perf tools: Move compiler and linker flags check into config/Makefile

Moving compiler and linker flags check into config/Makefile.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1369398928-9809-5-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf tools: Move programs check into config/Makefile

Moving programs check into config/Makefile.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1369398928-9809-4-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf tools: Move arch check into config/Makefile

Moving arch check into config/Makefile.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1369398928-9809-3-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf tools: Add automated make test suite

Adding automated test for testing the build process.

To run it you need to be in perf directory or specify one with PERF
variable. It's also possible to specify optional Makefile to test via MK
variable.

Whole suite is executed twice, the second time with O=/tmp/xxx option
added.

To run the whole suite:
  $ make -f tests/make
  - make_pure: cd . && make -f Makefile
    test: test -x ./perf
  - make_clean_all: cd . && make -f Makefile clean all
    test: test -x ./perf
  - make_python_perf_so: cd . && make -f Makefile python/perf.so
    test: test -f ./python/perf.so
  - make_debug: cd . && make -f Makefile DEBUG=1
    test: test -x ./perf
  - make_no_libperl: cd . && make -f Makefile NO_LIBPERL=1
    test: test -x ./perf

You see command line for 'make_pure' test right away, and the output is
stored into 'make_pure' file.

To run simple test:
  $ make -f tests/make make_debug
  - make_debug: cd . && make -f Makefile DEBUG=1
    test: test -x ./perf

At this moment tests checks for successfull build and for existence of
several built files. Additional after-build checks could be added.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1369398928-9809-2-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf  tools: Save parent pid in thread struct

Information is available, so why not save it in case some command wants
to use it.

Signed-off-by: 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@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1369543631-5106-1-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf stats: Fix divide by 0 in variance

Number of samples needs to be greater 1 to have a variance.

Fixes nan% in perf-kvm-live output.

Signed-off-by: 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@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Runzhen Wang <runzhen@linux.vnet.ibm.com>
Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1369527896-3650-9-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf kvm: Handle realloc failures

Save previous pointer and free on failure.

Signed-off-by: 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@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Runzhen Wang <runzhen@linux.vnet.ibm.com>
Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1369527896-3650-7-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf test: Fix typo

Its 'multiple', not 'mutliple', noticed while preparing a talk for
Linuxtag'13.

Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
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-dzy9nl1ku7a5umddvdic4ibl@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hists: Rename hist_entry__add_pair arguments

The current logic is to attach pair to the leader hist_entry.

Arguments of hist_entry__add_pair function were placed the other way
round.. driving me crazy.

I.e. list_add_tail expects (new_node, head).

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1355404152-16523-3-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf diff: Use internal rb tree for hists__precompute

There's missing change for hists__precompute to iterate either
entries_collapsed or entries_in tree. The change was initiated
for hists_compute_resort function in commit:

  66f97ed perf diff: Use internal rb tree for compute resort

but was missing for hists__precompute function changes.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1355404152-16523-2-git-send-email-jolsa@redhat.com
[ committer note: Reduce patch size, no functional change ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf report: Add report.percent-limit config variable

Now an user can set a default value of --percent-limit option into the
perfconfig file.

  $ cat ~/.perfconfig
  [report]
  percent-limit = 0.1

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Pekka Enberg <penberg@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1368497347-9628-9-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf top: Add --percent-limit option

The --percent-limit option is for not showing small overhead entries in
the output.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Pekka Enberg <penberg@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1368497347-9628-8-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf report: Add --percent-limit option

The --percent-limit option is for not showing small overhead entries in
the output.  Maybe we want to set a certain default value like 0.1.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Pekka Enberg <penberg@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1368497347-9628-7-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf report: Don't bother locking when adding hist entries

The 'perf report'command is single-threaded, so no need to grab a lock.

Although the fast path of pthread_mutex_[un]lock() is very fast, there's
a ~3% gain by eliminating it when we have huge sample data.

