GitHub/LineageOS/android_kernel_samsung_universal7580.git
16 years agomv643xx_eth: bump version to 1.4
Lennert Buytenhek [Sun, 24 Aug 2008 07:01:57 +0000 (09:01 +0200)]
mv643xx_eth: bump version to 1.4

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: convert to phylib
Lennert Buytenhek [Tue, 26 Aug 2008 11:34:19 +0000 (13:34 +0200)]
mv643xx_eth: convert to phylib

Switch mv643xx_eth from using drivers/net/mii.c to using phylib.

Since the mv643xx_eth hardware does all the link state handling and
PHY polling, the driver will use phylib in the "Doing it all yourself"
mode described in the phylib documentation.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Acked-by: Andy Fleming <afleming@freescale.com>
16 years agophylib: allow incremental scanning of an mii bus
Lennert Buytenhek [Tue, 26 Aug 2008 11:08:46 +0000 (13:08 +0200)]
phylib: allow incremental scanning of an mii bus

This patch splits the bus scanning code in mdiobus_register() off
into a separate function, and makes this function available for
calling from external code.  This allows incrementally scanning an
mii bus, e.g. as information about which addresses are 'safe' to
scan becomes available.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Acked-by: Andy Fleming <afleming@freescale.com>
16 years agomv643xx_eth: enforce frequent hardware statistics polling
Lennert Buytenhek [Fri, 19 Sep 2008 03:04:57 +0000 (05:04 +0200)]
mv643xx_eth: enforce frequent hardware statistics polling

If we don't poll the hardware statistics counters at least once every
~34 seconds, overflow might occur without us noticing.  So, set up a
timer to poll the statistics counters at least once every 30 seconds.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: deal with unexpected ethernet header sizes
Lennert Buytenhek [Fri, 19 Sep 2008 02:05:00 +0000 (04:05 +0200)]
mv643xx_eth: deal with unexpected ethernet header sizes

When the IP header doesn't start 14, 18, 22 or 26 bytes into the packet
(which are the only four cases that the hardware can deal with if asked
to do IP checksumming on transmit), invoke the software checksum helper
instead of letting the packet go out with a corrupt checksum inserted
into the packet in the wrong place.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: fix receive checksumming
Lennert Buytenhek [Fri, 19 Sep 2008 00:58:50 +0000 (02:58 +0200)]
mv643xx_eth: fix receive checksumming

We have to explicitly tell the hardware to include the pseudo-header
when doing receive checksumming, otherwise hardware checksumming will
fail for every received packet and we'll end up setting CHECKSUM_NONE
on every received packet.

While we're at it, when skb->ip_summed is set to CHECKSUM_UNNECESSARY
on received packets, skb->csum is supposed to be undefined, and thus
there is no need to set it.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: add support for chips without transmit bandwidth control
Lennert Buytenhek [Thu, 4 Sep 2008 10:22:34 +0000 (12:22 +0200)]
mv643xx_eth: add support for chips without transmit bandwidth control

Add support for mv643xx_eth versions that have no transmit bandwidth
control registers at all, such as the ethernet block found in the
Marvell 88F6183 ARM SoC.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: avoid reading ->byte_cnt twice during receive processing
Lennert Buytenhek [Sun, 14 Sep 2008 13:50:32 +0000 (15:50 +0200)]
mv643xx_eth: avoid reading ->byte_cnt twice during receive processing

Currently, the receive processing reads ->byte_cnt twice (once to
update interface statistics and once to properly size the data area
of the received skb), but since receive descriptors live in uncached
memory, caching this value in a local variable saves one uncached
access, and increases routing performance a tiny little bit more.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: shrink default receive and transmit queue sizes
Lennert Buytenhek [Sun, 14 Sep 2008 12:09:59 +0000 (14:09 +0200)]
mv643xx_eth: shrink default receive and transmit queue sizes

Since the size of the receive queue is directly related to the data
cache footprint of the driver (between refilling a receive ring entry
with a fresh skb and receiving a packet in that entry, queue_size - 1
other skbs will have been touched), shrink the default receive queue
size to a saner number of entries, as 400 is definite overkill for
almost all workloads.

While we are at it, trim the default transmit queue size a bit as well.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: replace array of skbs awaiting transmit completion with a queue
Lennert Buytenhek [Thu, 28 Aug 2008 03:53:18 +0000 (05:53 +0200)]
mv643xx_eth: replace array of skbs awaiting transmit completion with a queue

Get rid of the skb pointer array that we currently use for transmit
reclaim, and replace it with an skb queue, to which skbuffs are appended
when they are passed to the xmit function, and removed from the front
and freed when we do transmit queue reclaim and hit a descriptor with
the 'owned by device' bit clear and 'last descriptor' bit set.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: avoid dropping tx lock during transmit reclaim
Lennert Buytenhek [Sat, 13 Sep 2008 02:16:15 +0000 (04:16 +0200)]
mv643xx_eth: avoid dropping tx lock during transmit reclaim

By moving DMA unmapping during transmit reclaim back under the netif
tx lock, we avoid the situation where we read the DMA address and buffer
length from the descriptor under the lock and then not do anything with
that data after dropping the lock on platforms where the DMA unmapping
routines are all NOPs (which is the case on all ARM platforms that
mv643xx_eth is used on at least).

This saves two uncached reads, which makes a small but measurable
performance difference in routing benchmarks.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: switch to netif tx queue lock, get rid of private spinlock
Lennert Buytenhek [Thu, 28 Aug 2008 06:26:28 +0000 (08:26 +0200)]
mv643xx_eth: switch to netif tx queue lock, get rid of private spinlock

Since our ->hard_start_xmit() method is already called under spinlock
protection (the netif tx queue lock), we can simply make that lock
cover the private transmit state (descriptor ring indexes et al.) as
well, which avoids having to use a private lock to protect that state.

Since this was the last user of the driver-private spinlock, it can
be killed off.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: move all work to the napi poll handler
Lennert Buytenhek [Thu, 28 Aug 2008 06:09:10 +0000 (08:09 +0200)]
mv643xx_eth: move all work to the napi poll handler

Move link status handling, transmit reclaim and TX_END handling from
the interrupt handler to the napi poll handler.  This allows switching
->lock over to a non-IRQ-safe lock and removes all explicit interrupt
disabling from the driver.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: transmit multiqueue support
Lennert Buytenhek [Thu, 28 Aug 2008 04:26:23 +0000 (06:26 +0200)]
mv643xx_eth: transmit multiqueue support

As all the infrastructure for multiple transmit queues already exists
in the driver, this patch is entirely trivial.

