Saves us one cycle of alloc-add-free if the queue was full.
Signed-off-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (modified)

Now we can tell the theoretical capacity remaining in the output
queue, virtio_net can waste entries by stopping the queue early.

It doesn't work in the case of indirect buffers and kmalloc failure,
but that's rare (we could drop the packet in that case, but other
drivers return TX_BUSY for similar reasons).

For the record, I think this patch reflects poorly on the linux
network API.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Dinesh Subhraveti <dineshs@us.ibm.com>

We put the virtio_net_hdr into the skb's cb region; turn this into a
union to clean up the code slightly and allow future expansion.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Mark McLoughlin <markmc@redhat.com>
Cc: Dinesh Subhraveti <dineshs@us.ibm.com>

The virtio_net driver is complicated by the two methods of freeing old
xmit buffers (in addition to freeing old ones at the start of the xmit
path).

The original code used a 1/10 second timer attached to xmit_free(),
reset on every xmit.  Before we orphaned skbs on xmit, the
transmitting userspace could block with a full socket until the timer
fired, the skb destructor was called, and they were re-woken.

So we added the VIRTIO_F_NOTIFY_ON_EMPTY feature: supporting devices
send an interrupt (even if normally suppressed) on an empty xmit ring
which makes us schedule xmit_tasklet().  This was a benchmark win.

Unfortunately, VIRTIO_F_NOTIFY_ON_EMPTY makes quite a lot of work: a
host which is faster than the guest will fire the interrupt every xmit
packet (slowing the guest down further).  Attempting mitigation in the
host adds overhead of userspace timers (possibly with the additional
pain of signals), and risks increasing latency anyway if you get it
wrong.

In practice, this effect was masked by benchmarks which take advantage
of GSO (with its inherent transmit batching), but it's still there.

Now we orphan xmitted skbs, the pressure is off: remove both paths and
no longer request VIRTIO_F_NOTIFY_ON_EMPTY.  Note that the current
QEMU will notify us even if we don't negotiate this feature (legal,
but suboptimal); a patch is outstanding to improve that.

Move the skb_orphan/nf_reset to after we've done the send and notified
the other end, for a slight optimization.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Mark McLoughlin <markmc@redhat.com>

This effectively reverts 99ffc696
"virtio: wean net driver off NETDEV_TX_BUSY".

The complexity of queuing an skb (setting a tasklet to re-xmit) is
questionable, especially once we get rid of the other reason for the
tasklet in the next patch.

If the skb won't fit in the tx queue, just return NETDEV_TX_BUSY.
This is frowned upon, so a followup patch uses a more complex solution.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Herbert Xu <herbert@gondor.apana.org.au>

The complex transmit free logic was introduced to avoid hangs on
removing the ip_conntrack module and also because drivers aren't
generally supposed to keep stale skbs for unbounded times.

After some debate, it was decided that while doing skb_orphan()
generally is a rat's nest, we can do it in this driver.  Following
patches take advantage of this.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

* 'ixp4xx' of git://git.kernel.org/pub/scm/linux/kernel/git/chris/linux-2.6:
  Add MAINTAINERS entry for ARM/INTEL IXP4xx arch support.
  ixp4xx: arch_idle() documentation fixup
  ixp4xx: timer and clocks cleanups

Fix new kernel-doc warnings in serial_core.[hc] files.

  Warning(include/linux/serial_core.h:485): No description found for parameter 'uport'
  Warning(include/linux/serial_core.h:485): Excess function parameter 'port' description in 'uart_handle_dcd_change'
  Warning(include/linux/serial_core.h:511): No description found for parameter 'uport'
  Warning(include/linux/serial_core.h:511): Excess function parameter 'port' description in 'uart_handle_cts_change'
  Warning(drivers/serial/serial_core.c:2437): No description found for parameter 'uport'
  Warning(drivers/serial/serial_core.c:2437): Excess function parameter 'port' description in 'uart_add_one_port'
  Warning(drivers/serial/serial_core.c:2509): No description found for parameter 'uport'
  Warning(drivers/serial/serial_core.c:2509): Excess function parameter 'port' description in 'uart_remove_one_port'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Commit ebd2c8f6 removed struct uart_info and
commit bdc04e31 further moved delta_msr_wait.
Fix up the MN10300 on-chip serial port drivers to comply with this.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Use asm/generic-hardirq.h to build asm/hardirq.h and also remove the unused
idle_timestamp field in irq_cpustat whilst we're at it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Merge branch 'x86/ptrace-syscall-exit' of git://git.kernel.org/pub/scm/linux/kernel/git/frob/linux-2.6-roland