  $ perf record -a -F 100000 -o perf.data.bench -- perf bench sched all
  $ perf record -e cycles:upp -o perf.data.before -- \
  > perf report -i perf.data.bench --stdio > /dev/null
  ... apply this patch ...
  $ perf record -e cycles:upp -o perf.data.after -- \
  > perf report -i perf.data.bench --stdio > /dev/null
  $ perf diff perf.data.{before,after} | grep pthread
             +0.02%  libpthread-2.15.so  [.] _pthread_cleanup_push_defer
             +0.02%  libpthread-2.15.so  [.] _pthread_cleanup_pop_restore
     0.05%   -0.05%  perf                [.] pthread_mutex_unlock@plt
     0.05%   -0.05%  perf                [.] pthread_mutex_lock@plt
     1.01%   -1.01%  libpthread-2.15.so  [.] pthread_mutex_lock
     1.68%   -1.68%  libpthread-2.15.so  [.] __pthread_mutex_unlock_usercnt
     0.05%   -0.05%  libpthread-2.15.so  [.] pthread_mutex_unlock

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1368497347-9628-6-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hists: Move locking to its call-sites

It's a preparation patch to eliminate unneeded locking in the perf
report path.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1368497347-9628-5-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf top: Get rid of *_threaded() functions

Those _threaded() functions are needed to make hist tree handling
thread-safe, but AFAICS the only thing it does is forcing it to use
the intermediate 'collapsed' tree.

This can be acheived by setting sort__need_collapse to 1 in cmd_top() so
no need to keep those _threaded() variants.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1368497347-9628-4-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf top: Fix percent output when no samples collected

If there's no sample, kernel and exact percent output at the header
looked like "-nan%".

Tested-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1368497347-9628-3-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf top: Fix -E option behavior

The -E/--entries option controls how many lines to be printed on stdio
output but it doesn't work as it should be:

If -E option is specified, print that many lines regardless of current
window size, if not automatically adjust number of lines printed to fit
into the window size.

Reported-by: Minchan Kim <minchan@kernel.org>
Tested-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1368497347-9628-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf record: handle death by SIGTERM

Perf data files cannot be processed until the header is updated which is
done via an on_exit handler.

If perf is killed due to a SIGTERM it does not run the on_exit hooks
leaving the perf.data file in a random state which perf-report will
happily spin on trying to read.

As noted by Mike an easy reproducer is:

perf record -a -g & sleep 1; killall perf

Fix by catching SIGTERM like it does SIGINT.

Also need to remove the kill which was added via commit f7b7c26e.

Acked-by: Stephane Eranian <eranian@google.com>
Signed-off-by: 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@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1367864663-1309-1-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf tests: Fix compile errors in bp_signal files

When building on powerpc,  we get compile errors in bp_signal.c and
bp_signal_overflow.c due to __u64 and '%llx'.

Powerpc, needs __SANE_USERSPACE_TYPES__ to be defined so we pick up
<asm-generic/int-ll64.h> and define __u64 as unsigned long long.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/20130426173320.GA7029@us.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf tools: Fix tab vs spaces issue in Makefile ifdef/endif

Unmatched spaces/tabs Makefile indentation could make the
Makefile fails. While the tabed line could be considered
sometimes as follow up for rule command, the mixed space
tab meses up with makefile if conditions.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1366796273-4780-3-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hists browser: Use sort__has_sym

The TUI hist browser had a similar variable has_symbols for the same
purpose.  Let's get rid of the duplication.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1365125198-8334-9-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf top: Use sort__has_sym

perf top had a similar variable sort_has_symbols for the same purpose.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1365125198-8334-8-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf sort: Cleanup sort__has_sym setting

The sort__has_sym variable is set only if a symbol-related sort key was
added.  Since branch stack and memory sort dimensions are separated, it
doesn't need to be checked from common dimension.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1365125198-8334-7-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf sort: Reorder HISTC_SRCLINE index

It's in common sort dimension so it'd be more natural to place it with
other common column index.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1365125198-8334-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf archive: Fix typo on Documentation

It is analysis, not analisys.

