EEH may happen during a PCI driver probe. If the driver is trying to
access some register in a loop, the EEH code will try to print the
driver name. But the driver pointer in struct pci_dev is not set until
probe returns successfully.

Use a function to test if the device and the driver pointer is NULL
before accessing the driver's name.

Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

This fixes a hang that was observed during live partition migration.
Since stop_topology_update must not be called from an interrupt
context, call it earlier in the migration process. The hang observed
can be seen below:

WARNING: at kernel/timer.c:1011
Modules linked in: ip6t_LOG xt_tcpudp xt_pkttype ipt_LOG xt_limit ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_raw xt_NOTRACK ipt_REJECT xt_state iptable_raw iptable_filter ip6table_mangle nf_conntrack_netbios_ns nf_conntrack_broadcast nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables ip6table_filter ip6_tables x_tables ipv6 fuse loop ibmveth sg ext3 jbd mbcache raid456 async_raid6_recov async_pq raid6_pq async_xor xor async_memcpy async_tx raid10 raid1 raid0 scsi_dh_alua scsi_dh_rdac scsi_dh_hp_sw scsi_dh_emc dm_round_robin dm_multipath scsi_dh sd_mod crc_t10dif ibmvfc scsi_transport_fc scsi_tgt scsi_mod dm_snapshot dm_mod
NIP: c0000000000c52d8 LR: c00000000004be28 CTR: 0000000000000000
REGS: c00000005ffd77d0 TRAP: 0700   Not tainted  (3.2.0-git-00001-g07d106d)
MSR: 8000000000021032 <ME,CE,IR,DR>  CR: 48000084  XER: 00000001
CFAR: c00000000004be20
TASK = c00000005ec78860[0] 'swapper/3' THREAD: c00000005ec98000 CPU: 3
GPR00: 0000000000000001 c00000005ffd7a50 c000000000fbbc98 c000000000ec8340
GPR04: 00000000282a0020 0000000000000000 0000000000004000 0000000000000101
GPR08: 0000000000000012 c00000005ffd4000 0000000000000020 c000000000f3ba88
GPR12: 0000000000000000 c000000007f40900 0000000000000001 0000000000000004
GPR16: 0000000000000001 0000000000000000 0000000000000000 c000000001022310
GPR20: 0000000000000001 0000000000000000 0000000000200200 c000000001029e14
GPR24: 0000000000000000 0000000000000001 0000000000000040 c00000003f74bc80
GPR28: c00000003f74bc84 c000000000f38038 c000000000f16b58 c000000000ec8340
NIP [c0000000000c52d8] .del_timer_sync+0x28/0x60
LR [c00000000004be28] .stop_topology_update+0x20/0x38
Call Trace:
[c00000005ffd7a50] [c00000005ec78860] 0xc00000005ec78860 (unreliable)
[c00000005ffd7ad0] [c00000000004be28] .stop_topology_update+0x20/0x38
[c00000005ffd7b40] [c000000000028378] .__rtas_suspend_last_cpu+0x58/0x260
[c00000005ffd7bf0] [c0000000000fa230] .generic_smp_call_function_interrupt+0x160/0x358
[c00000005ffd7cf0] [c000000000036ec8] .smp_ipi_demux+0x88/0x100
[c00000005ffd7d80] [c00000000005c154] .icp_hv_ipi_action+0x5c/0x80
[c00000005ffd7e00] [c00000000012a088] .handle_irq_event_percpu+0x100/0x318
[c00000005ffd7f00] [c00000000012e774] .handle_percpu_irq+0x84/0xd0
[c00000005ffd7f90] [c000000000022ba8] .call_handle_irq+0x1c/0x2c
[c00000005ec9ba20] [c00000000001157c] .do_IRQ+0x22c/0x2a8
[c00000005ec9bae0] [c0000000000054bc] hardware_interrupt_entry+0x18/0x1c
Exception: 501 at .cpu_idle+0x194/0x2f8
    LR = .cpu_idle+0x194/0x2f8
[c00000005ec9bdd0] [c000000000017e58] .cpu_idle+0x188/0x2f8 (unreliable)
[c00000005ec9be90] [c00000000067ec18] .start_secondary+0x3e4/0x524
[c00000005ec9bf90] [c0000000000093e8] .start_secondary_prolog+0x10/0x14
Instruction dump:
ebe1fff8 4e800020 fbe1fff8 7c0802a6 f8010010 7c7f1b78 f821ff81 78290464
80090014 5400019e 7c0000d0 78000fe0 <0b000000> 4800000c 7c210b78 7c421378