The individual transmit queues are still serialised by the driver's
per-port private spinlock, but that will disappear (i.e. be replaced
by the per-subqueue ->_xmit_lock) in a subsequent patch.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: delete unused and uninteresting interrupt source mask bits
Lennert Buytenhek [Thu, 28 Aug 2008 04:10:04 +0000 (06:10 +0200)]
mv643xx_eth: delete unused and uninteresting interrupt source mask bits

Delete a couple of unused and uninteresting interrupt source mask bits:
- The receive resource underrun interrupt sources are uninteresting
  because if we are in out-of-memory mode, we are already dealing with
  the issue, and we don't need the hardware to remind us again that we
  are out of memory.
- The LINK and PHY interrupt sources can be coalesced into one define,
  since we always use them together.
- The transmit resource underrun interrupt source can be disabled since
  we never activate the head descriptor of a paged skb until the
  fragments are all activated, so transmit underrun during a packet
  should never happen.
- The INT_EXT_TX_0 define is never used.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: get rid of netif_{stop,wake}_queue() calls on link down/up
Lennert Buytenhek [Thu, 28 Aug 2008 03:39:13 +0000 (05:39 +0200)]
mv643xx_eth: get rid of netif_{stop,wake}_queue() calls on link down/up

There is no need to call netif_{stop,wake}_queue() when the link goes
down/up, as the networking already takes care of this internally.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: remove force_phy_addr field
Lennert Buytenhek [Tue, 26 Aug 2008 12:06:47 +0000 (14:06 +0200)]
mv643xx_eth: remove force_phy_addr field

Currently, there are two different fields in the
mv643xx_eth_platform_data struct that together describe the PHY
address -- one field (phy_addr) has the address of the PHY, but if
that address is zero, a second field (force_phy_addr) needs to be
set to distinguish the actual address zero from a zero due to not
having filled in the PHY address explicitly (which should mean
'use the default PHY address').

If we are a bit smarter about the encoding of the phy_addr field,
we can avoid the need for a second field -- this patch does that.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: smi sharing is a per-unit property, not a per-port one
Lennert Buytenhek [Tue, 26 Aug 2008 10:56:56 +0000 (12:56 +0200)]
mv643xx_eth: smi sharing is a per-unit property, not a per-port one

Which top-level unit's SMI interface to use should be a property of
the top-level unit, not of the individual ports.  This patch moves the
->shared_smi pointer from the per-port platform data to the global
platform data.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: require contiguous receive and transmit queue numbering
Lennert Buytenhek [Tue, 26 Aug 2008 08:23:22 +0000 (10:23 +0200)]
mv643xx_eth: require contiguous receive and transmit queue numbering

Simplify receive and transmit queue handling by requiring the set
of queue numbers to be contiguous starting from zero.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: get rid of compile-time configurable transmit checksumming
Lennert Buytenhek [Sun, 24 Aug 2008 03:33:55 +0000 (05:33 +0200)]
mv643xx_eth: get rid of compile-time configurable transmit checksumming

Get rid of the mv643xx_eth-internal MV643XX_ETH_CHECKSUM_OFFLOAD_TX
compile-time option.  Using transmit checksumming is the sane default,
and anyone wanting to disable it should use ethtool(8) instead of
recompiling their kernels.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: get rid of receive-side locking
Lennert Buytenhek [Sun, 24 Aug 2008 02:33:36 +0000 (04:33 +0200)]
mv643xx_eth: get rid of receive-side locking

By having the receive out-of-memory handling timer schedule the napi
poll handler and then doing oom processing from the napi poll handler,
all code that touches receive state moves to napi context, letting us
get rid of all explicit locking in the receive paths since the only
mutual exclusion we need anymore at that point is protection against
reentering ourselves, which is provided by napi synchronisation.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: make napi unconditional
Lennert Buytenhek [Sat, 23 Aug 2008 23:03:57 +0000 (01:03 +0200)]
mv643xx_eth: make napi unconditional

Make napi unconditional on the receive side, so that we can get rid
of all the locking and local interrupt disabling in the receive path.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: use the SMI done interrupt to wait for SMI access completion
Lennert Buytenhek [Tue, 26 Aug 2008 02:42:59 +0000 (04:42 +0200)]
mv643xx_eth: use the SMI done interrupt to wait for SMI access completion

If the platform code has passed us the IRQ number of the mv643xx_eth
top-level error interrupt, use the error interrupt to wait for SMI
access completion instead of polling the SMI busy bit, since SMI bus
accesses can take up to tens of milliseconds.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: switch ->phy_lock from a spinlock to a mutex
Lennert Buytenhek [Sun, 24 Aug 2008 03:41:09 +0000 (05:41 +0200)]
mv643xx_eth: switch ->phy_lock from a spinlock to a mutex

Since commit 81600eea98789da09a32de69ca9d3be8b9503c54 ("mv643xx_eth:
use auto phy polling for configuring (R)(G)MII interface"),
mv643xx_eth no longer does SMI accesses from interrupt context.  The
only other callers that do SMI accesses all do them from process
context, which means we can switch the PHY lock from a spinlock to a
mutex, and get rid of the extra locking in some ethtool methods.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: get rid of modulo operations
Lennert Buytenhek [Sat, 23 Aug 2008 21:45:28 +0000 (23:45 +0200)]
mv643xx_eth: get rid of modulo operations

Get rid of the modulo operations that are currently used for
computing successive TX/RX descriptor ring indexes.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: get rid of IRQF_SAMPLE_RANDOM
Lennert Buytenhek [Sat, 23 Aug 2008 21:43:38 +0000 (23:43 +0200)]
mv643xx_eth: get rid of IRQF_SAMPLE_RANDOM

Using IRQF_SAMPLE_RANDOM for the mv643xx_eth interrupt handler
significantly increases interrupt processing overhead, so get rid
of it.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: fix receive buffer DMA unmapping
Lennert Buytenhek [Sun, 24 Aug 2008 05:19:48 +0000 (07:19 +0200)]
mv643xx_eth: fix receive buffer DMA unmapping

When tearing down a DMA mapping for a receive buffer, we should pass
dma_unmap_single() the exact same address that dma_map_single() gave
us when we originally set up the mapping.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agomv643xx_eth: fix 'netdev_priv(dev) == dev->priv' assumption
Lennert Buytenhek [Sun, 24 Aug 2008 03:59:16 +0000 (05:59 +0200)]
mv643xx_eth: fix 'netdev_priv(dev) == dev->priv' assumption