Reported-by: William Cohen <wcohen@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
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-s7476m0irq0naxkzd9iekbr3@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf sort: Consolidate sort_entry__setup_elide()

The same code was duplicate to places, factor them out to common
sort__setup_elide().

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1364991979-3008-11-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf sort: Separate out memory-specific sort keys

Since they're used only for perf mem, separate out them to a different
dimension so that normal user cannot access them by any chance.

For global/local weights, I'm not entirely sure to place them into the
memory dimension.  But it's the only user at this time.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1364991979-3008-3-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf sort: Factor out common code in sort_dimension__add()

Let's remove duplicate code.

Suggested-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1364991979-3008-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf sort: Introduce sort__mode variable

It's used for determining current sort mode which can be one of
NORMAL, BRANCH and new MEMORY.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1364816125-12212-5-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf report: Fix alignment of symbol column when -v is given

When -v option is given, the symbol sort key prints its address also but
it wasn't properly aligned since hists__calc_col_len() misses the
additional part.  Also it missed 2 spaces for 0x prefix when printing.

  $ perf report --stdio -v -s sym
  # Samples: 133  of event 'cycles'
  # Event count (approx.): 50536717
  #
  # Overhead                          Symbol
  # ........  ..............................
  #
      12.20%  0xffffffff81384c50 v [k] intel_idle
       7.62%  0xffffffff8170976a v [k] ftrace_caller
       7.02%  0x2d986d         B [.] 0x00000000002d986d

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1364816125-12212-4-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hists: Free unused mem info of a matched hist entry

The mem info is shared between matched entries so one should be freed.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1364816125-12212-3-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hists: Fix an invalid memory free on he->branch_info

The branch info was allocated for the whole stack and passed matching
hist entry for each level during processing samples.  Thus when a hist
entry tries to free its branch info like in hists__collapse_insert_entry
it'll face following error.

  *** glibc detected *** perf: munmap_chunk(): invalid pointer: 0x00000000014e9d20 ***
  ======= Backtrace: =========
  /lib64/libc.so.6[0x387d47ae16]
  perf[0x4923bd]
  perf(cmd_report+0xd68)[0x432a08]
  perf[0x41a663]
  perf(main+0x58f)[0x419eaf]
  /lib64/libc.so.6(__libc_start_main+0xf5)[0x387d421735]
  perf[0x419f95]

Fix it by allocating and copying branch info for each new hist entry.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.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/1364816125-12212-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf tools: Fix bug in isupper() and islower()

One of the reasons 'perf test' is failing on Power appears to be due to
a bug in isupper().

isupper(c) and islower(c) should be checking 'c' against the mask 0x20.
Instead they are checking sane_ctype[c] which causes isupper() to be
true for lower case letters.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/20130329192950.GA9312@us.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

drm/exynos: add cache operation backend callbacks

Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos: export a gem into dmabuf fd with O_RDWR

user process needs read/write permissions when the user process tires
buffer lock with a exported dmabuf fd through fcntl system call.

Signed-off-by: Inki Dae <inki.dae@samsung.com>

dma-buf: add lock callback for fcntl system call.

This patch adds lock callback to dma buf file operations,
and this callback will be called by fcntl system call.

With this patch, fcntl system call can be used for buffer
synchronization between CPU and CPU, and CPU and DMA in user mode.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

dmabuf-sync: add cache operation feature

With this patch, all cache operations will be done in kernel side
instead of user side.

P.S. basically, not only user shouldn't need to request cache operation
to kernel but also kernel should't need such things. However, we would
have performance overhead by unnecessary cache operations if we conform
with mainline style: cache sync just before memory ownership moves from
CPU to DMA, and just after memory ownership moves from DMA to CPU.

Signed-off-by: Inki Dae <inki.dae@samsung.com>

dmabuf-sync: add buffer synchronization framework

This patch adds a buffer synchronization framework based on DMA BUF[1]
and and based on ww-mutexes[2] for lock mechanism.

The purpose of this framework is to provide not only buffer access control
to CPU and DMA but also easy-to-use interfaces for device drivers and
user application. This framework can be used for all dma devices using
system memory as dma buffer, especially for most ARM based SoCs.