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

We need to disable interrupts when taking the phb->lock. Otherwise
we could deadlock with pci_lock taken from an interrupt.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

We use __get_cpu_var() which triggers a false positive warning
in smp_processor_id() thinking interrupts are enabled (at this
point, they are soft-enabled but hard-disabled).

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

We call the cache_hwirq_map() function with a linux IRQ number
but it expects a HW irq number. This triggers a BUG on multic-chip
setups in addition to not doing the right thing.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

With this change, helpers such as instruction_pointer() et al, get defined
in the generic header in terms of GET_IP

Removed the unnecessary definition of profile_pc in !CONFIG_SMP case as
suggested by Mike Frysinger.

Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

It appears that on the Chroma card, the class code of the root
complex is still wrong even on DD2 or later chips. This could
be a firmware issue, but that breaks resource allocation so let's
unconditionally fix it up.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

Says Jens:

 "Time to push off some of the pending items.  I really wanted to wait
  until we had the regression nailed, but alas it's not quite there yet.
  But I'm very confident that it's "just" a missing expire on exit, so
  fix from Tejun should be fairly trivial.  I'm headed out for a week on
  the slopes.

  - Killing the barrier part of mtip32xx.  It doesn't really support
    barriers, and it doesn't need them (writes are fully ordered).

  - A few fixes from Dan Carpenter, preventing overflows of integer
    multiplication.

  - A fixup for loop, fixing a previous commit that didn't quite solve
    the partial read problem from Dave Young.

  - A bio integer overflow fix from Kent Overstreet.

  - Improvement/fix of the door "keep locked" part of the cdrom shared
    code from Paolo Benzini.

  - A few cfq fixes from Shaohua Li.

  - A fix for bsg sysfs warning when removing a file it did not create
    from Stanislaw Gruszka.

  - Two fixes for floppy from Vivek, preventing a crash.

  - A few block core fixes from Tejun.  One killing the over-optimized
    ioc exit path, cleaning that up nicely.  Two others fixing an oops
    on elevator switch, due to calling into the scheduler merge check
    code without holding the queue lock."

* 'for-linus' of git://git.kernel.dk/linux-block:
  block: fix lockdep warning on io_context release put_io_context()
  relay: prevent integer overflow in relay_open()
  loop: zero fill bio instead of return -EIO for partial read
  bio: don't overflow in bio_get_nr_vecs()
  floppy: Fix a crash during rmmod
  floppy: Cleanup disk->queue before caling put_disk() if add_disk() was never called
  cdrom: move shared static to cdrom_device_info
  bsg: fix sysfs link remove warning
  block: don't call elevator callbacks for plug merges
  block: separate out blk_rq_merge_ok() and blk_try_merge() from elevator functions
  mtip32xx: removed the irrelevant argument of mtip_hw_submit_io() and the unused member of struct driver_data
  block: strip out locking optimization in put_io_context()
  cdrom: use copy_to_user() without the underscores
  block: fix ioc locking warning
  block: fix NULL icq_cache reference
  block,cfq: change code order

11a3122f

 "block: strip out locking optimization in put_io_context()"
removed ioc_lock depth lockdep annoation along with locking
optimization; however, while recursing from put_io_context() is no
longer possible, ioc_release_fn() may still end up putting the last
reference of another ioc through elevator, which wlil grab ioc->lock
triggering spurious (as the ioc is always different one) A-A deadlock
warning.