* 'x86/ptrace-syscall-exit' of git://git.kernel.org/pub/scm/linux/kernel/git/frob/linux-2.6-roland:
  x86: ptrace: sysret path should reach syscall_trace_leave

* git://git.infradead.org/battery-2.6:
  power_supply: Add driver for the PMU on WM831x PMICs
  ds2760_battery: Fix integer overflow for time_to_empty_now
  wm97xx_battery: Convert to dev_pm_ops
  wm97xx_battery: Use irq to detect charger state
  wm97xx_battery: Use platform_data
  wm97xx-core: Pass platform_data to battery
  ds2760_battery: implement set_charged() feature
  power_supply: get_by_name and set_charged functionality
  power_supply: EXPORT_SYMBOL cleanups
  ds2760_battery: add current_accum module parameter
  ds2760_battery: handle full_active_uAh == 0 case correctly
  ds2760_battery: add rated_capacity module parameter
  ds2760_battery: export more features
  ds2760_battery: delay power supply registration
  wm8350_power: Implement charge type property
  power_supply: Add a charge_type property, and use it for olpc driver
  olpc_battery: Add an 'error' sysfs device that displays raw errors
  Revert "power: remove POWER_SUPPLY_PROP_CAPACITY_LEVEL"

* 'drm-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/radeon/r600: set correct pitch for 4 byte copy
  drm/radeon: consolidate family flags used in pciids.

* git://git.infradead.org/mtd-2.6: (58 commits)
  mtd: jedec_probe: add PSD4256G6V id
  mtd: OneNand support for Nomadik 8815 SoC (on NHK8815 board)
  mtd: nand: driver for Nomadik 8815 SoC (on NHK8815 board)
  m25p80: Add Spansion S25FL129P serial flashes
  jffs2: Use SLAB_HWCACHE_ALIGN for jffs2_raw_{dirent,inode} slabs
  mtd: sh_flctl: register sh_flctl using platform_driver_probe()
  mtd: nand: txx9ndfmc: transfer 512 byte at a time if possible
  mtd: nand: fix tmio_nand ecc correction
  mtd: nand: add __nand_correct_data helper function
  mtd: cfi_cmdset_0002: add 0xFF intolerance for M29W128G
  mtd: inftl: fix fold chain block number
  mtd: jedec: fix compilation problem with I28F640C3B definition
  mtd: nand: fix ECC Correction bug for SMC ordering for NDFC driver
  mtd: ofpart: Check availability of reg property instead of name property
  driver/Makefile: Initialize "mtd" and "spi" before "net"
  mtd: omap: adding DMA mode support in nand prefetch/post-write
  mtd: omap: add support for nand prefetch-read and post-write
  mtd: add nand support for w90p910 (v2)
  mtd: maps: add mtd-ram support to physmap_of
  mtd: pxa3xx_nand: add single-bit error corrections reporting
  ...

* 'vgaarb-fix' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  vgaarb: make client interface config invariant.

