deliverable/linux.git
9 years agosfc: add support for skb->xmit_more
Edward Cree [Fri, 17 Oct 2014 14:32:25 +0000 (15:32 +0100)] 
sfc: add support for skb->xmit_more

Don't ring the doorbell, and don't do PIO.  This will also prevent
 TX Push, because there will be more than one buffer waiting when
 the doorbell is rung.

Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agor8152: return -EBUSY for runtime suspend
hayeswang [Fri, 17 Oct 2014 08:55:08 +0000 (16:55 +0800)] 
r8152: return -EBUSY for runtime suspend

Remove calling cancel_delayed_work_sync() for runtime suspend,
because it would cause dead lock. Instead, return -EBUSY to
avoid the device enters suspending if the net is running and
the delayed work is pending or running. The delayed work would
try to wake up the device later, so the suspending is not
necessary.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: fix a potential use after free in fou.c
Li RongQing [Fri, 17 Oct 2014 08:53:47 +0000 (16:53 +0800)] 
ipv4: fix a potential use after free in fou.c

pskb_may_pull() maybe change skb->data and make uh pointer oboslete,
so reload uh and guehdr

Fixes: 37dd0247 ("gue: Receive side for Generic UDP Encapsulation")
Cc: Tom Herbert <therbert@google.com>
Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: fix a potential use after free in ip_tunnel_core.c
Li RongQing [Fri, 17 Oct 2014 08:53:23 +0000 (16:53 +0800)] 
ipv4: fix a potential use after free in ip_tunnel_core.c

pskb_may_pull() maybe change skb->data and make eth pointer oboslete,
so set eth after pskb_may_pull()

Fixes:3d7b46cd("ip_tunnel: push generic protocol handling to ip_tunnel module")
Cc: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agohyperv: Add handling of IP header with option field in netvsc_set_hash()
Haiyang Zhang [Thu, 16 Oct 2014 21:47:58 +0000 (14:47 -0700)] 
hyperv: Add handling of IP header with option field in netvsc_set_hash()

In case that the IP header has optional field at the end, this patch will
get the port numbers after that field, and compute the hash. The general
parser skb_flow_dissect() is used here.

Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Reviewed-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years ago[CIFS] Remove obsolete comment
Steve French [Fri, 17 Oct 2014 22:17:12 +0000 (17:17 -0500)] 
[CIFS] Remove obsolete comment

Signed-off-by: Steven French <smfrench@gmail.com>
9 years agoopenvswitch: Create right mask with disabled megaflows
Pravin B Shelar [Fri, 17 Oct 2014 04:55:45 +0000 (21:55 -0700)] 
openvswitch: Create right mask with disabled megaflows

If megaflows are disabled, the userspace does not send the netlink attribute
OVS_FLOW_ATTR_MASK, and the kernel must create an exact match mask.

sw_flow_mask_set() sets every bytes (in 'range') of the mask to 0xff, even the
bytes that represent padding for struct sw_flow, or the bytes that represent
fields that may not be set during ovs_flow_extract().
This is a problem, because when we extract a flow from a packet,
we do not memset() anymore the struct sw_flow to 0.

This commit gets rid of sw_flow_mask_set() and introduces mask_set_nlattr(),
which operates on the netlink attributes rather than on the mask key. Using
this approach we are sure that only the bytes that the user provided in the
flow are matched.

Also, if the parse_flow_mask_nlattrs() for the mask ENCAP attribute fails, we
now return with an error.

This bug is introduced by commit 0714812134d7dcadeb7ecfbfeb18788aa7e1eaac
("openvswitch: Eliminate memset() from flow_extract").

Reported-by: Alex Wang <alexw@nicira.com>
Signed-off-by: Daniele Di Proietto <ddiproietto@vmware.com>
Signed-off-by: Andy Zhou <azhou@nicira.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agovxlan: fix a free after use
Li RongQing [Fri, 17 Oct 2014 06:06:16 +0000 (14:06 +0800)] 
vxlan: fix a free after use

pskb_may_pull maybe change skb->data and make eth pointer oboslete,
so eth needs to reload

Fixes: 91269e390d062 ("vxlan: using pskb_may_pull as early as possible")
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoopenvswitch: fix a use after free
Li RongQing [Fri, 17 Oct 2014 06:03:08 +0000 (14:03 +0800)] 
openvswitch: fix a use after free

pskb_may_pull() called by arphdr_ok can change skb->data, so put the arp
setting after arphdr_ok to avoid the use the freed memory

Fixes: 0714812134d7d ("openvswitch: Eliminate memset() from flow_extract.")
Cc: Jesse Gross <jesse@nicira.com>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: dst_entry leak in ip_send_unicast_reply()
Vasily Averin [Wed, 15 Oct 2014 12:24:02 +0000 (16:24 +0400)] 
ipv4: dst_entry leak in ip_send_unicast_reply()

ip_setup_cork() called inside ip_append_data() steals dst entry from rt to cork
and in case errors in __ip_append_data() nobody frees stolen dst entry

Fixes: 2e77d89b2fa8 ("net: avoid a pair of dst_hold()/dst_release() in ip_append_data()")
Signed-off-by: Vasily Averin <vvs@parallels.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoperf script: Add period as a default output column
Jiri Olsa [Mon, 25 Aug 2014 14:45:43 +0000 (16:45 +0200)] 
perf script: Add period as a default output column

Adding period as a default output column in script command fo hardware,
software and raw events.

If PERF_SAMPLE_PERIOD sample type is defined in perf.data, following
will be displayed in perf script output:

  $ perf script
              ls  8034 57477.887209:     250000 task-clock:  ffffffff81361d72 memset ([kernel.kallsyms])
              ls  8034 57477.887464:     250000 task-clock:  ffffffff816f6d92 _raw_spin_unlock_irqrestore ([kernel.kallsyms])
              ls  8034 57477.887708:     250000 task-clock:  ffffffff811a94f0 do_munmap ([kernel.kallsyms])
              ls  8034 57477.887959:     250000 task-clock:        34080916c6 get_next_seq (/usr/lib64/libc-2.17.so)
              ls  8034 57477.888208:     250000 task-clock:        3408079230 _IO_doallocbuf (/usr/lib64/libc-2.17.so)
              ls  8034 57477.888717:     250000 task-clock:  ffffffff814242c8 n_tty_write ([kernel.kallsyms])
              ls  8034 57477.889285:     250000 task-clock:        3408076402 fwrite_unlocked (/usr/lib64/libc-2.17.so)

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: "Jen-Cheng(Tommy) Huang" <tommy24@gatech.edu>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jen-Cheng(Tommy) Huang <tommy24@gatech.edu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1408977943-16594-10-git-send-email-jolsa@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf script: Add period data column
Jiri Olsa [Mon, 25 Aug 2014 14:45:42 +0000 (16:45 +0200)] 
perf script: Add period data column

Adding period data column to be displayed in perf script.  It's possible
to get period values using -f option, like:

  $ perf script -f comm,tid,time,period,ip,sym,dso
          :26019 26019 52414.329088:       3707  ffffffff8105443a native_write_msr_safe ([kernel.kallsyms])
          :26019 26019 52414.329088:         44  ffffffff8105443a native_write_msr_safe ([kernel.kallsyms])
          :26019 26019 52414.329093:       1987  ffffffff8105443a native_write_msr_safe ([kernel.kallsyms])
          :26019 26019 52414.329093:          6  ffffffff8105443a native_write_msr_safe ([kernel.kallsyms])
              ls 26019 52414.329442:     537558        3407c0639c _dl_map_object_from_fd (/usr/lib64/ld-2.17.so)
              ls 26019 52414.329442:       2099        3407c0639c _dl_map_object_from_fd (/usr/lib64/ld-2.17.so)
              ls 26019 52414.330181:    1242100        34080917bb get_next_seq (/usr/lib64/libc-2.17.so)
              ls 26019 52414.330181:       3774        34080917bb get_next_seq (/usr/lib64/libc-2.17.so)
              ls 26019 52414.331427:    1083662  ffffffff810c7dc2 update_curr ([kernel.kallsyms])
              ls 26019 52414.331427:        360  ffffffff810c7dc2 update_curr ([kernel.kallsyms])

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: David Ahern <dsahern@gmail.com>
Cc: "Jen-Cheng(Tommy) Huang" <tommy24@gatech.edu>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jen-Cheng(Tommy) Huang <tommy24@gatech.edu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1408977943-16594-9-git-send-email-jolsa@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoipv4: clean up cookie_v4_check()
Cong Wang [Wed, 15 Oct 2014 21:33:22 +0000 (14:33 -0700)] 
ipv4: clean up cookie_v4_check()

We can retrieve opt from skb, no need to pass it as a parameter.
And opt should always be non-NULL, no need to check.