As this can only happen one time from ioc_release_fn(), using non-zero
subclass from ioc_release_fn() is enough.  Use subclass 1.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

Quoth David:

1) GRO MAC header comparisons were ethernet specific, breaking other
   link types.  This required a multi-faceted fix to cure the originally
   noted case (Infiniband), because IPoIB was lying about it's actual
   hard header length.  Thanks to Eric Dumazet, Roland Dreier, and
   others.

2) Fix build failure when INET_UDP_DIAG is built in and ipv6 is modular.
   From Anisse Astier.

3) Off by ones and other bug fixes in netprio_cgroup from Neil Horman.

4) ipv4 TCP reset generation needs to respect any network interface
   binding from the socket, otherwise route lookups might give a
   different result than all the other segments received.  From Shawn
   Lu.

5) Fix unintended regression in ipv4 proxy ARP responses, from Thomas
   Graf.

6) Fix SKB under-allocation bug in sh_eth, from Yoshihiro Shimoda.

7) Revert skge PCI mapping changes that are causing crashes for some
   folks, from Stephen Hemminger.

8) IPV4 route lookups fill in the wildcarded fields of the given flow
   lookup key passed in, which is fine most of the time as this is
   exactly what the caller's want.  However there are a few cases that
   want to retain the original flow key values afterwards, so handle
   those cases properly.  Fix from Julian Anastasov.

9) IGB/IXGBE VF lookup bug fixes from Greg Rose.

10) Properly null terminate filename passed to ethtool flash device
    method, from Ben Hutchings.

11) S3 resume fix in via-velocity from David Lv.

12) Fix double SKB free during xmit failure in CAIF, from Dmitry
    Tarnyagin.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (72 commits)
  net: Don't proxy arp respond if iif == rt->dst.dev if private VLAN is disabled
  ipv4: Fix wrong order of ip_rt_get_source() and update iph->daddr.
  netprio_cgroup: fix wrong memory access when NETPRIO_CGROUP=m
  netprio_cgroup: don't allocate prio table when a device is registered
  netprio_cgroup: fix an off-by-one bug
  bna: fix error handling of bnad_get_flash_partition_by_offset()
  isdn: type bug in isdn_net_header()
  net: Make qdisc_skb_cb upper size bound explicit.
  ixgbe: ethtool: stats user buffer overrun
  ixgbe: dcb: up2tc mapping lost on disable/enable CEE DCB state
  ixgbe: do not update real num queues when netdev is going away
  ixgbe: Fix broken dependency on MAX_SKB_FRAGS being related to page size
  ixgbe: Fix case of Tx Hang in PF with 32 VFs
  ixgbe: fix vf lookup
  igb: fix vf lookup
  e1000: add dropped DMA receive enable back in for WoL
  gro: more generic L2 header check
  IPoIB: Stop lying about hard_header_len and use skb->cb to stash LL addresses
  zd1211rw: firmware needs duration_id set to zero for non-pspoll frames
  net: enable TC35815 for MIPS again
  ...

Commit 653241 (net: RFC3069, private VLAN proxy arp support) changed
the behavior of arp proxy to send arp replies back out on the interface
the request came in even if the private VLAN feature is disabled.

Previously we checked rt->dst.dev != skb->dev for in scenarios, when
proxy arp is enabled on for the netdevice and also when individual proxy
neighbour entries have been added.

This patch adds the check back for the pneigh_lookup() scenario.

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Jesper Dangaard Brouer <hawk@comx.dk>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch fix a bug which introduced by commit ac8a4810

 (ipv4: Save
nexthop address of LSRR/SSRR option to IPCB.).In that patch, we saved
the nexthop of SRR in ip_option->nexthop and update iph->daddr until
we get to ip_forward_options(), but we need to update it before
ip_rt_get_source(), otherwise we may get a wrong src.