* git://git.infradead.org/iommu-2.6: (23 commits)
  intel-iommu: Disable PMRs after we enable translation, not before
  intel-iommu: Kill DMAR_BROKEN_GFX_WA option.
  intel-iommu: Fix integer wrap on 32 bit kernels
  intel-iommu: Fix integer overflow in dma_pte_{clear_range,free_pagetable}()
  intel-iommu: Limit DOMAIN_MAX_PFN to fit in an 'unsigned long'
  intel-iommu: Fix kernel hang if interrupt remapping disabled in BIOS
  intel-iommu: Disallow interrupt remapping if not all ioapics covered
  intel-iommu: include linux/dmi.h to use dmi_ routines
  pci/dmar: correct off-by-one error in dmar_fault()
  intel-iommu: Cope with yet another BIOS screwup causing crashes
  intel-iommu: iommu init error path bug fixes
  intel-iommu: Mark functions with __init
  USB: Work around BIOS bugs by quiescing USB controllers earlier
  ia64: IOMMU passthrough mode shouldn't trigger swiotlb init
  intel-iommu: make domain_add_dev_info() call domain_context_mapping()
  intel-iommu: Unify hardware and software passthrough support
  intel-iommu: Cope with broken HP DC7900 BIOS
  iommu=pt is a valid early param
  intel-iommu: double kfree()
  intel-iommu: Kill pointless intel_unmap_single() function
  ...

Fixed up trivial include lines conflict in drivers/pci/intel-iommu.c

Impact: cleanup

No need for redeclaration.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lrg/voltage-2.6: (41 commits)
  regulator: Add some brief design documentation
  regulator: fix voltage range in da9034 ldo12
  regulator/driver: be more specific in nanodoc for is_enabled
  regulator/lp3971: drop unnecessary initialization
  regulator: drop 'default n'
  regulator: fix typos
  regulator: fix calculation of voltage range in da9034_set_ldo12_voltage()
  regulator: update a filename in documentation
  drivers/regulator/Kconfig: fix typo (s/Usersapce/Userspace/) in REGULATOR_USERSPACE_CONSUMER description
  REGULATOR Handle positive returncode from enable
  regulator: tps650xx - build fixes for x86_64
  Fix some regulator documentation
  Regulator: Adding TPS65023 and TPS6507x in Kconfig and Makefile
  Regulator: Add TPS6507x regulator driver
  Regulator: Add TPS65023 regulator driver
  regulator: userspace: use sysfs_create_group
  regulator: Add GPIO enable control to fixed voltage regulator driver
  Regulator: Implement list_voltage for pcf50633 regulator driver.
  regulator: regulator_enable() permission checking
  regulator: Push locking for regulator_is_enabled() out
  ...

* 'fix/misc' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ALSA: lx6464es - remove unused struct member
  ALSA: lx6464es - cleanup of rmh message bus function
  ALSA: pcm - Simplify snd_pcm_drain() implementation

* 'fix/asoc' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ASoC: wm8753: fix mapping when MONOMIX is set to Stereo
  ASoC: some minor changes for AD1836 and AD1938 codec drivers
  ASoC: DaVinci: Fixes to McASP configuration
  ASoC: Blackfin I2S: fix resuming when device hasn't been used
  ASoC: Blackfin I2S: add lost platform_device parameter to resume function
  ASoC: fix typos in Blackfin headers
  ASoC: bf5xx-sport: the irq save/restore funcs take an unsigned long
  ASoC: Blackfin AC97: add a few missing multichannel define handling

* 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6: (22 commits)
  [S390] Update default configuration.
  [S390] hibernate: Do real CPU swap at resume time
  [S390] dasd: tolerate devices that have no feature codes
  [S390] zcrypt: Do not add/remove devices in s/r callbacks
  [S390] hibernate: make sure pfn_is_nosave handles lowcore pages
  [S390] smp: introduce LC_ORDER and simplify lowcore handling
  [S390] ptrace: use common code for simple peek/poke operations
  [S390] fix disabled_wait inline assembly clobber list
  [S390] Change kernel_page_present coding style.
  [S390] hibernation: reset system after resume
  [S390] hibernation: fix guest page hinting related crash
  [S390] Get rid of init_module/delete_module compat functions.
  [S390] Convert sys_execve to function with parameters.
  [S390] Convert sys_clone to function with parameters.
  [S390] qdio: change state of all primed input buffers
  [S390] qdio: reduce per device debug messages
  [S390] cio: introduce consistent subchannel scanning
  [S390] cio: idset use actual number of ssids
  [S390] cio: dont kfree vmalloced memory
  [S390] cio: introduce css_settle
  ...