Cc: Krzysztof Kolasa <kkolasa@winsoft.pl>
Cc: Eric Dumazet <edumazet@google.com>
Tested-by: Krzysztof Kolasa <kkolasa@winsoft.pl>
Signed-off-by: Cong Wang <cwang@twopensource.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: share tcp_v4_save_options() with cookie_v4_check()
Cong Wang [Wed, 15 Oct 2014 21:33:21 +0000 (14:33 -0700)] 
ipv4: share tcp_v4_save_options() with cookie_v4_check()

cookie_v4_check() allocates ip_options_rcu in the same way
with tcp_v4_save_options(), we can just make it a helper function.

Cc: Krzysztof Kolasa <kkolasa@winsoft.pl>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Cong Wang <cwang@twopensource.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: call __ip_options_echo() in cookie_v4_check()
Cong Wang [Wed, 15 Oct 2014 21:33:20 +0000 (14:33 -0700)] 
ipv4: call __ip_options_echo() in cookie_v4_check()

commit 971f10eca186cab238c49da ("tcp: better TCP_SKB_CB layout to reduce cache line misses")
missed that cookie_v4_check() still calls ip_options_echo() which uses
IPCB(). It should use TCPCB() at TCP layer, so call __ip_options_echo()
instead.

Fixes: commit 971f10eca186cab238c49da ("tcp: better TCP_SKB_CB layout to reduce cache line misses")
Cc: Krzysztof Kolasa <kkolasa@winsoft.pl>
Cc: Eric Dumazet <edumazet@google.com>
Reported-by: Krzysztof Kolasa <kkolasa@winsoft.pl>
Tested-by: Krzysztof Kolasa <kkolasa@winsoft.pl>
Signed-off-by: Cong Wang <cwang@twopensource.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoatm: simplify lanai.c by using module_pci_driver
Michael Opdenacker [Wed, 15 Oct 2014 07:45:50 +0000 (09:45 +0200)] 
atm: simplify lanai.c by using module_pci_driver

This simplifies the lanai.c driver by using
the module_pci_driver() macro, at the expense
of losing only debugging messages.

Signed-off-by: Michael Opdenacker <michael.opdenacker@free-electrons.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoperf evsel: No need to drag util/cgroup.h
Arnaldo Carvalho de Melo [Fri, 17 Oct 2014 15:17:40 +0000 (12:17 -0300)] 
perf evsel: No need to drag util/cgroup.h

The only thing we need is a forward declaration for 'struct cgroup_sel',
that is inside 'struct perf_evsel'.

Include cgroup.h instead on the tools that support cgroups.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-b7kuymbgf0zxi5viyjjtu5hk@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf evlist: Add missing 'struct option' forward declaration
Arnaldo Carvalho de Melo [Fri, 17 Oct 2014 15:16:00 +0000 (12:16 -0300)] 
perf evlist: Add missing 'struct option' forward declaration

It was being found, by chance, because evsel.h needlessly includes
util/cgroup.h, which will be sorted out in a following patch.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-xsvxr747wkkpg1ay9dramorr@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf evsel: Move exit stuff from __delete to __exit
Arnaldo Carvalho de Melo [Thu, 16 Oct 2014 16:25:01 +0000 (13:25 -0300)] 
perf evsel: Move exit stuff from __delete to __exit

So that when an evsel is embedded into other struct it can free up
resources calling perf_evsel__exit().

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-n1w68pfe9m2vkhm4sqs8y1en@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agontb: Adding split BAR support for Haswell platforms
Dave Jiang [Thu, 28 Aug 2014 20:53:23 +0000 (13:53 -0700)] 
ntb: Adding split BAR support for Haswell platforms

On the Haswell platform, a split BAR option to allow creation of 2
32bit BARs (4 and 5) from the 64bit BAR 4. Adding support for this
new option.

Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
9 years agontb: use errata flag set via DID to implement workaround
Dave Jiang [Thu, 28 Aug 2014 20:53:18 +0000 (13:53 -0700)] 
ntb: use errata flag set via DID to implement workaround

Instead of using a module parameter, we should detect the errata via
PCI DID and then set an appropriate flag. This will be used for additional
errata later on.

Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
9 years agontb: conslidate reading of PPD to move platform detection earlier
Dave Jiang [Thu, 28 Aug 2014 20:53:13 +0000 (13:53 -0700)] 
ntb: conslidate reading of PPD to move platform detection earlier

To simplify some of the platform detection code. Move the platform detection
to a function to be called earlier.

Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
9 years agontb: move platform detection to separate function
Dave Jiang [Thu, 28 Aug 2014 20:53:07 +0000 (13:53 -0700)] 
ntb: move platform detection to separate function

Move the platform detection function to separate functions to allow
easier maintenence.

Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
9 years agoNTB: debugfs device entry
Jon Mason [Mon, 7 Apr 2014 17:55:47 +0000 (10:55 -0700)] 
NTB: debugfs device entry

Create a debugfs entry for the NTB device to log the basic device info,
as well as display the error count on a number of registers.

Signed-off-by: Jon Mason <jon.mason@intel.com>
9 years agoRevert "Btrfs: race free update of commit root for ro snapshots"
Chris Mason [Wed, 15 Oct 2014 20:50:56 +0000 (13:50 -0700)] 
Revert "Btrfs: race free update of commit root for ro snapshots"

This reverts commit 9c3b306e1c9e6be4be09e99a8fe2227d1005effc.

Switching only one commit root during a transaction is wrong because it
leads the fs into an inconsistent state. All commit roots should be
switched at once, at transaction commit time, otherwise backref walking
can often miss important references that were only accessible through
the old commit root.  Plus, the root item for the snapshot's root wasn't
getting updated and preventing the next transaction commit to do it.

This made several users get into random corruption issues after creation
of readonly snapshots.

A regression test for xfstests will follow soon.

Cc: stable@vger.kernel.org # 3.17
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
9 years agokprobes/x86: Remove stale ARCH_SUPPORTS_KPROBES_ON_FTRACE define
Anton Blanchard [Thu, 16 Oct 2014 20:29:50 +0000 (07:29 +1100)] 
kprobes/x86: Remove stale ARCH_SUPPORTS_KPROBES_ON_FTRACE define

Commit e7dbfe349d12 ("kprobes/x86: Move ftrace-based kprobe code
into kprobes-ftrace.c") switched from using
ARCH_SUPPORTS_KPROBES_ON_FTRACE to CONFIG_KPROBES_ON_FTRACE but
missed removing the define.

Signed-off-by: Anton Blanchard <anton@samba.org>
Cc: masami.hiramatsu.pt@hitachi.com
Cc: ananth@in.ibm.com
Cc: a.p.zijlstra@chello.nl
Cc: fweisbec@gmail.com
Cc: rostedt@goodmis.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agoInput: xpad - add Thrustmaster as Xbox 360 controller vendor
Tommi Rantala [Thu, 16 Oct 2014 21:02:07 +0000 (14:02 -0700)] 
Input: xpad - add Thrustmaster as Xbox 360 controller vendor

Add Thrustmaster as Xbox 360 controller vendor. This is required for
example to make the GP XID (044f:b326) gamepad work.