Signed-off-by: Li Wei <lw@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When the netprio_cgroup module is not loaded, net_prio_subsys_id
is -1, and so sock_update_prioidx() accesses cgroup_subsys array
with negative index subsys[-1].

Make the code resembles cls_cgroup code, which is bug free.

Origionally-authored-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: "David S. Miller" <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

So we delay the allocation till the priority is set through cgroup,
and this makes skb_update_priority() faster when it's not set.

This also eliminates an off-by-one bug similar with the one fixed
in the previous patch.

Origionally-authored-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: "David S. Miller" <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

# mount -t cgroup xxx /mnt
  # mkdir /mnt/tmp
  # cat /mnt/tmp/net_prio.ifpriomap
  lo 0
  eth0 0
  virbr0 0
  # echo 'lo 999' > /mnt/tmp/net_prio.ifpriomap
  # cat /mnt/tmp/net_prio.ifpriomap
  lo 999
  eth0 0
  virbr0 4101267344

We got weired output, because we exceeded the boundary of the array.
We may even crash the kernel..

Origionally-authored-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: "David S. Miller" <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/i915: fixup interlaced bits clearing in PIPECONF on PCH_SPLIT (v2)
  drm/i915: no lvds quirk for AOpen MP45
  drm/i915: Force explicit bpp selection for intel_dp_link_required
  drm/radeon: do not continue after error from r600_ib_test
  drivers/gpu/drm/drm_ioc32.c: initialize all fields
  drm/i915: fixup interlaced bits clearing in PIPECONF on PCH_SPLIT
  drm/i915:: Disable FBC on SandyBridge

fix 1 mysterious divide error
fix 3 NULL dereference bugs in writeback tracing, on SD card removal w/o umount

* tag 'writeback-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/wfg/linux:
  writeback: fix dereferencing NULL bdi->dev on trace_writeback_queue
  lib: proportion: lower PROP_MAX_SHIFT to 32 on 64-bit kernel
  writeback: fix NULL bdi->dev in trace writeback_single_inode
  backing-dev: fix wakeup timer races with bdi_unregister()

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf: Fix double start/stop in x86_pmu_start()
  perf evsel: Fix an issue where perf report fails to show the proper percentage
  perf tools: Fix prefix matching for kernel maps
  perf tools: Fix perf stack to non executable on x86_64
  perf: Remove deprecated WARN_ON_ONCE()

Couple of regressions, couple of zero-day bugs, a minor enhancement.
Nothing really major.

* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (f75375s) Let f75375_update_device treat pwmX as a measured value
  hwmon: (f75375s) Fix bit shifting in f75375_write16
  hwmon: (f75375s) Fix reading of wrong register when initializing the F75387
  hwmon: (f75375s) Fix automatic pwm mode setting for F75373 & F75375
  hwmon: (w83627ehf) Remove duplicate code

Merge branch 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/keithp/linux into drm-fixes

* 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/keithp/linux:
  drm/i915: fixup interlaced bits clearing in PIPECONF on PCH_SPLIT (v2)
  drm/i915: no lvds quirk for AOpen MP45
  drm/i915: Force explicit bpp selection for intel_dp_link_required
  drm/i915: fixup interlaced bits clearing in PIPECONF on PCH_SPLIT
  drm/i915:: Disable FBC on SandyBridge

"subbuf_size" and "n_subbufs" come from the user and they need to be
capped to prevent an integer overflow.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Cc: stable@kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>

Serial/TTY fixes for the 3.3-rc3 tree

Just a few new device ids, omap serial driver regression fixes, and a
build fix for the 8250 driver movement.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