* 'timers-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  itimers: Add tracepoints for itimer
  hrtimer: Add tracepoint for hrtimers
  timers: Add tracepoints for timer_list timers
  cputime: Optimize jiffies_to_cputime(1)
  itimers: Simplify arm_timer() code a bit
  itimers: Fix periodic tics precision
  itimers: Merge ITIMER_VIRT and ITIMER_PROF

Trivial header file include conflicts in kernel/fork.c

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6:
  ieee1394: sbp2: remove a workaround for Momobay FX-3A
  firewire: sbp2: remove a workaround for Momobay FX-3A
  firewire: sbp2: fix status reception
  firewire: core: fix topology map response handler
  firewire: core: fix race with parallel PCI device probe
  firewire: core: header file cleanup
  firewire: ohci: fix Self ID Count register mask (safeguard against buffer overflow)
  ieee1394: raw1394: Do not leak memory on failed trylock.

* 'sfi-release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-sfi-2.6:
  SFI: remove unneeded includes
  sfi: Remove unused code
  SFI: Hook PCI MMCONFIG
  x86: add arch-specific SFI support
  SFI: add capability to parse ACPI tables
  SFI: add platform-independent core support
  SFI: create linux/sfi.h
  SFI: Simple Firmware Interface - MAINTAINERS, Kconfig

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6: (119 commits)
  ACPI: don't pass handle for fixed hardware notifications
  ACPI: remove null pointer checks in deferred execution path
  ACPI: simplify deferred execution path
  acerhdf: additional BIOS versions
  acerhdf: convert to dev_pm_ops
  acerhdf: fix fan control for AOA150 model
  thermal: add missing Kconfig dependency
  acpi: switch /proc/acpi/{debug_layer,debug_level} to seq_file
  hp-wmi: fix rfkill memory leak on unload
  ACPI: remove unnecessary #ifdef CONFIG_DMI
  ACPI: linux/acpi.h should not include linux/dmi.h
  hwmon driver for ACPI 4.0 power meters
  topstar-laptop: add new driver for hotkeys support on Topstar N01
  thinkpad_acpi: fix rfkill memory leak on unload
  thinkpad-acpi: report brightness events when required
  thinkpad-acpi: don't poll by default any of the reserved hotkeys
  thinkpad-acpi: Fix procfs hotkey reset command
  thinkpad-acpi: deprecate hotkey_bios_mask
  thinkpad-acpi: hotkey poll fixes
  thinkpad-acpi: be more strict when detecting a ThinkPad
  ...

* 'i2c-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
  i2c: Clearly mark ACPI drivers as such
  i2c: Add driver for SMBus Control Method Interface
  i2c-pnx: Correct use of request_region/request_mem_region
  MAINTAINERS: Add maintainer for AT24 and PCA9564/PCA9665
  i2c-piix4: Add AMD SB900 SMBus device ID
  i2c/chips: Remove deprecated pcf8574 driver
  i2c/chips: Remove deprecated pca9539 driver
  i2c/chips: Remove deprecated pcf8575 driver
  gpio/pcf857x: Copy i2c_device_id from old pcf8574 driver
  i2c/scx200_acb: Provide more information on bus errors
  i2c: Provide compatibility links for i2c adapters
  i2c: Convert i2c adapters to bus devices
  i2c: Convert i2c clients to a device type
  i2c/tsl2550: Use combined SMBus transactions
  i2c-taos-evm: Switch echo off to improve performance
  i2c: Drop unused i2c_driver.id field