Signed-off-by: Tommi Rantala <tt.rantala@gmail.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
9 years agoInput: xpad - add USB ID for Thrustmaster Ferrari 458 Racing Wheel
Tommi Rantala [Thu, 16 Oct 2014 21:01:43 +0000 (14:01 -0700)] 
Input: xpad - add USB ID for Thrustmaster Ferrari 458 Racing Wheel

Add the USB ID for the Xbox 360 Thrustmaster Ferrari 458 Racing Wheel.

Signed-off-by: Tommi Rantala <tt.rantala@gmail.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
9 years agoInput: max77693-haptic - fix state check in imax77693_haptic_disable()
Jaewon Kim [Thu, 16 Oct 2014 20:56:52 +0000 (13:56 -0700)] 
Input: max77693-haptic - fix state check in imax77693_haptic_disable()

The check to see whether the device is already disabled in
max77693_haptic_disable() was inversed, this change corrects it.

Signed-off-by: Jaewon Kim <jaewon02.kim@samsung.com>
Reviewed-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
9 years agoInput: xen-kbdfront - free grant table entry in xenkbd_disconnect_backend
Chang Huaixin [Thu, 16 Oct 2014 20:33:24 +0000 (13:33 -0700)] 
Input: xen-kbdfront - free grant table entry in xenkbd_disconnect_backend

xenkbd_disconnect_backend doesn't free grant table entry.  This bug affects
live migration.

xenkbd_disconnect_backend uses gnttab_end_foreign_access_ref to handle
grant table entry which doesn't really free an entry.

Thus every time we do xenkbd_resume, grant table entry increses by one. As
an grant table entry occupies 8 bytes, an grant table page has at most 512
entries. Every 512 times we do xenkdb_resume, grant table pages increses by
one.

After around 3500 times of live migration, grant table pages will increase
by 7, causing too many pages to populate and hitting max_pages limit when
assigning pages.Thus assign_pages will fail, so will live migration.

Signed-off-by: Chang Huaixin <huaixin.chx@alibaba-inc.com>
Acked-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
9 years agoCheck minimum response length on query_network_interface
Steve French [Thu, 16 Oct 2014 20:13:14 +0000 (15:13 -0500)] 
Check minimum response length on query_network_interface

Acked-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Steve French <smfrench@gmail.com>
9 years agoWorkaround Mac server problem
Steve French [Mon, 6 Oct 2014 06:01:03 +0000 (01:01 -0500)] 
Workaround Mac server problem

Mac server returns that they support CIFS Unix Extensions but
doesn't actually support QUERY_FILE_UNIX_BASIC so mount fails.

Workaround this problem by disabling use of Unix CIFS protocol
extensions if server returns an EOPNOTSUPP error on
QUERY_FILE_UNIX_BASIC during mount.

Signed-off-by: Steve French <smfrench@gmail.com>
9 years agoRemap reserved posix characters by default (part 3/3)
Steve French [Sat, 27 Sep 2014 07:19:01 +0000 (02:19 -0500)] 
Remap reserved posix characters by default (part 3/3)

This is a bigger patch, but its size is mostly due to
a single change for how we check for remapping illegal characters
in file names - a lot of repeated, small changes to
the way callers request converting file names.

The final patch in the series does the following:

1) changes default behavior for cifs to be more intuitive.
Currently we do not map by default to seven reserved characters,
ie those valid in POSIX but not in NTFS/CIFS/SMB3/Windows,
unless a mount option (mapchars) is specified.  Change this
to by default always map and map using the SFM maping
(like the Mac uses) unless the server negotiates the CIFS Unix
Extensions (like Samba does when mounting with the cifs protocol)
when the remapping of the characters is unnecessary.  This should
help SMB3 mounts in particular since Samba will likely be
able to implement this mapping with its new "vfs_fruit" module
as it will be doing for the Mac.
2) if the user specifies the existing "mapchars" mount option then
use the "SFU" (Microsoft Services for Unix, SUA) style mapping of
the seven characters instead.
3) if the user specifies "nomapposix" then disable SFM/MAC style mapping
(so no character remapping would be used unless the user specifies
"mapchars" on mount as well, as above).
4) change all the places in the code that check for the superblock
flag on the mount which is set by mapchars and passed in on all
path based operation and change it to use a small function call
instead to set the mapping type properly (and check for the
mapping type in the cifs unicode functions)

Signed-off-by: Steve French <smfrench@gmail.com>
9 years agoAllow conversion of characters in Mac remap range (part 2)
Steve French [Thu, 25 Sep 2014 19:01:34 +0000 (14:01 -0500)] 
Allow conversion of characters in Mac remap range (part 2)

The previous patch allowed remapping reserved characters from directory
listenings, this patch adds conversion the other direction, allowing
opening of files with any of the seven reserved characters.

Signed-off-by: Steve French <smfrench@gmail.com>
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
9 years agoAllow conversion of characters in Mac remap range. Part 1
Steve French [Thu, 25 Sep 2014 18:20:05 +0000 (13:20 -0500)] 
Allow conversion of characters in Mac remap range. Part 1

This allows directory listings to Mac to display filenames
correctly which have been created with illegal (to Windows)
characters in their filename. It does not allow
converting the other direction yet ie opening files with
these characters (followon patch).

There are seven reserved characters that need to be remapped when
mounting to Windows, Mac (or any server without Unix Extensions) which
are valid in POSIX but not in the other OS.

: \ < > ? * |

We used the normal UCS-2 remap range for this in order to convert this
to/from UTF8 as did Windows Services for Unix (basically add 0xF000 to
any of the 7 reserved characters), at least when the "mapchars" mount
option was specified.

Mac used a very slightly different "Services for Mac" remap range
0xF021 through 0xF027.  The attached patch allows cifs.ko (the kernel
client) to read directories on macs containing files with these
characters and display their names properly.  In theory this even
might be useful on mounts to Samba when the vfs_catia or new
"vfs_fruit" module is loaded.

Currently the 7 reserved characters look very strange in directory
listings from cifs.ko to Mac server.  This patch allows these file
name characters to be read (requires specifying mapchars on mount).

Two additional changes are needed:
1) Make it more automatic: a way of detecting enough info so that
we know to try to always remap these characters or not. Various
have suggested that the SFM approach be made the default when
the server does not support POSIX Unix extensions (cifs mounts
to Samba for example) so need to make SFM remapping the default
unless mapchars (SFU style mapping) specified on mount or no
mapping explicitly requested or no mapping needed (cifs mounts to Samba).

2) Adding a patch to map the characters the other direction
(ie UTF-8 to UCS-2 on open).  This patch does it for translating
readdir entries (ie UCS-2 to UTF-8)

Signed-off-by: Steve French <smfrench@gmail.com>
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
9 years agomfsymlinks support for SMB2.1/SMB3. Part 2 query symlink
Steve French [Tue, 16 Sep 2014 12:18:19 +0000 (07:18 -0500)] 
mfsymlinks support for SMB2.1/SMB3. Part 2 query symlink

Adds support on SMB2.1 and SMB3 mounts for emulation of symlinks
via the "Minshall/French" symlink format already used for cifs
mounts when mfsymlinks mount option is used (and also used by Apple).
  http://wiki.samba.org/index.php/UNIX_Extensions#Minshall.2BFrench_symlinks
This second patch adds support to query them (recognize them as symlinks
and read them).  Third version of patch makes minor corrections
to error handling.

Signed-off-by: Steve French <smfrench@gmail.com>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
9 years agoAdd mfsymlinks support for SMB2.1/SMB3. Part 1 create symlink
Steve French [Mon, 15 Sep 2014 09:49:28 +0000 (04:49 -0500)] 
Add mfsymlinks support for SMB2.1/SMB3. Part 1 create symlink

Adds support on SMB2.1 and SMB3 mounts for emulation of symlinks
via the "Minshall/French" symlink format already used for cifs
mounts when mfsymlinks mount option is used (and also used by Apple).
http://wiki.samba.org/index.php/UNIX_Extensions#Minshall.2BFrench_symlinks
This first patch adds support to create them.  The next patch will
add support for recognizing them and reading them.  Although CIFS/SMB3
have other types of symlinks, in the many use cases they aren't
practical (e.g. either require cifs only mounts with unix extensions
to Samba, or require the user to be Administrator to Windows for SMB3).
This also helps enable running additional xfstests over SMB3 (since some
xfstests directly or indirectly require symlink support).