* tag 'tty-3.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  tty: serial: omap-serial: wakeup latency constraint is in microseconds, not milliseconds
  tty: serial: OMAP: block idle while the UART is transferring data in PIO mode
  tty: serial: OMAP: use a 1-byte RX FIFO threshold in PIO mode
  m32r: relocate drivers back out of 8250 dir
  tty: fix a build failure on sparc
  serial: samsung: Add support for EXYNOS5250
  serial: samsung: Add support for EXYNOS4212 and EXYNOS4412
  drivers/tty/vt/vt_ioctl.c: fix KDFONTOP 32bit compatibility layer

Staging tree patches for 3.3-rc3

Big things here is the deletion of the Android pmem driver, as it's
obsolete and no one uses it, the gma500 driver as it's already in the
drm portion of the kernel tree, and the pohmelfs filesystem as it's
obsolete and a rewritten version is being proposed for the fs/ section
of the kernel.

Other than that, a smattering of different bugfixes and regressions, and
some omap drm api merge fixups that were needed due to api changes in
the main portion of the drm tree, allowing this code to build properly
again.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

* tag 'staging-3.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging: (28 commits)
  staging: pohmelfs: remove drivers/staging/pohmelfs
  staging: android/ram_console: Don't build on arches w/o ioremap
  staging: r8712u: Use asynchronous firmware loading
  staging: usbip: fix to prevent potentially using uninitialized spinlock
  staging: r8712u: Fix problem when CONFIG_R8712_AP is set
  staging: tidspbridge: fix incorrect free to drv_datap
  staging: tidspbridge: fix bridge_open memory leaks
  staging: android: lowmemorykiller: Don't wait more than one second for a process to die
  MAINTAINERS: staging: iio: add iio information
  staging: zcache: fix serialization bug in zv stats
  staging: fix go7007-usb license
  Staging: android: binder: Fix crashes when sharing a binder file between processes
  Staging: android: Remove pmem driver
  Staging: asus_oled: fix NULL-ptr crash on unloading
  Staging: asus_oled: fix image processing
  Staging: android: binder: Don't call dump_stack in binder_vma_open
  staging: r8712u: Add new Sitecom UsB ID
  zcache: Set SWIZ_BITS to 8 to reduce tmem bucket lock contention.
  zcache: fix deadlock condition
  staging: drm/omap: fix locking issue
  ...

Driver core fixes for the 3.3-rc3 tree.

A few fixes for kobject warnings that have popped up in the cpu hotplug path,
and a regression fix for the speed of the hotplug memory code.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

* tag 'driver-core-3.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  driver-core: cpu: fix kobject warning when hotplugging a cpu
  ACPI: remove duplicated lines of merging problems with acpi_processor_add
  docbook: fix fatal errors in device-drivers docbook and add DMA Management section
  drivers/base/memory.c: fix memory_dev_init() long delay
  driver core: cpu: remove kernel warning when removing a cpu

Minor char-misc fixes for 3.3-rc3

Nothing big here, some Kconfig fixes for the MISC_DEVICES config option
that was being used incorrectly, and some other minor bug fixes.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

* tag 'char-misc-3.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  mmc: cb710 core: Add missing spin_lock_init for irq_lock of struct cb710_chip
  cs5535-mfgpt: don't call __init function from __devinit
  vmw_balloon: fix for a -Wuninitialized warning
  drivers: misc: Remove MISC_DEVICES config option
  c2port: fix build error for duramar2150 due to missing header.

USB fixes for 3.3-rc3

Here are a few minor USB fixes and a bunch of device id updates for the
USB drivers.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

* tag 'usb-3.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  USB: usbserial: add new PID number (0xa951) to the ftdi driver
  usb: ch9.h: usb_endpoint_maxp() uses __le16_to_cpu()
  usb: musb: fix a build error on mips
  uwb & wusb & usb wireless controllers: fix kconfig error & build errors
  usb: Skip PCI USB quirk handling for Netlogic XLP
  powerpc/usb: fix issue of CPU halt when missing USB PHY clock
  usb: otg: mv_otg: Add dependence
  usb: host: Distinguish Kconfig text for Freescale controllers
  USB: add new zte 3g-dongle's pid to option.c
  usb: ch9.h: usb_endpoint_maxp() uses __le16_to_cpu()
  USB: qcserial: don't enable autosuspend
  USB: qcserial: add several new serial devices
  usb: otg: mv_otg: Add dependence
  usb: gadget: zero: fix bug in loopback autoresume handling