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
16 years agoLinux 2.6.27-rc5
Linus Torvalds [Thu, 28 Aug 2008 22:52:02 +0000 (15:52 -0700)]
Linux 2.6.27-rc5

16 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Thu, 28 Aug 2008 19:34:27 +0000 (12:34 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm:
  [ARM] 5226/1: remove unmatched comment end.
  [ARM] Skip memory holes in FLATMEM when reading /proc/pagetypeinfo
  [ARM] use bcd2bin/bin2bcd
  [ARM] use the new byteorder headers
  [ARM] OMAP: Fix 2430 SMC91x ethernet IRQ
  [ARM] OMAP: Add and update OMAP default configuration files
  [ARM] OMAP: Change mailing list for OMAP in MAINTAINERS
  [ARM] S3C2443: Fix the S3C2443 clock register definitions
  [ARM] JIVE: Fix the spi bus numbering
  [ARM] S3C24XX: pwm.c: stop debugging output
  [ARM] S3C24XX: Fix sparse warnings in pwm.c
  [ARM] S3C24XX: Fix spare errors in pwm-clock driver
  [ARM] S3C24XX: Fix sparse warnings in arch/arm/plat-s3c24xx/gpiolib.c
  [ARM] S3C24XX: Fix nor-simtec driver sparse errors
  [ARM] 5225/1: zaurus: Register I2C controller for audio codecs
  [ARM] orion5x: update defconfig to v2.6.27-rc4
  [ARM] Orion: register UART1 on QNAP TS-209 and TS-409
  [ARM] Orion: activate lm75 driver on DNS-323
  [ARM] Orion: fix MAC detection on QNAP TS-209 and TS-409
  [ARM] Orion: Fix boot crash on Kurobox Pro

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cooloney...
Linus Torvalds [Thu, 28 Aug 2008 19:34:01 +0000 (12:34 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/cooloney/blackfin-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cooloney/blackfin-2.6:
  Blackfin arch: Fix PM building on BF52x: No ROTWE on BF52x, add USBWE
  Blackfin arch: sram: use 'unsigned long' for irqflags
  Blackfin arch: let PCI depend on BROKEN
  Blackfin arch: move include/asm-blackfin header files to arch/blackfin
  Blackfin arch: fix bug - MPU crashes under stress
  Blackfin arch: Fix bug - when to rmmod the L1_module, it stucks and then reboot the board.
  Blackfin arch: dont actually need to muck with EMAC_SYSTAT for BF52x for demuxing
  Blackfin arch: Add MTD Partitions for MTD_DATAFLASH, increase max SPI SCLK

16 years agoMerge branch 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 28 Aug 2008 19:31:49 +0000 (12:31 -0700)]
Merge branch 'core-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  exit signals: use of uninitialized field notify_count
  lockdep: fix invalid list_del_rcu in zap_class
  lockstat: repair erronous contention statistics
  lockstat: fix numerical output rounding error

16 years agoMerge branch 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 28 Aug 2008 19:31:12 +0000 (12:31 -0700)]
Merge branch 'sched-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  sched: rt-bandwidth accounting fix
  sched: fix sched_rt_rq_enqueue() resched idle

16 years agoMerge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 28 Aug 2008 19:30:59 +0000 (12:30 -0700)]
Merge branch 'x86-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  x86: update defconfigs
  x86: msr: fix bogus return values from rdmsr_safe/wrmsr_safe
  x86: cpuid: correct return value on partial operations
  x86: msr: correct return value on partial operations
  x86: cpuid: propagate error from smp_call_function_single()
  x86: msr: propagate errors from smp_call_function_single()
  smp: have smp_call_function_single() detect invalid CPUs

16 years agoMerge branch 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6
Linus Torvalds [Thu, 28 Aug 2008 19:28:50 +0000 (12:28 -0700)]
Merge branch 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6

* 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6:
  i2c: Prevent log spam on some DVB adapters
  i2c: Add missing kerneldoc descriptions
  i2c: Fix device_init_wakeup place

16 years agoftrace: disable tracing for hibernation
Rafael J. Wysocki [Thu, 28 Aug 2008 12:39:12 +0000 (14:39 +0200)]
ftrace: disable tracing for hibernation

In accordance with commit f42ac38c59e0a03d6da0c24a63fb211393f484b0
("ftrace: disable tracing for suspend to ram"), disable tracing
around the suspend code in hibernation code paths.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years ago[ARM] 5226/1: remove unmatched comment end.
Jean-Christophe DUBOIS [Wed, 27 Aug 2008 23:00:23 +0000 (00:00 +0100)]
[ARM] 5226/1: remove unmatched comment end.

remove unmatched comment end.

Signed-off-by: Jean-Christophe DUBOIS <jcd@tribudubois.net>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years agosched: rt-bandwidth accounting fix
Peter Zijlstra [Tue, 19 Aug 2008 10:33:03 +0000 (12:33 +0200)]
sched: rt-bandwidth accounting fix

It fixes an accounting bug where we would continue accumulating runtime
even though the bandwidth control is disabled. This would lead to very long
throttle periods once bandwidth control gets turned on again.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agoBlackfin arch: Fix PM building on BF52x: No ROTWE on BF52x, add USBWE
Michael Hennerich [Thu, 28 Aug 2008 09:32:01 +0000 (17:32 +0800)]
Blackfin arch: Fix PM building on BF52x: No ROTWE on BF52x, add USBWE

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
16 years agoBlackfin arch: sram: use 'unsigned long' for irqflags
Vegard Nossum [Thu, 28 Aug 2008 09:28:46 +0000 (17:28 +0800)]
Blackfin arch: sram: use 'unsigned long' for irqflags

Using just 'unsigned' will make flags an unsigned int. While this is
arguably not an error on blackfin where sizeof(int) == sizeof(long),
the patch is still justified on the grounds of principle.

The patch was generated using the Coccinelle semantic patch framework.

Cc: Julia Lawall <julia@diku.dk>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Vegard Nossum <vegard.nossum@gmail.com>
Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
16 years agosched: fix sched_rt_rq_enqueue() resched idle
John Blackwood [Tue, 26 Aug 2008 19:09:43 +0000 (15:09 -0400)]
sched: fix sched_rt_rq_enqueue() resched idle

When sysctl_sched_rt_runtime is set to something other than -1 and the
CONFIG_RT_GROUP_SCHED kernel parameter is NOT enabled, we get into a state
where we see one or more CPUs idling forvever even though there are
real-time
tasks in their rt runqueue that are able to run (no longer throttled).