Signed-off-by: Steve French <smfrench@gmail.com>
CC: Stefan Metzmacher <metze@samba.org>
9 years agoAllow mknod and mkfifo on SMB2/SMB3 mounts
Steve French [Mon, 22 Sep 2014 10:13:55 +0000 (05:13 -0500)] 
Allow mknod and mkfifo on SMB2/SMB3 mounts

The "sfu" mount option did not work on SMB2/SMB3 mounts.
With these changes when the "sfu" mount option is passed in
on an smb2/smb2.1/smb3 mount the client can emulate (and
recognize) fifo and device (character and device files).

In addition the "sfu" mount option should not conflict
with "mfsymlinks" (symlink emulation) as we will never
create "sfu" style symlinks, but using "sfu" mount option
will allow us to recognize existing symlinks, created with
Microsoft "Services for Unix" (SFU and SUA).

To enable the "sfu" mount option for SMB2/SMB3 the calling
syntax of the generic cifs/smb2/smb3 sync_read and sync_write
protocol dependent function needed to be changed (we
don't have a file struct in all cases), but this actually
ended up simplifying the code a little.

Signed-off-by: Steve French <smfrench@gmail.com>
9 years agoadd defines for two new file attributes
Steve French [Wed, 24 Sep 2014 00:25:42 +0000 (19:25 -0500)] 
add defines for two new file attributes

Signed-off-by: Steve French <smfrench@gmail.com>
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
9 years agonetlink: fix description of portid
Nicolas Dichtel [Thu, 16 Oct 2014 13:47:51 +0000 (15:47 +0200)] 
netlink: fix description of portid

Avoid confusion between pid and portid.

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net
David S. Miller [Thu, 16 Oct 2014 18:42:51 +0000 (14:42 -0400)] 
Merge branch 'master' of git://git./linux/kernel/git/jkirsher/net

Jeff Kirsher says:

====================
Intel Wired LAN Driver Updates 2014-10-16

This series contains updates to fm10k and ixgbe.

Matthew provides two fixes for fm10k, first sets the flag to fetch the
host state before kicking off the service task that reads the host
state when bringing the interface up.  The second makes sure that we
release the mailbox lock after detecting an error and before we return
the error code.

Andy Zhou provides a compile fix for fm10k, when the driver is compiled
into the kernel and the VXLAN driver is compiled as a module.

Emil provides a fix for ixgbe to prevent against a panic by trying
to dereference a NULL pointer in ixgbe_ndo_set_vf_spoofchk().
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoNTFS: Bump version to 2.1.31.
Anton Altaparmakov [Thu, 16 Oct 2014 11:53:35 +0000 (12:53 +0100)] 
NTFS: Bump version to 2.1.31.

Signed-off-by: Anton Altaparmakov <anton@tuxera.com>
9 years agoNTFS: Add bmap address space operation needed for FIBMAP ioctl.
Anton Altaparmakov [Thu, 16 Oct 2014 11:50:52 +0000 (12:50 +0100)] 
NTFS: Add bmap address space operation needed for FIBMAP ioctl.

Signed-off-by: Anton Altaparmakov <anton@tuxera.com>
9 years agoNTFS: Remove changelog from Documentation/filesystems/ntfs.txt.
Anton Altaparmakov [Thu, 16 Oct 2014 11:43:57 +0000 (12:43 +0100)] 
NTFS: Remove changelog from Documentation/filesystems/ntfs.txt.

Changelog is in git history, no need to have a copy in the documentation.

Signed-off-by: Anton Altaparmakov <anton@tuxera.com>
9 years agoNTFS: Split ntfs_aops into ntfs_normal_aops and ntfs_compressed_aops
Anton Altaparmakov [Thu, 16 Oct 2014 11:28:03 +0000 (12:28 +0100)] 
NTFS: Split ntfs_aops into ntfs_normal_aops and ntfs_compressed_aops
in preparation for them diverging.

Signed-off-by: Anton Altaparmakov <anton@tuxera.com>
9 years agoMerge branch 'stable-3.18' of git://git.infradead.org/users/pcmoore/selinux into...
James Morris [Thu, 16 Oct 2014 10:04:18 +0000 (21:04 +1100)] 
Merge branch 'stable-3.18' of git://git.infradead.org/users/pcmoore/selinux into for-linus2

9 years agoixgbe: check for vfs outside of sriov_num_vfs before dereference
Emil Tantilov [Thu, 16 Oct 2014 15:49:02 +0000 (15:49 +0000)] 
ixgbe: check for vfs outside of sriov_num_vfs before dereference

The check for vfinfo is not sufficient because it does not protect
against specifying vf that is outside of sriov_num_vfs range.
All of the ndo functions have a check for it except for
ixgbevf_ndo_set_spoofcheck().

The following patch is all we need to protect against this panic:

ip link set p96p1 vf 0 spoofchk off
BUG: unable to handle kernel NULL pointer dereference at 0000000000000052
IP: [<ffffffffa044a1c1>]
ixgbe_ndo_set_vf_spoofchk+0x51/0x150 [ixgbe]

Reported-by: Thierry Herbelot <thierry.herbelot@6wind.com>
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Acked-by: Thierry Herbelot <thierry.herbelot@6wind.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agofm10k: Add CONFIG_FM10K_VXLAN configuration option
Andy Zhou [Sat, 4 Oct 2014 06:19:11 +0000 (06:19 +0000)] 
fm10k: Add CONFIG_FM10K_VXLAN configuration option

Compiling with CONFIG_FM10K=y and VXLAN=m resulting in linking error:

   drivers/built-in.o: In function `fm10k_open':
   (.text+0x1f9d7a): undefined reference to `vxlan_get_rx_port'
   make: *** [vmlinux] Error 1

The fix follows the same strategy as I40E.

Signed-off-by: Andy Zhou <azhou@nicira.com>
Acked-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoarm/arm64: KVM: Fix BE accesses to GICv2 EISR and ELRSR regs
Christoffer Dall [Sun, 28 Sep 2014 14:04:26 +0000 (16:04 +0200)] 
arm/arm64: KVM: Fix BE accesses to GICv2 EISR and ELRSR regs

The EIRSR and ELRSR registers are 32-bit registers on GICv2, and we
store these as an array of two such registers on the vgic vcpu struct.
However, we access them as a single 64-bit value or as a bitmap pointer
in the generic vgic code, which breaks BE support.

Instead, store them as u64 values on the vgic structure and do the
word-swapping in the assembly code, which already handles the byte order
for BE systems.

Tested-by: Victor Kamensky <victor.kamensky@linaro.org>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Christoffer Dall <christoffer.dall@linaro.org>
9 years agofm10k: Unlock mailbox on VLAN addition failures
Matthew Vick [Fri, 3 Oct 2014 00:43:35 +0000 (00:43 +0000)] 
fm10k: Unlock mailbox on VLAN addition failures

After grabbing the mailbox lock and detecting an error, the lock must be
released before the error code can be returned.

Signed-off-by: Matthew Vick <matthew.vick@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoserial: atmel: add missing dmaengine header
Vinod Koul [Thu, 16 Oct 2014 07:29:06 +0000 (12:59 +0530)] 
serial: atmel: add missing dmaengine header

The atmel serial driver uses dmaengine APIs but never included the dmaengine
header as it was getting inculded thru one of driver headers.

commit 3d588f83e4d6a5230d9094b97d38621cbaa9a972 - "dmaengine: dw: split
dma-dw.h to platform and private parts" broke this as it moved headers
around.  Fix this by doing the right thing to include the dmaengine header

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Fixes: 08f738be88bb (serial: at91: add tx dma support)
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agofm10k: Check the host state when bringing the interface up
Matthew Vick [Thu, 2 Oct 2014 05:10:18 +0000 (05:10 +0000)] 
fm10k: Check the host state when bringing the interface up

Set the flag to fetch the host state before kicking off the service task
that reads the host state when bringing the interface back up.