The current error handling doesn't work because we flash_part is a u32
so the checks for negative error codes don't work.  I considered making
things signed but I don't know the hardware enough to say if that's a
problem.  Really, we don't use the error codes so just returning zero
for all problems is fine.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We use len to store the return value from eth_header().  eth_header()
can return -ETH_HLEN (-14).  We want to pass this back instead of
truncating it to 65522 and returning that.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Treat pwmX as a measured value, not as a (mostly static) limit value, so
that it is updated more frequently from the device register.

Signed-off-by: Nikolaus Schulz <mail@microschulz.de>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

Just like skb->cb[], so that qdisc_skb_cb can be encapsulated inside
of other data structures.

This is intended to be used by IPoIB so that it can remember
addressing information stored at hard_header_ops->create() time that
it can fetch when the packet gets to the transmit routine.

Signed-off-by: David S. Miller <davem@davemloft.net>

The receive FIFO wakeup latency estimate in the omap-serial driver is
three orders of magnitude too small.  This effectively prevents the
MPU from going to a low-power state when CONFIG_CPU_IDLE=y.  This is a
major power management regression and masks some other FIFO-related
bugs in the driver.

Fix by correcting the most egregious problem in the RX wakeup latency
estimate.  There are several other flaws in the estimator; these will
be fixed by a separate patch series intended for 3.4.

The difference in low-power states with this patch can be observed via
debugfs in pm_debug/count.

This estimate does not have any effect when CONFIG_CPU_IDLE=n.

Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: Alan Cox <alan@linux.intel.com>
Acked-by: Govindraj.R <govindraj.raja@ti.com>
Reviewed-by: Kevin Hilman <khilman@ti.com>
Tested-by: Kevin Hilman <khilman@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Prevent OMAP UARTs from going idle while they are still transferring
data in PIO mode.  This works around an oversight in the OMAP UART
hardware present in OMAP34xx and earlier: an idle UART won't send a
wakeup when the TX FIFO threshold is reached.  This causes long delays
during data transmission when the MPU powerdomain enters a low-power
mode.  The MPU interrupt controller is not able to respond to
interrupts when it's in a low-power state, so the TX buffer is not
refilled until another wakeup event occurs.

This fix changes the erratum i291 DMA idle workaround.  Rather than
toggling between force-idle and no-idle, it will toggle between
smart-idle and no-idle.  The important part of the workaround is the
no-idle part, so this shouldn't result in any change in behavior.

This fix should work on all OMAP UARTs.  Future patches intended for
the 3.4 merge window will make this workaround conditional on a
"feature" flag, and will use the OMAP36xx+ TX event wakeup support.

Thanks to Kevin Hilman <khilman@ti.com> for mentioning the erratum i291
workaround, which led to the development of this approach.

Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: Alan Cox <alan@linux.intel.com>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: Govindraj.R <govindraj.raja@ti.com>
Reviewed-by: Kevin Hilman <khilman@ti.com>
Tested-by: Kevin Hilman <khilman@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In the (default) PIO mode, use a one-byte RX FIFO threshold.  The OMAP
UART IP blocks do not appear to be capable of waking the system under
an RX timeout condition.  Since the previous RX FIFO threshold was 16
bytes, this meant that omap-serial.c did not become aware of any
received data until all those bytes arrived or until another UART
interrupt occurred.  This made the serial console and presumably other
serial applications (GPS, serial Bluetooth) unusable or extremely
slow.  A 1-byte RX FIFO threshold also allows the MPU to enter a
low-power consumption state while waiting for the FIFO to fill.