The sequence is:

- A real-time task is running when the timer sets the rt runqueue
    to throttled, and the rt task is resched_task()ed and switched
    out, and idle is switched in since there are no non-rt tasks to
    run on that cpu.

- Eventually the do_sched_rt_period_timer() runs and un-throttles
    the rt runqueue, but we just exit the timer interrupt and go back
    to executing the idle task in the idle loop forever.

If we change the sched_rt_rq_enqueue() routine to use some of the code
from the CONFIG_RT_GROUP_SCHED enabled version of this same routine and
resched_task() the currently executing task (idle in our case) if it is
a lower priority task than the higher rt task in the now un-throttled
runqueue, the problem is no longer observed.

Signed-off-by: John Blackwood <john.blackwood@ccur.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agoi2c: Prevent log spam on some DVB adapters
Jean Delvare [Thu, 28 Aug 2008 06:33:23 +0000 (08:33 +0200)]
i2c: Prevent log spam on some DVB adapters

Some DVB adapters do not support the special I2C transaction that we
use for probing purposes. There's no point in logging this event, as
there's nothing the user can do and in general there is no actual
problem. So, degrade one of these messages to a debug message, and
move the other one around so that it is only printed on bogus drivers.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Uwe Bugla <uwe.bugla@gmx.de>
16 years agoi2c: Add missing kerneldoc descriptions
Jean Delvare [Thu, 28 Aug 2008 06:33:23 +0000 (08:33 +0200)]
i2c: Add missing kerneldoc descriptions

Add missing kernel descriptions of struct i2c_driver members.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Randy Dunlap <rdunlap@xenotime.net>
Cc: David Brownell <david-b@pacbell.net>
16 years agoi2c: Fix device_init_wakeup place
Marc Pignat [Thu, 28 Aug 2008 06:33:22 +0000 (08:33 +0200)]
i2c: Fix device_init_wakeup place

device_init_wakeup must be called after device_register.

Signed-off-by: Marc Pignat <marc.pignat@hevs.ch>
Acked-by: David Brownell <david-b@pacbell.net>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Thu, 28 Aug 2008 00:38:07 +0000 (17:38 -0700)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (55 commits)
  sctp: fix random memory dereference with SCTP_HMAC_IDENT option.
  sctp: correct bounds check in sctp_setsockopt_auth_key
  wan: Missing capability checks in sbni_ioctl()
  e100, fix iomap read
  qeth: preallocated header account offset
  qeth: l2 write unicast list to hardware
  qeth: use -EOPNOTSUPP instead of -ENOTSUPP.
  ibm_newemac: Don't call dev_mc_add() before device is registered
  net: don't grab a mutex within a timer context in gianfar
  forcedeth: fix checksum flag
  net/usb/mcs7830: add set_mac_address
  net/usb/mcs7830: new device IDs
  [netdrvr] smc91x: fix resource removal (null ptr deref)
  ibmveth: fix bad UDP checksums
  [netdrvr] hso: dev_kfree_skb crash fix
  [netdrvr] hso: icon 322 detection fix
  atl1: disable TSO by default
  atl1e: multistatement if missing braces
  igb: remove 82576 quad adapter
  drivers/net/skfp/ess.c: fix compile warnings
  ...

16 years agosctp: fix random memory dereference with SCTP_HMAC_IDENT option.
Vlad Yasevich [Wed, 27 Aug 2008 23:09:49 +0000 (16:09 -0700)]
sctp: fix random memory dereference with SCTP_HMAC_IDENT option.

The number of identifiers needs to be checked against the option
length.  Also, the identifier index provided needs to be verified
to make sure that it doesn't exceed the bounds of the array.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agosctp: correct bounds check in sctp_setsockopt_auth_key
Vlad Yasevich [Wed, 27 Aug 2008 23:08:54 +0000 (16:08 -0700)]
sctp: correct bounds check in sctp_setsockopt_auth_key

The bonds check to prevent buffer overlflow was not exactly
right.  It still allowed overflow of up to 8 bytes which is
sizeof(struct sctp_authkey).

Since optlen is already checked against the size of that struct,
we are guaranteed not to cause interger overflow either.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoMerge branch 'omap-rmk'
Russell King [Wed, 27 Aug 2008 22:07:46 +0000 (23:07 +0100)]
Merge branch 'omap-rmk'

16 years agoIB/mlx4: Actually return L_Key and R_Key for fast register MRs
Vladimir Sokolovsky [Wed, 27 Aug 2008 21:29:57 +0000 (14:29 -0700)]
IB/mlx4: Actually return L_Key and R_Key for fast register MRs

Initialize the L_Key and R_Key for memory regions returned from
mlx4_ib_alloc_fast_reg_mr().  Otherwise callers just get garbage for
the memory keys and can't do anything useful with these MRs.

Signed-off-by: Vladimir Sokolovsky <vlad@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog
Linus Torvalds [Wed, 27 Aug 2008 21:37:28 +0000 (14:37 -0700)]
Merge git://git./linux/kernel/git/wim/linux-2.6-watchdog

* git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog:
  [WATCHDOG] removed unused #include <version.h>
  [WATCHDOG] at91rm9200_wdt.c: fix misleading indentation
  [WATCHDOG] mpc8xxx_wdt: fix modular build
  [WATCHDOG] hpwdt.c kdebug support
  [WATCHDOG] Add support for the IDT RC32434 watchdog
  [WATCHDOG] Add support for the built-int RDC R-321x SoC watchdog
  [WATHDOG] delete unused driver mpc8xx_wdt.c
  [WATCHDOG] Fix s3c2410_wdt driver coding style issues
  [WATCHDOG] Clean out header of s3c2410_wdt driver.
  [WATCHDOG] Fix NULL usage in s3c2410_wdt driver.

16 years agotty: Fix termios tty window resize race with pty/tty pair
Alan Cox [Tue, 26 Aug 2008 18:55:58 +0000 (19:55 +0100)]
tty: Fix termios tty window resize race with pty/tty pair

Kanru Chen posted a patch versus the old code which deals with the case
where you resize the pty side of a pty/tty pair. In that situation the
termios data is updated for both pty and tty but the locks are not held
for the right side.