Signed-off-by: Matthew Vick <matthew.vick@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoi2c: i801: Add Device IDs for Intel Sunrise Point PCH
james.d.ralston@intel.com [Mon, 13 Oct 2014 22:20:24 +0000 (15:20 -0700)] 
i2c: i801: Add Device IDs for Intel Sunrise Point PCH

This patch adds the I2C/SMBus Device IDs for the Intel Sunrise Point PCH.

Signed-off-by: James Ralston <james.d.ralston@intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
9 years agoi2c: hix5hd2: add i2c controller driver
Wei Yan [Tue, 7 Oct 2014 01:04:12 +0000 (09:04 +0800)] 
i2c: hix5hd2: add i2c controller driver

I2C drivers for hix5hd2 soc series, including following chipset
Hi3716CV200, Hi3719CV100, Hi3718CV100, Hi3719MV100, Hi3718MV100.

Signed-off-by: Wei Yan <sledge.yanwei@huawei.com>
Signed-off-by: Zhangfei Gao <zhangfei.gao@linaro.org>
[wsa: folded dt docs into this patch]
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
9 years agoMerge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git...
Ingo Molnar [Thu, 16 Oct 2014 05:15:45 +0000 (07:15 +0200)] 
Merge tag 'perf-core-for-mingo' of git://git./linux/kernel/git/acme/linux into perf/urgent

Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo:

 User visible changes:

  * Add a visual cue for toggle zeroing of samples in 'perf top' (Taeung Song)

  * Fix for double free in 'perf stat' when using some specific invalid
    command line combo (Yasser Shalabi)

 Infrastructure changes:

  * Add option to copy events when queuing for sorting across cpu buffers
    and enable it for 'perf kvm stat live', to avoid having events left
    in the queue pointing to the ring buffer be rewritten in high volume
    sessions.  (Alexander Yarygin, improving work done by David Ahern):

  * Document sysfs events/ interfaces (Cody P Schafer)

  * Add support to new style format of kernel PMU event. (Kan Liang)

  * Fix typos in perf/Documentation (Masanari Iida)

  * Improve callchains when using libunwind (Namhyung Kim)

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agovxlan: using pskb_may_pull as early as possible
Li RongQing [Thu, 16 Oct 2014 01:17:18 +0000 (09:17 +0800)] 
vxlan: using pskb_may_pull as early as possible

pskb_may_pull should be used to check if skb->data has enough space,
skb->len can not ensure that.

Cc: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agovxlan: fix a use after free in vxlan_encap_bypass
Li RongQing [Thu, 16 Oct 2014 00:49:41 +0000 (08:49 +0800)] 
vxlan: fix a use after free in vxlan_encap_bypass

when netif_rx() is done, the netif_rx handled skb maybe be freed,
and should not be used.

Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoopenvswitch: use vport instead of p
Fabian Frederick [Wed, 15 Oct 2014 19:03:41 +0000 (21:03 +0200)] 
openvswitch: use vport instead of p

All functions used struct vport *vport except
ovs_vport_find_upcall_portid.

This fixes 1 kerneldoc warning

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Acked-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoopenvswitch: kerneldoc warning fix
Fabian Frederick [Wed, 15 Oct 2014 19:03:18 +0000 (21:03 +0200)] 
openvswitch: kerneldoc warning fix

s/sock/gs

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Acked-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agopowerpc/pci: Fix IO space breakage after of_pci_range_to_resource() change
Michael Ellerman [Thu, 16 Oct 2014 01:29:46 +0000 (12:29 +1100)] 
powerpc/pci: Fix IO space breakage after of_pci_range_to_resource() change

Commit 0b0b0893d49b "of/pci: Fix the conversion of IO ranges into IO
resources" changed the behaviour of of_pci_range_to_resource().

Previously it simply populated the resource based on the arguments. Now
it calls pci_register_io_range() and pci_address_to_pio(). These both
have two implementations depending on whether PCI_IOBASE is defined,
which it is not for powerpc.

Further complicating matters, both routines are weak, and powerpc
implements it's own version of one - pci_address_to_pio(). However
powerpc's implementation depends on other initialisations which are done
later in boot.

The end result is incorrectly initialised IO space. Often we can get
away with that, because we don't make much use of IO space. However
virtio requires it, so we see eg:

  pci_bus 0000:00: root bus resource [io  0xffff] (bus address [0xffffffffffffffff-0xffffffffffffffff])
  PCI: Cannot allocate resource region 0 of device 0000:00:01.0, will remap
  virtio-pci 0000:00:01.0: can't enable device: BAR 0 [io  size 0x0020] not assigned

The simplest fix for now is to just stop using of_pci_range_to_resource(),
and open-code the original implementation, that's all we want it to do.

Fixes: 0b0b0893d49b ("of/pci: Fix the conversion of IO ranges into IO resources")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
9 years agogianfar: Add FCS to rx buffer size (fix)
Claudiu Manoil [Wed, 15 Oct 2014 16:11:46 +0000 (19:11 +0300)] 
gianfar: Add FCS to rx buffer size (fix)

For each Rx frame the eTSEC writes its FCS (Frame Check Sequence)
to the Rx buffer.

The eTSEC h/w manual states in the "Receive Buffer Descriptor Field
Descriptions" table:
"Data length is the number of octets written by the eTSEC into this BD's
data buffer if L is cleared (the value is equal to MRBLR), or, if L is
set, the length of the frame including *CRC*, FCB (if RCTRL[PRSDEP > 00),
preamble (if MACCFG2[PreAmRxEn]=1), time stamp (if RCTRL[TS] = 1) and
any padding (RCTRL[PAL])."

Though the FCS bytes are removed by the driver before passing the skb
to the net stack, the Rx buffer size computation does not currently
take into account the FCS bytes (4 bytes).
Because the Rx buffer size is multiple of 512 bytes, leaving out the
FCS is not a problem for the default MTU of 1500, as the Rx buffer size
is 1536 in this case.  However, for custom MTUs, where the difference
between the MTU size and the Rx buffer size is less, this can be a
problem as the computed Rx buffer size won't be enough to accomodate
the FCS for a received frame that is big enough (close to MTU size).
In such case the received frame is considered to be incomplete (L flag
not set in the RxBD status) and silently dropped.

Note that the driver does not currently support S/G on Rx, so it has to
compute its Rx buffer size based on the MTU of the device.

Reported-by: Kristian Otnes <kotnes@cisco.com>
Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agopstore: Fix duplicate {console,ftrace}-efi entries
Valdis Kletnieks [Mon, 13 Oct 2014 03:09:08 +0000 (23:09 -0400)] 
pstore: Fix duplicate {console,ftrace}-efi entries

The pstore filesystem still creates duplicate filename/inode pairs for
some pstore types.  Add the id to the filename to prevent that.

Before patch:

[/sys/fs/pstore] ls -li
total 0
1250 -r--r--r--. 1 root root 67 Sep 29 17:09 console-efi
1250 -r--r--r--. 1 root root 67 Sep 29 17:09 console-efi
1250 -r--r--r--. 1 root root 67 Sep 29 17:09 console-efi
1250 -r--r--r--. 1 root root 67 Sep 29 17:09 console-efi
1250 -r--r--r--. 1 root root 67 Sep 29 17:09 console-efi
1250 -r--r--r--. 1 root root 67 Sep 29 17:09 console-efi
1250 -r--r--r--. 1 root root 67 Sep 29 17:09 console-efi
1250 -r--r--r--. 1 root root 67 Sep 29 17:09 console-efi
1250 -r--r--r--. 1 root root 67 Sep 29 17:09 console-efi

After:

[/sys/fs/pstore] ls -li
total 0
1232 -r--r--r--. 1 root root 148 Sep 29 17:09 console-efi-141202499100000
1231 -r--r--r--. 1 root root  67 Sep 29 17:09 console-efi-141202499200000
1230 -r--r--r--. 1 root root 148 Sep 29 17:44 console-efi-141202705400000
1229 -r--r--r--. 1 root root  67 Sep 29 17:44 console-efi-141202705500000
1228 -r--r--r--. 1 root root  67 Sep 29 20:42 console-efi-141203772600000
1227 -r--r--r--. 1 root root 148 Sep 29 23:42 console-efi-141204854900000
1226 -r--r--r--. 1 root root  67 Sep 29 23:42 console-efi-141204855000000
1225 -r--r--r--. 1 root root 148 Sep 29 23:59 console-efi-141204954200000
1224 -r--r--r--. 1 root root  67 Sep 29 23:59 console-efi-141204954400000

Signed-off-by: Valdis Kletnieks <valdis.kletnieks@vt.edu>
Acked-by: Kees Cook <keescook@chromium.org>
Cc: stable@vger.kernel.org # 3.6+
Signed-off-by: Tony Luck <tony.luck@intel.com>
9 years agovirtio_net: fix use after free
Michael S. Tsirkin [Wed, 15 Oct 2014 13:23:28 +0000 (16:23 +0300)] 
virtio_net: fix use after free

commit 0b725a2ca61bedc33a2a63d0451d528b268cf975
    net: Remove ndo_xmit_flush netdev operation, use signalling instead.

added code that looks at skb->xmit_more after the skb has
been put in TX VQ. Since some paths process the ring and free the skb
immediately, this can cause use after free.