This can be verified using the serial console by comparing the
behavior when "0123456789abcde" is pasted in from another window, with
the behavior when "0123456789abcdef" is pasted in.  Since the former
string is less than sixteen bytes long, the string is not echoed for
some time, while the latter string is echoed immediately.

DMA operation is unaffected by this patch.

Thanks to Russell King - ARM Linux <linux@arm.linux.org.uk> for some
additional information on the standard behavior of the RX timeout
event, which was used to improve this commit description.

Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: Govindraj Raja <govindraj.r@ti.com>
Cc: Alan Cox <alan@linux.intel.com>
Cc: Russell King <linux@arm.linux.org.uk>
Reviewed-by: Kevin Hilman <khilman@ti.com>
Tested-by: Kevin Hilman <khilman@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

_parse_integer() does one or two division instructions (which are slow)
per digit parsed to perform the overflow check.

Furthermore, these are particularly expensive examples of division
instruction as the number of clock cycles required to complete them may
go up with the position of the most significant set bit in the dividend:

	if (*res > div_u64(ULLONG_MAX - val, base))

which is as maximal as possible.

Worse, on 32-bit arches, more than one of these division instructions
may be required per digit.

So, assuming we don't support a base of more than 16, skip the check if the
top nibble of the result is not set at this point.

Signed-off-by: David Howells <dhowells@redhat.com>
[ Changed it to not dereference the pointer all the time - even if the
  compiler can and does optimize it away, the code just looks cleaner.
  And edited the top nybble test slightly to make the code generated on
  x86-64 better in the loop - test against a hoisted constant instead of
  shifting and testing the result ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

If the number of tx/rx queues changes the ethtool ioctl
ETHTOOL_GSTATS may overrun the userspace buffer. This
occurs because the general practice in user space to
query stats is to issue a ETHTOOL_GSSET cmd to learn the
buffer size needed, allocate the buffer, then call
ETHTOOL_GSTIRNGS and ETHTOOL_GSTATS. If the number of
real_num_queues is changed or flow control attributes
are changed after ETHTOOL_GSSET but before the
ETHTOOL_GSTRINGS/ETHTOOL_GSTATS a user space buffer
overrun occurs.

To fix the overrun always return the max buffer size
needed from get_sset_count() then return all strings
and stats from get_strings()/get_ethtool_stats().

This _will_ change the output from the ioctl() call
which could break applications and script parsing in
theory. I believe these changes should not break existing
tools because the only changes will be more {tx|rx}_queues
and the {tx|rx}_pb_* stats will always be returned.
Existing scripts already need to handle changing number
of queues because this occurs today depending on system
and current features. The {tx|rx}_pb_* stats are at the
end of the output and should be handled by scripts today
regardless.

Finally get_ethtool_stats and get_strings are free-form
outputs tools parsing these outputs should be defensive
anyways. In the end these updates are better then
having a tool segfault because of a buffer overrun.

Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Users expect the up2tc mapping to be maintained across a DCB
enable/disable/enable transition. And since we maintain all
the other DCB attributes we should do this for up2tc mappings
as well just to be consistent. Also without this we break
user space applications that expect this to occur that
previously worked.

Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

If the netdev is already in NETREG_UNREGISTERING/_UNREGISTERED state, do not
update the real num tx queues. netdev_queue_update_kobjects() is already
called via remove_queue_kobjects() at NETREG_UNREGISTERING time. So, when
upper layer driver, e.g., FCoE protocol stack is monitoring the netdev
event of NETDEV_UNREGISTER and calls back to LLD ndo_fcoe_disable() to remove
extra queues allocated for FCoE, the associated txq sysfs kobjects are already
removed, and trying to update the real num queues would cause something like
below:

...
PID: 25138  TASK: ffff88021e64c440  CPU: 3   COMMAND: "kworker/3:3"
 #0 [ffff88021f007760] machine_kexec at ffffffff810226d9
 #1 [ffff88021f0077d0] crash_kexec at ffffffff81089d2d
 #2 [ffff88021f0078a0] oops_end at ffffffff813bca78
 #3 [ffff88021f0078d0] no_context at ffffffff81029e72
 #4 [ffff88021f007920] __bad_area_nosemaphore at ffffffff8102a155
 #5 [ffff88021f0079f0] bad_area_nosemaphore at ffffffff8102a23e
 #6 [ffff88021f007a00] do_page_fault at ffffffff813bf32e
 #7 [ffff88021f007b10] page_fault at ffffffff813bc045
    [exception RIP: sysfs_find_dirent+17]
    RIP: ffffffff81178611  RSP: ffff88021f007bc0  RFLAGS: 00010246
    RAX: ffff88021e64c440  RBX: ffffffff8156cc63  RCX: 0000000000000004
    RDX: ffffffff8156cc63  RSI: 0000000000000000  RDI: 0000000000000000
    RBP: ffff88021f007be0   R8: 0000000000000004   R9: 0000000000000008
    R10: ffffffff816fed00  R11: 0000000000000004  R12: 0000000000000000
    R13: ffffffff8156cc63  R14: 0000000000000000  R15: ffff8802222a0000
    ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018
 #8 [ffff88021f007be8] sysfs_get_dirent at ffffffff81178c07
 #9 [ffff88021f007c18] sysfs_remove_group at ffffffff8117ac27
#10 [ffff88021f007c48] netdev_queue_update_kobjects at ffffffff813178f9
#11 [ffff88021f007c88] netif_set_real_num_tx_queues at ffffffff81303e38
#12 [ffff88021f007cc8] ixgbe_set_num_queues at ffffffffa0249763 [ixgbe]
#13 [ffff88021f007cf8] ixgbe_init_interrupt_scheme at ffffffffa024ea89 [ixgbe]
#14 [ffff88021f007d48] ixgbe_fcoe_disable at ffffffffa0267113 [ixgbe]
#15 [ffff88021f007d68] vlan_dev_fcoe_disable at ffffffffa014fef5 [8021q]
#16 [ffff88021f007d78] fcoe_interface_cleanup at ffffffffa02b7dfd [fcoe]
#17 [ffff88021f007df8] fcoe_destroy_work at ffffffffa02b7f08 [fcoe]
#18 [ffff88021f007e18] process_one_work at ffffffff8105d7ca
#19 [ffff88021f007e68] worker_thread at ffffffff81060513
#20 [ffff88021f007ee8] kthread at ffffffff810648b6
#21 [ffff88021f007f48] kernel_thread_helper at ffffffff813c40f4

Signed-off-by: Yi Zou <yi.zou@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch fixes an issue in which RSC will generate corrupted frames when
PAGE_SIZE is larger than 8K.  Specifically it looks like that in 2.6.39 a
change was made so that GRO would always have at least 16 frags available
for coalescing, but the ixgbe RSC logic was not updated.  As such the RSC
feature would generate a frame larger than 64K and then overflow the value
in the IP length field.

To correct that I am now basing things on the PAGE_SIZE.

Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

A check for the number of VFs allocated should have used a greater than
equal operator instead of just greater than.  This caused allocation of
exactly 32 VFs to not enable the PF transmit and receive enables.

Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Robert E Garrett <robertX.e.garrett@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Recent addition of code to find already allocated VFs failed to take
account that systems with 2 or more multi-port SR-IOV capable controllers
might have already enabled VFs.  Make sure that the VFs the function is
finding are actually subordinate to the particular instance of the adapter
that is looking for them and not subordinate to some device that has
previously enabled SR-IOV.

This bug exists in 3.2 stable as well as 3.3 release candidates.

CC: stable@vger.kernel.org
Reported-by: David Ahern <daahern@cisco.com>
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Robert E Garrett <robertX.e.garrett@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>