This implements the fix differently against the updated tty code. Patch
by self but the hard bit (noticing and fixing the bug) is thanks to Kanru
Chen.

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agotty: TIOCGSOFTCAR/SSOFTCAR on pty is wron
Alan Cox [Tue, 26 Aug 2008 18:52:47 +0000 (19:52 +0100)]
tty: TIOCGSOFTCAR/SSOFTCAR on pty is wron

The termios settings ioctls on a pty should affect the bound tty side not
the pty. The SOFTCAR ioctls use the wrong device file.

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6
Linus Torvalds [Wed, 27 Aug 2008 21:34:49 +0000 (14:34 -0700)]
Merge git://git./linux/kernel/git/sfrench/cifs-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
  [CIFS] Add destroy routine for dns_resolver
  [CIFS] Reorder cifs config item for better clarity
  [CIFS] Correct keys dependency for cifs kerberos support

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg...
Linus Torvalds [Wed, 27 Aug 2008 21:33:06 +0000 (14:33 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/penberg/slab-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/slab-2.6:
  slub: Disable NUMA remote node defragmentation by default

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6
Linus Torvalds [Wed, 27 Aug 2008 21:31:44 +0000 (14:31 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
  [PATCH] deal with the first call of ->show() generating no output
  [PATCH] fix ->llseek() for a bunch of directories
  [PATCH] fix regular readdir() and friends
  [PATCH] fix hpux_getdents()
  [PATCH] fix osf_getdirents()
  [PATCH] ntfs: use d_add_ci
  [PATCH] change d_add_ci argument ordering
  [PATCH] fix efs_lookup()
  [PATCH] proc: inode number fixlet

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Wed, 27 Aug 2008 21:31:27 +0000 (14:31 -0700)]
Merge git://git./linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  sparc: Add target for a stripped kernel
  sparc64: Make NUMA depend upon SMP.

16 years agoMerge branch 'drm-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Wed, 27 Aug 2008 21:28:45 +0000 (14:28 -0700)]
Merge branch 'drm-patches' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm: don't set the signal blocker on the master process.
  drm: don't call the vblank tasklet with irqs disabled.
  r300: Fix cliprect emit
  drm/radeon: r300_cmdbuf: Always emit INDX_BUFFER immediately after DRAW_INDEX
  radeon: fix some hard lockups on r3/4/500s

16 years ago[CIFS] Add destroy routine for dns_resolver
Jeff Layton [Wed, 27 Aug 2008 17:53:30 +0000 (17:53 +0000)]
[CIFS] Add destroy routine for dns_resolver

Otherwise, we're leaking the payload memory.

CC: Stable Kernel <stable@vger.kernel.org>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
16 years agoe1000: fix stack size
Linus Torvalds [Wed, 27 Aug 2008 08:14:46 +0000 (01:14 -0700)]
e1000: fix stack size

Here's the patch. It shrinks the stack from 1152 bytes to 192 bytes (the
first version, that only did the e1000_option part, got it down to 600
bytes). About half comes from not using multiple "e1000_option"
structures, the other half comes from turning the "e1000_opt_list[]"
arrays into "static const" instead, so that gcc doesn't copy them onto the
stack.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Reveiewed-by: Auke Kok <auke-jan.h.kok@intel.com>
Tested-by: Emil Tantilov <emil.s.tantilov@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block
Linus Torvalds [Wed, 27 Aug 2008 20:55:35 +0000 (13:55 -0700)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  block: remove blk_queue_tag_depth() and blk_queue_tag_queue()
  block: remove unused ->busy part of the block queue tag map
  bio: fix __bio_copy_iov() handling of bio->bv_len
  bio: fix bio_copy_kern() handling of bio->bv_len
  block: submit_bh() inadvertently discards barrier flag on a sync write
  block: clean up cmdfilter sysfs interface
  block: rename blk_scsi_cmd_filter to blk_cmd_filter
  sg: restore command permission for TYPE_SCANNER
  block: move cmdfilter from gendisk to request_queue

16 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfashe...
Linus Torvalds [Wed, 27 Aug 2008 20:54:55 +0000 (13:54 -0700)]
Merge branch 'upstream-linus' of git://git./linux/kernel/git/mfasheh/ocfs2

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh/ocfs2:
  ocfs2: Increment the reference count of an already-active stack.
  [PATCH] configfs: Consolidate locking around configfs_detach_prep() in configfs_rmdir()
  ocfs2: correctly set i_blocks after inline dir gets expanded
  ocfs2: Jump to correct label in ocfs2_expand_inline_dir()
  ocfs2: Fix sleep-with-spinlock recovery regression
  [PATCH] ocfs2/cluster/netdebug.c: fix warning
  [PATCH] ocfs2/cluster/tcp.c: make some functions static

16 years agoMerge branch 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6
Linus Torvalds [Wed, 27 Aug 2008 20:54:43 +0000 (13:54 -0700)]
Merge branch 'for-linus' of git://git390.osdl.marist.edu/linux-2.6

* 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6:
  [S390] dcss: fix build bug.
  [S390] Fix linker script.

16 years agoftrace: disable tracing for suspend to ram
Steven Rostedt [Wed, 27 Aug 2008 13:14:40 +0000 (09:14 -0400)]
ftrace: disable tracing for suspend to ram

I've been painstakingly debugging the issue with suspend to ram and
ftraced. The 2.6.28 code does not have this issue, but since the mcount
recording is not going to be in 27, this must be solved for the ftrace
daemon version.

The resume from suspend to ram would reboot because it was triple
faulting. Debugging further, I found that calling the mcount function
itself was not an issue, but it would fault when it incremented
preempt_count. preempt_count is on the tasks info structure that is on the
low memory address of the task's stack.  For some reason, it could not
write to it. Resuming out of suspend to ram does quite a lot of funny
tricks to get to work, so it is not surprising at all that simply doing a
preempt_disable() would cause a fault.

Thanks to Rafael for suggesting to add a "while (1);" to find the place in
resuming that is causing the fault. I would place the loop somewhere in
the code, compile and reboot and see if it would either reboot (hit the
fault) or simply hang (hit the loop).  Doing this over and over again, I
narrowed it down that it was happening in enable_nonboot_cpus.

At this point, I found that it is easier to simply disable tracing around
the suspend code, instead of searching for the particular function that
can not handle doing a preempt_disable.

This patch disables the tracer as it suspends and reenables it on resume.

I tested this patch on my Laptop, and it can resume fine with the patch.

Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoRevert "pktcdvd: push BKL down into driver"
Linus Torvalds [Wed, 27 Aug 2008 20:35:31 +0000 (13:35 -0700)]
Revert "pktcdvd: push BKL down into driver"

This reverts commit 5b6155ee70e9c4d2ad7e6f514c8eee06e2711c3a, because
the block device ioctl's really aren't ready for it.

In particular, the "struct file *" and the "struct inode *" arguments do
not necessarily match, which means that the unlocked version of the
ioctl (that only gets a "struct file *") isn't actually able to handle
the cases it needs to handle.

This fixes bugzilla

http://bugzilla.kernel.org/show_bug.cgi?id=11401

Reported-and-bisected-by: Laurent Riffard <laurent.riffard@free.fr>
Acked-by: Peter Osterlund <petero2@telia.com>
Cc: Alan Cox <alan@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years ago[ARM] Skip memory holes in FLATMEM when reading /proc/pagetypeinfo
Mel Gorman [Thu, 14 Aug 2008 10:10:14 +0000 (11:10 +0100)]
[ARM] Skip memory holes in FLATMEM when reading /proc/pagetypeinfo

Ordinarily, memory holes in flatmem still have a valid memmap and is safe
to use. However, an architecture (ARM) frees up the memmap backing memory
holes on the assumption it is never used. /proc/pagetypeinfo reads the
whole range of pages in a zone believing that the memmap is valid and that
pfn_valid will return false if it is not. On ARM, freeing the memmap breaks
the page->zone linkages even though pfn_valid() returns true and the kernel
can oops shortly afterwards due to accessing a bogus struct zone *.

This patch lets architectures say when FLATMEM can have holes in the
memmap. Rather than an expensive check for valid memory, /proc/pagetypeinfo
will confirm that the page linkages are still valid by checking page->zone
is still the expected zone. The lookup of page_zone is safe as there is a
limited range of memory that is accessed when calling page_zone.  Even if
page_zone happens to return the correct zone, the impact is that the counters
in /proc/pagetypeinfo are slightly off but fragmentation monitoring is
unlikely to be relevant on an embedded system.

Reported-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Tested-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] use bcd2bin/bin2bcd
Adrian Bunk [Fri, 8 Aug 2008 16:34:35 +0000 (19:34 +0300)]
[ARM] use bcd2bin/bin2bcd

This patch changes arm to use the new bcd2bin/bin2bcd functions instead
of the obsolete BCD_TO_BIN/BIN_TO_BCD macros.

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] use the new byteorder headers
Harvey Harrison [Tue, 19 Aug 2008 00:47:55 +0000 (17:47 -0700)]
[ARM] use the new byteorder headers

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years agoMerge branch 'for-rmk' of git://git.marvell.com/orion
Russell King [Wed, 27 Aug 2008 18:30:13 +0000 (19:30 +0100)]
Merge branch 'for-rmk' of git://git.marvell.com/orion

16 years ago[ARM] OMAP: Fix 2430 SMC91x ethernet IRQ
Russell King [Wed, 27 Aug 2008 13:29:40 +0000 (14:29 +0100)]
[ARM] OMAP: Fix 2430 SMC91x ethernet IRQ

Taken from omap 97b705ad835f1481270c4b67b402d6e37fa8ad15:
  ARM: OMAP: Misc compile fixes after syncing with mainline

  Also fix 2430 smc91x to use IRQ_LOWLEVEL.

Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] OMAP: Add and update OMAP default configuration files
Russell King [Wed, 27 Aug 2008 13:22:13 +0000 (14:22 +0100)]
[ARM] OMAP: Add and update OMAP default configuration files

Update omap_h2_1610 and omap_osk_5912 default configurations.

Add ams delta, n770, omap 2430sdp, apollon_2420, generic 1510, 1610,
1710, 2420, h4 2420, innovator 1510 and 1610, perseus2 730, palte,
palmtt, palmz71 and sx1 default configurations.

Pulled out of the omap zoom tree.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years ago[ARM] OMAP: Change mailing list for OMAP in MAINTAINERS
Dirk Behme [Sat, 15 Dec 2007 06:47:46 +0000 (07:47 +0100)]
[ARM] OMAP: Change mailing list for OMAP in MAINTAINERS

OMAP has now a list at vger.

Signed-off-by: Dirk Behme <dirk.behme@gmail.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
16 years agowan: Missing capability checks in sbni_ioctl()
Eugene Teo [Wed, 27 Aug 2008 11:50:30 +0000 (04:50 -0700)]
wan: Missing capability checks in sbni_ioctl()

There are missing capability checks in the following code:

1300 static int
1301 sbni_ioctl( struct net_device  *dev,  struct ifreq  *ifr,  int  cmd)
1302 {
[...]
1319     case  SIOCDEVRESINSTATS :
1320         if( current->euid != 0 )    /* root only */
1321             return  -EPERM;
[...]
1336     case  SIOCDEVSHWSTATE :
1337         if( current->euid != 0 )    /* root only */
1338             return  -EPERM;
[...]
1357     case  SIOCDEVENSLAVE :
1358         if( current->euid != 0 )    /* root only */
1359             return  -EPERM;
[...]
1372     case  SIOCDEVEMANSIPATE :
1373         if( current->euid != 0 )    /* root only */
1374             return  -EPERM;

Here's my proposed fix:

Missing capability checks.

Signed-off-by: Eugene Teo <eugeneteo@kernel.sg>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoMerge branch 'no-iwlwifi' of git://git.kernel.org/pub/scm/linux/kernel/git/linville...
David S. Miller [Wed, 27 Aug 2008 11:29:50 +0000 (04:29 -0700)]
Merge branch 'no-iwlwifi' of git://git./linux/kernel/git/linville/wireless-2.6

16 years agoMerge branch 'davem-fixes' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
David S. Miller [Wed, 27 Aug 2008 10:54:45 +0000 (03:54 -0700)]
Merge branch 'davem-fixes' of /linux/kernel/git/jgarzik/netdev-2.6

16 years agoe100, fix iomap read
Jiri Slaby [Thu, 14 Aug 2008 16:27:23 +0000 (18:27 +0200)]
e100, fix iomap read

There were 2 omitted readb's used on an iomap space. eliminate them
by using ioread8 instead.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Cc: Bruce Allan <bruce.w.allan@intel.com>
Cc: PJ Waskiewicz <peter.p.waskiewicz.jr@intel.com>
Cc: John Ronciak <john.ronciak@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoqeth: preallocated header account offset
Frank Blaschka [Fri, 15 Aug 2008 06:03:00 +0000 (08:03 +0200)]
qeth: preallocated header account offset

When a preallocated header qdio buffer is filled we have to account
the offset for the data length.

Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoqeth: l2 write unicast list to hardware
Frank Blaschka [Fri, 15 Aug 2008 06:02:59 +0000 (08:02 +0200)]
qeth: l2 write unicast list to hardware

In case the netdev unicast list contains additional entries we have
to register/deregister them.

Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoqeth: use -EOPNOTSUPP instead of -ENOTSUPP.
Ursula Braun [Fri, 15 Aug 2008 06:02:58 +0000 (08:02 +0200)]
qeth: use -EOPNOTSUPP instead of -ENOTSUPP.

return value -ENOTSUPP is not valid in userspace context, use
-EOPNOTSUPP instead.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoibm_newemac: Don't call dev_mc_add() before device is registered
Benjamin Herrenschmidt [Sun, 17 Aug 2008 21:29:23 +0000 (07:29 +1000)]
ibm_newemac: Don't call dev_mc_add() before device is registered

We must not call dev_mc_add() from within our HW configure which happens
before we initialize and register the netdev. Do it in open() instead.

Thanks to Sebastian Siewior for tracking it down.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agonet: don't grab a mutex within a timer context in gianfar
Sebastian Siewior [Tue, 19 Aug 2008 19:12:45 +0000 (21:12 +0200)]
net: don't grab a mutex within a timer context in gianfar

I got the following backtrace while network was unavailble:

|NETDEV WATCHDOG: eth0: transmit timed out
|BUG: sleeping function called from invalid context at /home/bigeasy/git/linux-2.6-powerpc/kernel/mutex.c:87
|in_atomic():1, irqs_disabled():0
|Call Trace:
|[c0383d90] [c0006dd8] show_stack+0x48/0x184 (unreliable)
|[c0383db0] [c001e938] __might_sleep+0xe0/0xf4
|[c0383dc0] [c025a43c] mutex_lock+0x24/0x3c
|[c0383de0] [c019005c] phy_stop+0x20/0x70
|[c0383df0] [c018d4ec] stop_gfar+0x28/0xf4
|[c0383e10] [c018e8c4] gfar_timeout+0x30/0x60
|[c0383e20] [c01fe7c0] dev_watchdog+0xa8/0x144
|[c0383e30] [c002f93c] run_timer_softirq+0x148/0x1c8
|[c0383e60] [c002b084] __do_softirq+0x5c/0xc4
|[c0383e80] [c00046fc] do_softirq+0x3c/0x54
|[c0383e90] [c002ac60] irq_exit+0x3c/0x5c
|[c0383ea0] [c000b378] timer_interrupt+0xe0/0xf8
|[c0383ec0] [c000e5ac] ret_from_except+0x0/0x18
|[c0383f80] [c000804c] cpu_idle+0xcc/0xdc
|[c0383fa0] [c025c07c] etext+0x7c/0x90
|[c0383fc0] [c0338960] start_kernel+0x294/0x2a8
|[c0383ff0] [c00003dc] skpinv+0x304/0x340
|------------[ cut here ]------------

The phylock was once a spinlock but got changed into a mutex via
commit 35b5f6b1a aka [PHYLIB: Locking fixes for PHY I/O potentially sleeping]

Signed-off-by: Sebastian Siewior <bigeasy@linutronix.de>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoforcedeth: fix checksum flag
Ayaz Abdulla [Wed, 20 Aug 2008 23:34:37 +0000 (16:34 -0700)]
forcedeth: fix checksum flag

Fix the checksum feature advertised in device flags.  The hardware support
TCP/UDP over IPv4 and TCP/UDP over IPv6 (without IPv6 extension headers).
However, the kernel feature flags do not distinguish IPv6 with/without
extension headers.

Therefore, the driver needs to use NETIF_F_IP_CSUM instead of
NETIF_F_HW_CSUM since the latter includes all IPv6 packets.

A future patch can be created to check for extension headers and perform
software checksum calculation.

Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Cc: Jeff Garzik <jgarzik@pobox.com>
Cc: Manfred Spraul <manfred@colorfullife.com
Cc: <stable@kernel.org> [2.6.25.x, 2.6.26.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agonet/usb/mcs7830: add set_mac_address
Oliver Martin [Sat, 23 Aug 2008 20:08:47 +0000 (22:08 +0200)]
net/usb/mcs7830: add set_mac_address

Implement set_mac_address for mcs7830. This enables me to use it with my
cable modem.

Signed-off-by: Oliver Martin <oliver.martin@student.tuwien.ac.at>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agonet/usb/mcs7830: new device IDs
Arnd Bergmann [Sat, 23 Aug 2008 20:02:23 +0000 (22:02 +0200)]
net/usb/mcs7830: new device IDs

This adds USB device IDs for MosChip 7730 and Sitecom LN030
to the mcs7830 driver. The IDs have been reported to work without
further modifications.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: David Brownell <david-b@pacbell.net>
Cc: Viktor Horvath <ViktorHorvath@gmx.net>
Cc: Robbert Wethmar <robbert@wethmar.nl>
Cc: Bart van der Klip <bklip@xs4all.nl>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years ago[netdrvr] smc91x: fix resource removal (null ptr deref)
Jeff Garzik [Wed, 27 Aug 2008 09:54:30 +0000 (05:54 -0400)]
[netdrvr] smc91x: fix resource removal (null ptr deref)

Properly handle resource cleanup on unplug/exit.

Spotted by Jonathan Cameron

Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoibmveth: fix bad UDP checksums
Santiago Leon [Wed, 20 Aug 2008 19:09:19 +0000 (13:09 -0600)]
ibmveth: fix bad UDP checksums

This patch fixes a ibmveth bug where bad UDP checksums are being transmitted
when checksum offloading is enabled.
The hypervisor does checksum offloading only on TCP packets, so ibmveth calls
skb_checksum_help() for any other protocol.  The bug happens because
the packet is being modified after the DMA map, so we would need a memory
barrier before making the hypervisor call.  Reordering the code so that the
DMA map happens after skb_checksum_help() has the additional advantage of
fixing a DMA map leak if skb_checksum_help() where to fail.