Fix by storing xmit_more in a local variable.

Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: fec: ptp: fix convergence issue to support LinuxPTP stack
Nimrod Andy [Wed, 15 Oct 2014 09:30:12 +0000 (17:30 +0800)] 
net: fec: ptp: fix convergence issue to support LinuxPTP stack

iMX6SX IEEE 1588 module has one hw issue in capturing the ATVR register.
The current SW flow is:
ENET0->ATCR |= ENET_ATCR_CAPTURE_MASK;
ts_counter_ns = ENET0->ATVR;
The ATVR value is not expected value that cause LinuxPTP stack cannot be convergent.

ENET Block Guide/ Chapter for the iMX6SX (PELE) address the issue:
After set ENET_ATCR[Capture], there need some time cycles before the counter
value is capture in the register clock domain. The wait-time-cycles is at least
6 clock cycles of the slower clock between the register clock and the 1588 clock.
So need something like:
ENET0->ATCR |= ENET_ATCR_CAPTURE_MASK;
wait();
ts_counter_ns = ENET0->ATVR;

For iMX6SX, the 1588 ts_clk is fixed to 25Mhz, register clock is 66Mhz, so the
wait-time-cycles must be greater than 240ns (40ns * 6). The patch add 1us delay
before cpu read ATVR register.

Changes V2:
Modify the commit/comments log to describe the issue clearly.

Signed-off-by: Fugang Duan <B38611@freescale.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoperf kvm stat live: Enable events copying
Alexander Yarygin [Fri, 3 Oct 2014 14:40:12 +0000 (18:40 +0400)] 
perf kvm stat live: Enable events copying

Process of analyzing events caused by 2 functions: mmap_read() and
finished_round().

During mmap_read(), perf receives events from shared memory, queues
their pointers for further processing in finished_round() and notifies
the kernel that the events have been processed.

By the time when finished_round() is invoked, queued events can be
overwritten by the kernel, so the finished_round() occurs on potentially
corrupted memory.

Since there is no place where the event can be safely consumed, let's
copy events when queueing.

Signed-off-by: Alexander Yarygin <yarygin@linux.vnet.ibm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1412347212-28237-3-git-send-email-yarygin@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf session: Add option to copy events when queueing
Alexander Yarygin [Fri, 3 Oct 2014 14:40:11 +0000 (18:40 +0400)] 
perf session: Add option to copy events when queueing

When processing events the session code has an ordered samples queue
which is used to time-sort events coming in across multiple mmaps. At a
later point in time samples on the queue are flushed up to some
timestamp at which point the event is actually processed.

When analyzing events live (ie., record/analysis path in the same
command) there is a race that leads to corrupted events and parse errors
which cause perf to terminate. The problem is that when the event is
placed in the ordered samples queue it is only a reference to the event
which is really sitting in the mmap buffer. Even though the event is
queued for later processing the mmap tail pointer is updated which
indicates to the kernel that the event has been processed. The race is
flushing the event from the queue before it gets overwritten by some
other event. For commands trying to process events live (versus just
writing to a file) and processing a high rate of events this leads to
parse failures and perf terminates.

Examples hitting this problem are 'perf kvm stat live', especially with
nested VMs which generate 100,000+ traces per second, and a command
processing scheduling events with a high rate of context switching --
e.g., running 'perf bench sched pipe'.

This patch offers live commands an option to copy the event when it is
placed in the ordered samples queue.

Based on a patch from David Ahern <dsahern@gmail.com>

Signed-off-by: Alexander Yarygin <yarygin@linux.vnet.ibm.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1412347212-28237-2-git-send-email-yarygin@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf Documentation: Fix typos in perf/Documentation
Masanari Iida [Tue, 9 Sep 2014 15:18:50 +0000 (00:18 +0900)] 
perf Documentation: Fix typos in perf/Documentation

This patch fix spelling typos found in tool/perf/Documentation.

Signed-off-by: Masanari Iida <standby24x7@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Randy Dunlap <rdunlap@infradead.org>
Link: http://lkml.kernel.org/r/1410275930-17207-1-git-send-email-standby24x7@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf trace: Use thread_{,_set}_priv helpers
Namhyung Kim [Mon, 6 Oct 2014 00:46:03 +0000 (09:46 +0900)] 
perf trace: Use thread_{,_set}_priv helpers

This is mechanical changes only for accounting access to thread->priv
properly in the source level.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1412556363-26229-6-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf kvm: Use thread_{,_set}_priv helpers
Namhyung Kim [Mon, 6 Oct 2014 00:46:02 +0000 (09:46 +0900)] 
perf kvm: Use thread_{,_set}_priv helpers

This is mechanical changes only for accounting access to thread->priv
properly in the source level.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1412556363-26229-5-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf callchain: Create an address space per thread
Namhyung Kim [Mon, 6 Oct 2014 00:46:00 +0000 (09:46 +0900)] 
perf callchain: Create an address space per thread

The unw_addr_space_t in libunwind represents an address space to be used
for stack unwinding.  It doesn't need to be create/destory everytime to
unwind callchain (as in get_entries) and can have a same lifetime as
thread (unless exec called).

So move the address space construction/destruction logic to the thread
lifetime handling functions.  This is a preparation to enable caching in
the unwind library.

Note that it saves unw_addr_space_t object using thread__set_priv().  It
seems currently only used by perf trace and perf kvm stat commands which
don't use callchain.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jean Pihet <jean.pihet@linaro.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Arun Sharma <asharma@fb.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1412556363-26229-3-git-send-email-namhyung@kernel.org
[ Fixup unwind-libunwind.c missing CALLCHAIN_DWARF definition, added
  missing __maybe_unused on unused parameters in stubs at util/unwind.h ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf report: Set callchain_param.record_mode for future use
Namhyung Kim [Mon, 6 Oct 2014 00:45:59 +0000 (09:45 +0900)] 
perf report: Set callchain_param.record_mode for future use

Normally the callchain_param.record_mode is used only for record path.
But as it might need to prepare something for dwarf unwinding, setup
this info for perf report too.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1412556363-26229-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf evlist: Fix for double free in tools/perf stat
Yasser Shalabi [Sat, 4 Oct 2014 15:37:57 +0000 (11:37 -0400)] 
perf evlist: Fix for double free in tools/perf stat

Fix for double free bug in tools/perf due to dangling thread_map pointer
in perf_evlist struct.

Code path excercised when perf stat -C switch is used but not set and is
followed by another switch.

Example:

  perf stat -C -e.

Signed-off-by: Yasser Shalabi <yassershalabi@gmail.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1412437077-13109-1-git-send-email-yassershalabi@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf test: Add test case for pmu event new style format
Kan Liang [Tue, 7 Oct 2014 15:08:52 +0000 (11:08 -0400)] 
perf test: Add test case for pmu event new style format

Add test case in automated tests suite. It checks not only the two types
of pmu event stytle formats "pmu_event_name" and "cpu/pmu_event_name/",
but also the different formats mixtures which are more likely to trigger
parse issue.