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2: (85 commits)
  ocfs2: Use buffer IO if we are appending a file.
  ocfs2: add spinlock protection when dealing with lockres->purge.
  dlmglue.c: add missed mlog lines
  ocfs2: __ocfs2_abort() should not enable panic for local mounts
  ocfs2: Add ioctl for reflink.
  ocfs2: Enable refcount tree support.
  ocfs2: Implement ocfs2_reflink.
  ocfs2: Add preserve to reflink.
  ocfs2: Create reflinked file in orphan dir.
  ocfs2: Use proper parameter for some inode operation.
  ocfs2: Make transaction extend more efficient.
  ocfs2: Don't merge in 1st refcount ops of reflink.
  ocfs2: Modify removing xattr process for refcount.
  ocfs2: Add reflink support for xattr.
  ocfs2: Create an xattr indexed block if needed.
  ocfs2: Call refcount tree remove process properly.
  ocfs2: Attach xattr clusters to refcount tree.
  ocfs2: Abstract ocfs2 xattr tree extend rec iteration process.
  ocfs2: Abstract the creation of xattr block.
  ocfs2: Remove inode from ocfs2_xattr_bucket_get_name_value.
  ...

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (142 commits)
  USB: Fix sysfs paths in documentation
  USB: skeleton: fix coding style issues.
  USB: O_NONBLOCK in read path of skeleton
  USB: make usb-skeleton honor O_NONBLOCK in write path
  USB: skel_read really sucks royally
  USB: Add hub descriptor update hook for xHCI
  USB: xhci: Support USB hubs.
  USB: xhci: Set multi-TT field for LS/FS devices under hubs.
  USB: xhci: Set route string for all devices.
  USB: xhci: Fix command wait list handling.
  USB: xhci: Change how xHCI commands are handled.
  USB: xhci: Refactor input device context setup.
  USB: xhci: Endpoint representation refactoring.
  USB: gadget: ether needs to select CRC32
  USB: fix USBTMC get_capabilities success handling
  USB: fix missing error check in probing
  USB: usbfs: add USBDEVFS_URB_BULK_CONTINUATION flag
  USB: support for autosuspend in sierra while online
  USB: ehci-dbgp,ehci: Allow dbpg to work with suspend/resume
  USB: ehci-dbgp,documentation: Documentation updates for ehci-dbgp
  ...

* git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-for-linus:
  lguest: don't force VIRTIO_F_NOTIFY_ON_EMPTY
  lguest: cleanup for map_switcher()
  lguest: use PGDIR_SHIFT for PAE code to allow different PAGE_OFFSET
  lguest: use set_pte/set_pmd uniformly for real page table entries
  lguest: move panic notifier registration to its expected place.
  virtio_blk: add support for cache flush
  virtio: add virtio IDs file
  virtio: get rid of redundant VIRTIO_ID_9P definition
  virtio: make add_buf return capacity remaining
  virtio_pci: minor MSI-X cleanups

My 58fa879e "mm: FOLL flags for GUP flags"
broke CONFIG_NOMMU build by forgetting to update nommu.c foll_flags type:

  mm/nommu.c:171: error: conflicting types for `__get_user_pages'
  mm/internal.h:254: error: previous declaration of `__get_user_pages' was here
  make[1]: *** [mm/nommu.o] Error 1

My 03f6462a "mm: move highest_memmap_pfn"
broke CONFIG_NOMMU build by forgetting to add a nommu.c highest_memmap_pfn:

  mm/built-in.o: In function `memmap_init_zone':
  (.meminit.text+0x326): undefined reference to `highest_memmap_pfn'
  mm/built-in.o: In function `memmap_init_zone':
  (.meminit.text+0x32d): undefined reference to `highest_memmap_pfn'

Fix both breakages, and give myself 30 lashes (ouch!)
Reported-by: Michal Simek <michal.simek@petalogix.com>
Signed-off-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

For this system call user space passes a signed long length parameter,
while the kernel side takes an unsigned long parameter and converts it
later to signed long again.

This has led to bugs in compat wrappers see e.g.  dd90bbd5 "powerpc: Add
compat_sys_truncate".  The s390 compat wrapper for this functions is
broken as well since it also performs zero extension instead of sign
extension for the length parameter.

In addition if hpa comes up with an automated way of generating
compat wrappers it would generate a wrong one here.

So change the length parameter from unsigned long to long.

Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

* 'for-linus' of git://gitserver.sunplusct.com/linux-2.6-score:
  score: update email address in MAINTAINERS.
  score: Cleanup linker script using new macros.
  score: Make THREAD_SIZE available to assembly files.
  score: Make PAGE_SIZE available to assembly.

* 'x86/orig_ax' of git://git.kernel.org/pub/scm/linux/kernel/git/frob/linux-2.6-roland:
  x86: ptrace: set TS_COMPAT when 32-bit ptrace sets orig_eax>=0
  x86: ptrace: do not sign-extend orig_ax on write
  x86: syscall_get_nr returns int
  asm-generic: syscall_get_nr returns int

bitfields should be unsigned.

This fixes sparse noise:
  error: dubious one-bit signed bitfield
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Jan Kara <jack@ucw.cz>
Cc: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Unlike on most other architectures ino_t is an unsigned int on s390.  So
add an explicit cast to avoid this compile warning:

fs/ext2/namei.c: In function 'ext2_lookup':
fs/ext2/namei.c:73: warning: format '%lu' expects type 'long unsigned int', but argument 4 has type 'ino_t'
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

There are a few places in the Minix FS code where the "inode" field of a
minix_dir_entry is used without checking first to see if the dirent is
really a minix3_dir_entry.  The inode number in a V1/V2 dirent is 16 bits,
whereas that in a V3 dirent is 32 bits.

Accessing it as a 16 bit field when it really should be accessed as a 32
bit field probably kinda sorta works on a little-endian machine, but leads
to some rather odd behaviour on big-endian machines.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Doug Graham <dgraham@nortel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

The patch enables the driver to be used on platforms such as ARM where an
I/O address is a 32-bit memory address.

The patch avoids the following kernel oops:

debian:~# modprobe sisfb
[   73.070000] sisfb: Video ROM found
[   73.080000] sisfb: Video RAM at 0x80000000, mapped to 0xe0a00000, size 1024k
[   73.090000] sisfb: MMIO at 0x84080000, mapped to 0xe0b80000, size 256k
[   73.090000] sisfb: Memory heap starting at 800K, size 32K
[   73.360000] Unable to handle kernel paging request at virtual address 6e000844
[   73.380000] pgd = df230000
[   73.380000] [6e000844] *pgd=00000000
[   73.380000] Internal error: Oops: 8f5 [#1]
[   73.380000] Modules linked in: sisfb(+) fb cfbcopyarea cfbimgblt cfbfillrect
[   73.380000] CPU: 0    Not tainted  (2.6.31-iop32x #1)
[   73.380000] PC is at SiS_SetRegANDOR+0x10/0x38 [sisfb]
[   73.380000] LR is at SiS_SetSCLKHigh+0x38/0x94 [sisfb]
[   73.380000] pc : [<bf01dc00>]    lr : [<bf0238f8>]    psr: 60000013
[   73.380000] sp : df38fd00  ip : 6e000000  fp : 00000002
[   73.380000] r10: 00000108  r9 : 00000000  r8 : 00000108
[   73.380000] r7 : df064258  r6 : 00000110  r5 : 6e000844  r4 : 0000010a
[   73.380000] r3 : 00000001  r2 : 0000000e  r1 : 00000011  r0 : 00000844
[   73.380000] Flags: nZCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment user
[   73.380000] Control: 0000397f  Table: bf230000  DAC: 00000015
[   73.380000] Process modprobe (pid: 1849, stack limit = 0xdf38e270)
[   73.380000] Stack: (0xdf38fd00 to 0xdf390000)
[   73.380000] fd00: 0000010a 00000108 df064258 df064258 df064258 00000000 00000000 bf02c4e0
[   73.380000] fd20: 00000114 bf02c50c 00000013 00000114 0000010a df064258 00000000 bf02c980
[   73.380000] fd40: 00009c66 00000004 00000001 df064250 a0010000 a6a2a0a0 df064250 00000003
[   73.380000] fd60: df064250 00000000 df064258 0000fffd 00000000 00000000 00000000 bf033948
[   73.380000] fd80: 00000000 00000000 00000000 bf019e2c 00000000 df064a70 bf03b470 00010000
[   73.380000] fda0: 00000000 df064250 00000000 df831c00 00000012 bf039f70 00000000 c00abed8
[   73.380000] fdc0: 000008a6 000008a4 df0649b0 df064878 df064258 df064000 00000000 00000000
[   73.380000] fde0: 00000001 00008000 00000001 00030000 df81c930 bf049f88 df831c00 00000000
[   73.380000] fe00: bf049f58 df3952a0 c0447708 bf049f88 bf049fe0 c0191980 df831c00 c0191b10
[   73.380000] fe20: df831c58 bf049f58 df831c00 bf04aca8 df3952a0 df831c58 df831c58 bf049f88
[   73.380000] fe40: c01ba1b4 c01ba0a0 df831c58 df831c8c bf049f88 c01ba1b4 df3952a0 00000000
[   73.380000] fe60: c03e265c c01ba240 00000000 df38fe78 bf049f88 c01b990c df812938 df81b8d0
[   73.380000] fe80: df3952a0 df807780 00000000 00000060 bf049f88 c01b9224 bf0429c8 00000000
[   73.380000] fea0: bf049f58 00000000 bf049f88 00000000 00000000 bf04aea8 00000000 c01ba4e4
[   73.380000] fec0: e09861a0 bf049f58 00000000 bf049f88 00000000 c0191f20 00000000 00000000
[   73.380000] fee0: c03f7bac bf04d418 0000fff2 0000fff1 bf04ad08 0002f260 0002f260 e0986038
[   73.380000] ff00: e0986150 e098568b df143340 e0990280 00000036 c03d8b00 fffffffd 00000000
[   73.380000] ff20: bf04acfc 00000000 fffffffc 0003cf4b 00018098 c03f7bac 00000000 bf04d000
[   73.380000] ff40: df38e000 00000000 bedc0984 c00272a4 ffffffff c005bc88 00000000 00000000
[   73.380000] ff60: 0003cf4b 0003cf4b 00018098 bf04acfc 00000000 c0027fe8 df38e000 00000000
[   73.380000] ff80: bedc0984 c006882c 00001000 00000003 00000000 00009064 00000000 00008edc
[   73.380000] ffa0: 00000080 c0027e20 00009064 00000000 4014e000 0003cf4b 00018098 0003cf4b
[   73.380000] ffc0: 00009064 00000000 00008edc 00000080 00000000 00000000 40025000 bedc0984
[   73.380000] ffe0: 00000000 bedc08fc 0000b6b0 400e8f34 60000010 4014e000 00000000 00000000
[   73.380000] [<bf01dc00>] (SiS_SetRegANDOR+0x10/0x38 [sisfb]) from [<df064258>] (0xdf064258)
[   73.380000] Code: e92d0030 e20110ff e280546e e3a0c46e (e5c51000)
[   73.680000] ---[ end trace 62a93e01df37a5f2 ]---
Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
Cc: Thomas Winischhofer <thomas@winischhofer.net>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Fixes `s3c_fb_remove' referenced in section `.data' of
drivers/built-in.o: defined in discarded section `.devexit.text' of
drivers/built-in.o

With CONFIG_HOTPLUG=n, functions marked with __devexit gets removed,
so make sure we use __devexit_p when referencing pointers to them.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Acked-by: Ben Dooks <ben-linux@fluff.org>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/video/tmiofb.c: In function 'tmiofb_resume':
drivers/video/tmiofb.c:977: warning: 'retval' may be used uninitialized in this function
Acked-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fbcon makes the (reasonable) assumption that it only needs to program the
hardware once, when fbcon_init() is called for the foreground console.

This doesn't always play well with vgacon because vgacon_deinit() is only
doing its job when the last console it owns is closed (when switching from
vgacon to fbcon, that's usually *after* fbcon_init() has set the new
mode).

Depending on the hardware this can cause the wrong framebuffer location to
be scanned out (e.g.  reproduced on nv05 with the nouveau framebuffer
driver).
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Krzysztof Helt <krzysztof.h1@poczta.fm>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>