Signed-off-by: Santiago Leon <santil@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years ago[netdrvr] hso: dev_kfree_skb crash fix
Denis Joseph Barrow [Wed, 20 Aug 2008 01:07:55 +0000 (18:07 -0700)]
[netdrvr] hso: dev_kfree_skb crash fix

Fixes dev_kfree_skb happening too many times when hso_start_net_device
is called from hso_resume.

Signed-off-by: Denis Joseph Barrow <D.Barow@option.com>
Cc: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years ago[netdrvr] hso: icon 322 detection fix
Denis Joseph Barrow [Wed, 20 Aug 2008 01:07:52 +0000 (18:07 -0700)]
[netdrvr] hso: icon 322 detection fix

Fixes Icon-322 detection.

Signed-off-by: Denis Joseph Barrow <D.Barow@option.com>
Cc: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoatl1: disable TSO by default
Jay Cliburn [Tue, 19 Aug 2008 00:28:13 +0000 (19:28 -0500)]
atl1: disable TSO by default

The atl1 driver is causing stalled connections and file corruption
whenever TSO is enabled.  Two examples are here:

http://lkml.org/lkml/2008/7/15/325
http://lkml.org/lkml/2008/8/18/543

Disable TSO by default until we can determine the source of the
problem.

Signed-off-by: Jay Cliburn <jacliburn@bellsouth.net>
cc: stable@kernel.org
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoatl1e: multistatement if missing braces
Ilpo Järvinen [Tue, 19 Aug 2008 07:32:06 +0000 (10:32 +0300)]
atl1e: multistatement if missing braces

Doesn't cause problems (yet) because err gets zeroed earlier.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoigb: remove 82576 quad adapter
Alexander Duyck [Fri, 8 Aug 2008 23:51:26 +0000 (16:51 -0700)]
igb: remove 82576 quad adapter

Disable support for device 8086:10E8.  Currently the result of loading the
driver with the device present causes system instability.

Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agodrivers/net/skfp/ess.c: fix compile warnings
Takashi Iwai [Wed, 20 Aug 2008 23:34:35 +0000 (16:34 -0700)]
drivers/net/skfp/ess.c: fix compile warnings

  CC [M]  drivers/net/skfp/ess.o
drivers/net/skfp/ess.c: In function 'ess_send_response':
drivers/net/skfp/ess.c:513: warning: cast from pointer to integer of different size
drivers/net/skfp/ess.c: In function 'ess_send_alc_req':
drivers/net/skfp/ess.c:609: warning: cast from pointer to integer of different size
drivers/net/skfp/ess.c:639: warning: cast from pointer to integer of different size

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoipv4: mode 0555 in ipv4_skeleton
Hugh Dickins [Wed, 27 Aug 2008 09:35:18 +0000 (02:35 -0700)]
ipv4: mode 0555 in ipv4_skeleton

vpnc on today's kernel says Cannot open "/proc/sys/net/ipv4/route/flush":
d--------- 0 root root 0 2008-08-26 11:32 /proc/sys/net/ipv4/route
d--------- 0 root root 0 2008-08-26 19:16 /proc/sys/net/ipv4/neigh

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Acked-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agotcp: fix tcp header size miscalculation when window scale is unused
Philip Love [Wed, 27 Aug 2008 09:33:50 +0000 (02:33 -0700)]
tcp: fix tcp header size miscalculation when window scale is unused

The size of the TCP header is miscalculated when the window scale ends
up being 0. Additionally, this can be induced by sending a SYN to a
passive open port with a window scale option with value 0.

Signed-off-by: Philip Love <love_phil@emc.com>
Signed-off-by: Adam Langley <agl@imperialviolet.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agopkt_sched: Fix sch_tree_lock()
Jarek Poplawski [Wed, 27 Aug 2008 09:27:10 +0000 (02:27 -0700)]
pkt_sched: Fix sch_tree_lock()

Use new qdisc_root_sleeping_lock() instead of qdisc_root_lock() as
sch_tree_lock() because this lock could be used while dev is
deactivated, but we never need to use this with noop_qdisc as a root.

Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agopkt_sched: Fix gen_estimator locks
Jarek Poplawski [Wed, 27 Aug 2008 09:25:17 +0000 (02:25 -0700)]
pkt_sched: Fix gen_estimator locks

While passing a qdisc root lock to gen_new_estimator() and
gen_replace_estimator() dev could be deactivated or even before
grafting proper root qdisc as qdisc_sleeping (e.g. qdisc_create), so
using qdisc_root_lock() is not enough. This patch adds
qdisc_root_sleeping_lock() for this, plus additional checks, where
necessary.

Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agopkt_sched: Use rcu_assign_pointer() to change dev_queue->qdisc
Jarek Poplawski [Wed, 27 Aug 2008 09:22:07 +0000 (02:22 -0700)]
pkt_sched: Use rcu_assign_pointer() to change dev_queue->qdisc

These pointers are RCU protected, so proper primitives should be used.

Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoLCS recovery dumps when cable reconnect
Klaus-D. Wacker [Thu, 21 Aug 2008 15:10:25 +0000 (17:10 +0200)]
LCS recovery dumps when cable reconnect

LCS recovery dumps in irq routine when CCW address in
Subchannel Status Word (SCSW) is zero. This occurs
when recovery is driven after cable reconnect.

Signed-off-by: Klaus-D. Wacker <kdwacker@de.ibm.com>
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoctcm: netdev->priv vs. netdev->ml_priv
Peter Tiedemann [Thu, 21 Aug 2008 15:10:24 +0000 (17:10 +0200)]
ctcm: netdev->priv vs. netdev->ml_priv

Use netdev->ml_priv instead of netdev->priv

Signed-off-by: Peter Tiedemann <ptiedem@de.ibm.com>
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoclaw: netdev->priv vs. netdev->ml_priv
Peter Tiedemann [Thu, 21 Aug 2008 15:10:23 +0000 (17:10 +0200)]
claw: netdev->priv vs. netdev->ml_priv

Use netdev->ml_priv instead of netdev->priv

Signed-off-by: Peter Tiedemann <ptiedem@de.ibm.com>
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoclaw: fix memory leak in claw_probe.
Martin Schwidefsky [Thu, 21 Aug 2008 15:10:22 +0000 (17:10 +0200)]
claw: fix memory leak in claw_probe.

probe_error() frees memory only, if cgdev->dev.driver_data refers
to the claw_privbk structure. Move forward its setting in claw_probe()
to ensure proper freeing of claw_privbk allocations.

Cc: Daniel <danielm77@spray.se>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>