The patch set including this one has been tested by the perf automated
test:

   ./perf test parse -v"

On haswell, ivybridge and Romley platform.

The patch set also has been tested on haswell by the following script.

Note: please make sure that your test system support TSX and
L1-dcache-loads events. Otherwise, you may want to change the events to
other pmu events.

  [lk@localhost ~]$ cat perf_style_test.sh
  # hardware events + kernel pmu event with different style
  perf stat -x, -e cycles,mem-stores,tx-start sleep 2
  perf stat -x, -e cpu-cycles,cycles-ct,cycles-t sleep 2
  perf stat -x, -e cycles,cpu/cycles-ct/,cpu/cycles-t/ sleep 2
  perf stat -x, -e instructions,cpu/tx-start/ sleep 2
  perf stat -x, -e '{cycles,tx-start}' sleep 2
  perf stat -x, -e '{cycles,cpu/tx-start/}' sleep 2

  # HW Cache event + kernel pmu event with different style
  perf stat -x, -e L1-dcache-loads,cpu/mem-stores/,tx-start sleep 2
  perf stat -x, -e L1-dcache-loads,mem-stores,cpu/tx-start/ sleep 2
  perf stat -x, -e '{L1-dcache-loads,mem-stores}' sleep 2
  perf stat -x, -e '{L1-dcache-loads,cpu/tx-start/}' sleep 2

  # Raw event + kernel pmu event with different style:
  perf stat -x, -e cpu/event=0xc0,umask=0x00/,mem-loads,cpu/mem-stores/ sleep 2
  perf stat -x, -e cpu/event=0xc0,umask=0x00/,tx-start,cpu/el-start/ sleep 2
  perf stat -x, -e '{cpu/event=0xc0,umask=0x00/,tx-start}' sleep 2

Signed-off-by: Kan Liang <kan.liang@intel.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1412694532-23391-5-git-send-email-kan.liang@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf tools: Add support to new style format of kernel PMU event
Kan Liang [Tue, 7 Oct 2014 15:08:51 +0000 (11:08 -0400)] 
perf tools: Add support to new style format of kernel PMU event

Add new rules for kernel PMU event.

Currently, the patch only want to handle the PMU event name as "a-b" and
"a".

event_pmu:
PE_KERNEL_PMU_EVENT sep_dc
|
PE_PMU_EVENT_PRE '-' PE_PMU_EVENT_SUF sep_dc

PE_KERNEL_PMU_EVENT token is for
cycles-ct/cycles-t/mem-loads/mem-stores.

The prefix cycles is mixed up with cpu-cycles.  loads and stores are
mixed up with cache event So they have to be hardcode in lex.

PE_PMU_EVENT_PRE and PE_PMU_EVENT_SUF tokens are for other PMU events.

The lex looks generic identifier up in the table and return the matched
token. If there is no match, generic PE_NAME token will be return.

Using the rules, kernel PMU event could use new style format without //

so you can use:

  perf record -e mem-loads ...

instead of:

  perf record -e cpu/mem-loads/

Signed-off-by: Kan Liang <kan.liang@intel.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1412694532-23391-4-git-send-email-kan.liang@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf tools: Parse the pmu event prefix and suffix
Kan Liang [Tue, 7 Oct 2014 15:08:50 +0000 (11:08 -0400)] 
perf tools: Parse the pmu event prefix and suffix

There are two types of event formats for PMU events. E.g. el-abort OR
cpu/el-abort/. However, the lexer mistakenly recognizes the simple style
format as two events.

The parse_events_pmu_check function uses bsearch to search the name in
known pmu event list. It can tell the lexer that the name is a PE_NAME
or a PMU event name prefix or a PMU event name suffix. All these
information will be used for accurately parsing kernel PMU events.

The pmu events list will be read from sysfs at runtime.

Note: Currently, the patch only want to handle the PMU event name as
"a-b" and "a". The only exception, "stalled-cycles-frontend" and
"stalled-cycles-fronted", are already hardcoded in lexer.

Signed-off-by: Kan Liang <kan.liang@intel.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1412694532-23391-3-git-send-email-kan.liang@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoRevert "perf tools: Default to cpu// for events v5"
Kan Liang [Tue, 7 Oct 2014 15:08:49 +0000 (11:08 -0400)] 
Revert "perf tools: Default to cpu// for events v5"

This reverts commit 50e200f07948 ("perf tools: Default to cpu// for
events v5")

The fixup cannot handle the case that
new style format(which without //) mixed with
other different formats.

For example,
group events with new style format: {mem-stores,mem-loads}
some hardware event + new style event: cycles,mem-loads
Cache event + new style event: LLC-loads,mem-loads
Raw event + new style event:
cpu/event=0xc8,umask=0x08/,mem-loads
old style event and new stytle mixture: mem-stores,cpu/mem-loads/

Signed-off-by: Kan Liang <kan.liang@intel.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1412694532-23391-2-git-send-email-kan.liang@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf Documentation: Remove Ruplicated docs for powerpc cpu specific events
Cody P Schafer [Wed, 1 Oct 2014 06:03:20 +0000 (23:03 -0700)] 
perf Documentation: Remove Ruplicated docs for powerpc cpu specific events

Listing specific events doesn't actually help us at all here because:
 - these events actually vary between different ppc processors, they
   aren't garunteed to be present.
 - the documentation of the (generic) file contents is now superceded by the
   docs for arbitrary event file contents.

Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Anshuman Khandual <khandual@linux.vnet.ibm.com>
Cc: Haren Myneni <hbabu@us.ibm.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Michael Ellerman <michaele@au1.ibm.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: linuxppc-dev@lists.ozlabs.org
Link: http://lkml.kernel.org/r/1412143402-26061-5-git-send-email-sukadev@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf Documentation: sysfs events/ interfaces
Cody P Schafer [Wed, 1 Oct 2014 06:03:19 +0000 (23:03 -0700)] 
perf Documentation: sysfs events/ interfaces

Add documentation for the <event>, <event>.scale, and <event>.unit
files in sysfs.

<event>.scale and <event>.unit were undocumented.
<event> was previously documented only for specific powerpc pmu events.

Signed-off-by: Cody P Schafer <dev@codyps.com>
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Anshuman Khandual <khandual@linux.vnet.ibm.com>
Cc: Cody P Schafer <dev@codyps.com>
Cc: Haren Myneni <hbabu@us.ibm.com>
Cc: Haren Myneni <hbabu@us.ibm.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Michael Ellerman <michaele@au1.ibm.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: linuxppc-dev@lists.ozlabs.org
Link: http://lkml.kernel.org/r/1412143402-26061-4-git-send-email-sukadev@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoperf top: Add a visual cue for toggle zeroing of samples
Taeung Song [Tue, 7 Oct 2014 07:13:15 +0000 (16:13 +0900)] 
perf top: Add a visual cue for toggle zeroing of samples

When 'perf top' is run, one can't easily find a difference
between -z option and normal output.
So I added a visual cue to know whether it is the zeroing or not.

Output is as below.

Before:
 $ perf top

 Samples: 61K of event 'cycles', Event count (approx.): 3908136933
 Overhead  Shared Object                       Symbol
    1.42%  firefox                             [.] 0x0000000000011e76
    1.32%  libpthread-2.17.so                  [.] pthread_mutex_lock

If you press key 'z' or run with zero option like '$ perf top --zero', it is as below.

After:
 Samples: 61K of event 'cycles', Event count (approx.): 3908136933 [z]
 Overhead  Shared Object                       Symbol
    1.42%  firefox                             [.] 0x0000000000011e76
    1.32%  libpthread-2.17.so                  [.] pthread_mutex_lock

Signed-off-by: Taeung Song <treeze.taeung@gmail.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1412665995-26359-1-git-send-email-treeze.taeung@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
9 years agoDrivers: ide: Remove typedef atiixp_ide_timing
Himangi Saraogi [Thu, 14 Aug 2014 16:44:30 +0000 (22:14 +0530)] 
Drivers: ide: Remove typedef atiixp_ide_timing

The Linux kernel coding style guidelines suggest not using typedefs
for structure types. This patch gets rid of the typedef for
atiixp_ide_timing.

