been

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

9 years agodrm/nv50: improve vram page table construction
Ben Skeggs [Thu, 11 Feb 2010 01:31:44 +0000]
drm/nv50: improve vram page table construction

This commit changes nouveau to construct PTEs which look very much like
the ones the binary driver creates.

I presume that filling multiple PTEs identically with length flags and
the physical address of the start of a block of VRAM is a hint to the
memory controller that it need not perform additional page table lookups
for that range of addresses.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

9 years agodrm/nv50: more efficient clearing of gpu page table entries
Ben Skeggs [Thu, 11 Feb 2010 00:25:53 +0000]
drm/nv50: more efficient clearing of gpu page table entries

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

9 years agodrm/nv50: make nv50_mem_vm_{bind,unbind} operate only on vram
Ben Skeggs [Thu, 11 Feb 2010 00:23:30 +0000]
drm/nv50: make nv50_mem_vm_{bind,unbind} operate only on vram

GART is handled elsewhere, no reason to have the code for it here too.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

9 years agodrm/nouveau: Fix up pre-nv17 analog load detection.
Francisco Jerez [Wed, 10 Feb 2010 18:25:58 +0000]
drm/nouveau: Fix up pre-nv17 analog load detection.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

9 years agoMIPS: BCM47xx: Fix 128MB RAM support
Hauke Mehrtens [Sat, 20 Feb 2010 18:51:20 +0000]
MIPS: BCM47xx: Fix 128MB RAM support

Ignoring the last page when ddr size is 128M. Cached accesses to last page
is causing the processor to prefetch using address above 128M stepping out
of the DDR address space.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/981/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

9 years agoMIPS: Highmem: Fix build error
Yoichi Yuasa [Sat, 20 Feb 2010 12:23:22 +0000]
MIPS: Highmem: Fix build error

arch/mips/mm/highmem.c: In function 'kmap_init':
arch/mips/mm/highmem.c:130: error: 'init_mm' undeclared (first use in this function)
arch/mips/mm/highmem.c:130: error: (Each undeclared identifier is reported only once
arch/mips/mm/highmem.c:130: error: for each function it appears in.)

Signed-off-by: Yoichi Yuasa <yuasa@linux-mips.org>
Cc: linux-mips <linux-mips@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/980/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

9 years agoRevert "parisc: HAVE_ARCH_TRACEHOOK"
Roland McGrath [Mon, 22 Feb 2010 18:43:43 +0000]
Revert "parisc: HAVE_ARCH_TRACEHOOK"

This reverts commit 81bf550d9cdfe0325eb1504b06c9f6511b442c1a.

HAVE_ARCH_TRACEHOOK requires defining the user_regset interfaces,
including task_user_regset_view().  parisc doesn't do that yet,
so don't lie about it.

Signed-off-by: Roland McGrath <roland@redhat.com>

9 years agoACPI: Be in TS_POLLING state during mwait based C-state entry
Pallipadi, Venkatesh [Wed, 10 Feb 2010 18:35:31 +0000]
ACPI: Be in TS_POLLING state during mwait based C-state entry

ACPI deep C-state entry had a long standing bug/missing feature, wherein we were sending
resched IPIs when an idle CPU is in mwait based deep C-state. Only mwait based C1 was using
the write to the monitored address to wake up mwait'ing CPU.

This patch changes the code to retain TS_POLLING bit if we are entering an mwait based
deep C-state.

The patch has been verified to reduce the number of resched IPIs in general and also
improves the performance/power on workloads with low system utilization (i.e., when mwait based
deep C-states are being used).

Fixes "netperf ~50% regression with 2.6.33-rc1, bisect to 1b9508f"

Reported-by: Lin Ming <ming.m.lin@intel.com>
Tested-by: Alex Shi <alex.shi@intel.com>
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>

9 years agoMerge branch 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Mon, 22 Feb 2010 16:55:32 +0000]
Merge branch 'perf-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  perf probe: Init struct probe_point and set counter correctly
  hw-breakpoint: Keep track of dr7 local enable bits
  hw-breakpoints: Accept breakpoints on NULL address
  perf_events: Fix FORK events

9 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Mon, 22 Feb 2010 16:48:06 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: usbtouchscreen - extend coordinate range for Generaltouch devices
  Input: polldev can cause crash in case when polling disabled

9 years agoi915 / PM: Fix hibernate regression caused by suspend/resume splitting
Rafael J. Wysocki [Thu, 18 Feb 2010 22:06:27 +0000]
i915 / PM: Fix hibernate regression caused by suspend/resume splitting

Commit 84b79f8d2882b0a84330c04839ed4d3cefd2ff77 (drm/i915: Fix crash
while aborting hibernation) attempted to fix a regression introduced
by commit cbda12d77ea590082edb6d30bd342a67ebc459e0 (drm/i915:
implement new pm ops for i915), but it went too far trying to split
the freeze/suspend and resume/thaw parts of the code.  As a result,
it introduced another regression, which only is visible on some systems.

Fix the problem by merging i915_drm_suspend() with
i915_drm_freeze() and moving some code from i915_resume()
into i915_drm_thaw(), so that intel_opregion_free() and
intel_opregion_init() are also executed in the freeze and thaw code
paths, respectively.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Reported-and-tested-by: Pedro Ribeiro <pedrib@gmail.com>
Tested-by: Tino Keitel <tino.keitel@tikei.de>
Acked-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agoiwlwifi: sanity check before counting number of tfds can be free
Stanislaw Gruszka [Fri, 19 Feb 2010 23:47:33 +0000]
iwlwifi: sanity check before counting number of tfds can be free

Check the frame control for ieee80211_is_data_qos() is true before
counting the number of tfds can be free, the tfds_in_queue only
increment when ieee80211_is_data_qos() is true before transmit; so it
should only decrement if the type match.

Remove ieee80211_is_data_qos check for frame_ctrl in tx_resp to avoid
invalid information pass from uCode.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
CC: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>