The following Coccinelle semantic patch detects the case:

@tn1@
type td;
@@

typedef struct { ... } td;

@script:python tf@
td << tn1.td;
tdres;
@@

coccinelle.tdres = td;

@@
type tn1.td;
identifier tf.tdres;
@@

-typedef
 struct
+  tdres
   { ... }
-td
 ;

@@
type tn1.td;
identifier tf.tdres;
@@

-td
+ struct tdres

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoInput: alps - fix v4 button press recognition
Andreas Bosch [Wed, 15 Oct 2014 17:44:50 +0000 (10:44 -0700)] 
Input: alps - fix v4 button press recognition

Since the change to struct input_mt_pos some variables are now bitfields
instead of integers. Automatic conversion from integer to bitfield entry
destroys information, therefore enforce boolean interpretation instead.

Link: https://bugzilla.redhat.com/show_bug.cgi?id=1114768
Fixes: 02d04254a5df ("Input: alps - use struct input_mt_pos to track coordinates")
Signed-off-by: Andreas Bosch <linux@progandy.de>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
9 years agocxgb4i : Fix -Wmaybe-uninitialized warning.
Anish Bhatt [Wed, 15 Oct 2014 07:26:47 +0000 (00:26 -0700)] 
cxgb4i : Fix -Wmaybe-uninitialized warning.

Identified by kbuild test robot. csk family is always set to be AF_INET or
AF_INET6, so skb will always be initialized to some value but there is no harm
in silencing the warning anyways.

Signed-off-by: Anish Bhatt <anish@chelsio.com>
Fixes : f42bb57c61fd ('cxgb4i : Fix -Wunused-function warning')
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: Add ndo_gso_check
Tom Herbert [Tue, 14 Oct 2014 22:19:06 +0000 (15:19 -0700)] 
net: Add ndo_gso_check

Add ndo_gso_check which a device can define to indicate whether is
is capable of doing GSO on a packet. This funciton would be called from
the stack to determine whether software GSO is needed to be done. A
driver should populate this function if it advertises GSO types for
which there are combinations that it wouldn't be able to handle. For
instance a device that performs UDP tunneling might only implement
support for transparent Ethernet bridging type of inner packets
or might have limitations on lengths of inner headers.

Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'topic/dma_control_fsl_acks' into for-linus
Vinod Koul [Wed, 15 Oct 2014 16:09:09 +0000 (21:39 +0530)] 
Merge branch 'topic/dma_control_fsl_acks' into for-linus

9 years agoMerge branch 'topic/dma_control_cleanup_acks' into for-linus
Vinod Koul [Wed, 15 Oct 2014 16:08:49 +0000 (21:38 +0530)] 
Merge branch 'topic/dma_control_cleanup_acks' into for-linus

9 years agodmaengine: remove FSLDMA_EXTERNAL_START
Vinod Koul [Sat, 11 Oct 2014 15:46:48 +0000 (21:16 +0530)] 
dmaengine: remove FSLDMA_EXTERNAL_START

as users have been converted, so no need of this custom method

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agodmaengine: freescale: remove FSLDMA_EXTERNAL_START control method
Vinod Koul [Sat, 11 Oct 2014 15:46:47 +0000 (21:16 +0530)] 
dmaengine: freescale: remove FSLDMA_EXTERNAL_START control method

since users have been move to fsl_dma_external_start() API, so remove this
now

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agocarma-fpga: move to fsl_dma_external_start()
Vinod Koul [Sat, 11 Oct 2014 15:46:46 +0000 (21:16 +0530)] 
carma-fpga: move to fsl_dma_external_start()

carma-fpga driver uses device control with custom FSLDMA_EXTERNAL_START
command. Since we wnat to deprecate the device control, move this driver to
use new fsl_dma_external_start() API

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agocarma-fpga: use dmaengine_xxx() API
Vinod Koul [Sat, 11 Oct 2014 15:46:45 +0000 (21:16 +0530)] 
carma-fpga: use dmaengine_xxx() API

The drivers should use dmaengine_slave_config() and dmaengine_prep_dma_sg()
API instead of accessing the device_control which will be deprecated soon

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agodmaengine: freescale: add and export fsl_dma_external_start()
Vinod Koul [Sat, 11 Oct 2014 15:46:44 +0000 (21:16 +0530)] 
dmaengine: freescale: add and export fsl_dma_external_start()

The freescale driver uses custom device control FSLDMA_EXTERNAL_START to
put the controller in external start mode.
Since we are planning to deprecate the device control, move this to exported
API. Subsequent patches will remove the FSLDMA_EXTERNAL_START

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agodmaengine: add dmaengine_prep_dma_sg() helper
Vinod Koul [Sat, 11 Oct 2014 15:46:43 +0000 (21:16 +0530)] 
dmaengine: add dmaengine_prep_dma_sg() helper

This was only prep API which didnt have an helper

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agovideo: mx3fb: use dmaengine_terminate_all() API
Vinod Koul [Sat, 11 Oct 2014 15:40:40 +0000 (21:10 +0530)] 
video: mx3fb: use dmaengine_terminate_all() API

The drivers should use dmaengine_terminate_all() API instead of
accessing the device_control which will be deprecated soon

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agoserial: sh-sci: use dmaengine_terminate_all() API
Vinod Koul [Sat, 11 Oct 2014 15:40:37 +0000 (21:10 +0530)] 
serial: sh-sci: use dmaengine_terminate_all() API

The drivers should use dmaengine_terminate_all() API instead of
accessing the device_control which will be deprecated soon

Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agonet: ks8842: use dmaengine_terminate_all() API
Vinod Koul [Sat, 11 Oct 2014 15:40:34 +0000 (21:10 +0530)] 
net: ks8842: use dmaengine_terminate_all() API

The drivers should use dmaengine_terminate_all() API instead of
accessing the device_control which will be deprecated soon

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agomtd: sh_flctl: use dmaengine_terminate_all() API
Vinod Koul [Sat, 11 Oct 2014 15:40:33 +0000 (21:10 +0530)] 
mtd: sh_flctl: use dmaengine_terminate_all() API

The drivers should use dmaengine_terminate_all() API instead of
accessing the device_control which will be deprecated soon

Acked-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agomtd: fsmc_nand: use dmaengine_terminate_all() API
Vinod Koul [Sat, 11 Oct 2014 15:40:32 +0000 (21:10 +0530)] 
mtd: fsmc_nand: use dmaengine_terminate_all() API

The drivers should use dmaengine_terminate_all() API instead of
accessing the device_control which will be deprecated soon

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agoV4L2: mx3_camer: use dmaengine_pause() API
Vinod Koul [Sat, 11 Oct 2014 15:40:31 +0000 (21:10 +0530)] 
V4L2: mx3_camer: use dmaengine_pause() API

The drivers should use dmaengine_pause() API instead of
accessing the device_control which will be deprecated soon

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agodmaengine: coh901318: use dmaengine_terminate_all() API
Vinod Koul [Sat, 11 Oct 2014 15:40:30 +0000 (21:10 +0530)] 
dmaengine: coh901318: use dmaengine_terminate_all() API

The drivers should use dmaengine_terminate_all() API instead of accessing
the device_control which will be deprecated soon

Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agopata_arasan_cf: use dmaengine_terminate_all() API
Vinod Koul [Sat, 11 Oct 2014 15:40:29 +0000 (21:10 +0530)] 
pata_arasan_cf: use dmaengine_terminate_all() API

The drivers should use dmaengine_terminate_all() API instead of accessing
the device_control which will be deprecated soon

Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
9 years agostmmac: fix sti compatibililies
Giuseppe CAVALLARO [Wed, 15 Oct 2014 05:30:41 +0000 (07:30 +0200)] 
stmmac: fix sti compatibililies

this patch is to fix the stmmac data compatibilities for
all the SoCs inside the platform file.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This page took 0.061323 seconds and 5 git commands to generate.