Changelog v5:
- Rmove a dependence on reservation_object: the reservation_object is used
  to hook up to ttm and dma-buf for easy sharing of reservations across
  devices. However, the dmabuf sync can be used for all dma devices; v4l2
  and drm based drivers, so doesn't need the reservation_object anymore.
  With regared to this, it adds 'void *sync' to dma_buf structure.
- All patches are rebased on mainline, Linux v3.10.

Changelog v4:
- Add user side interface for buffer synchronization mechanism and update
  descriptions related to the user side interface.

Changelog v3:
- remove cache operation relevant codes and update document file.

Changelog v2:
- use atomic_add_unless to avoid potential bug.
- add a macro for checking valid access type.
- code clean.

The mechanism of this framework has the following steps,
    1. Register dmabufs to a sync object - A task gets a new sync object and
    can add one or more dmabufs that the task wants to access.
    This registering should be performed when a device context or an event
    context such as a page flip event is created or before CPU accesses a shared
    buffer.

dma_buf_sync_get(a sync object, a dmabuf);

    2. Lock a sync object - A task tries to lock all dmabufs added in its own
    sync object. Basically, the lock mechanism uses ww-mutex[1] to avoid dead
    lock issue and for race condition between CPU and CPU, CPU and DMA, and DMA
    and DMA. Taking a lock means that others cannot access all locked dmabufs
    until the task that locked the corresponding dmabufs, unlocks all the locked
    dmabufs.
    This locking should be performed before DMA or CPU accesses these dmabufs.

dma_buf_sync_lock(a sync object);

    3. Unlock a sync object - The task unlocks all dmabufs added in its own sync
    object. The unlock means that the DMA or CPU accesses to the dmabufs have
    been completed so that others may access them.
    This unlocking should be performed after DMA or CPU has completed accesses
    to the dmabufs.

dma_buf_sync_unlock(a sync object);

    4. Unregister one or all dmabufs from a sync object - A task unregisters
    the given dmabufs from the sync object. This means that the task dosen't
    want to lock the dmabufs.
    The unregistering should be performed after DMA or CPU has completed
    accesses to the dmabufs or when dma_buf_sync_lock() is failed.

dma_buf_sync_put(a sync object, a dmabuf);
dma_buf_sync_put_all(a sync object);

    The described steps may be summarized as:
get -> lock -> CPU or DMA access to a buffer/s -> unlock -> put

This framework includes the following two features.
    1. read (shared) and write (exclusive) locks - A task is required to declare
    the access type when the task tries to register a dmabuf;
    READ, WRITE, READ DMA, or WRITE DMA.

    The below is example codes,
struct dmabuf_sync *sync;

sync = dmabuf_sync_init(NULL, "test sync");

dmabuf_sync_get(sync, dmabuf, DMA_BUF_ACCESS_R);
...

And the below can be used as access types:
DMA_BUF_ACCESS_R - CPU will access a buffer for read.
DMA_BUF_ACCESS_W - CPU will access a buffer for read or write.
DMA_BUF_ACCESS_DMA_R - DMA will access a buffer for read
DMA_BUF_ACCESS_DMA_W - DMA will access a buffer for read or
write.

    2. Mandatory resource releasing - a task cannot hold a lock indefinitely.
    A task may never try to unlock a buffer after taking a lock to the buffer.
    In this case, a timer handler to the corresponding sync object is called
    in five (default) seconds and then the timed-out buffer is unlocked by work
    queue handler to avoid lockups and to enforce resources of the buffer.

The below is how to use interfaces for device driver:
1. Allocate and Initialize a sync object:
struct dmabuf_sync *sync;

sync = dmabuf_sync_init(NULL, "test sync");
...

2. Add a dmabuf to the sync object when setting up dma buffer relevant
   registers:
dmabuf_sync_get(sync, dmabuf, DMA_BUF_ACCESS_READ);
...

3. Lock all dmabufs of the sync object before DMA or CPU accesses
   the dmabufs:
dmabuf_sync_lock(sync);
...

4. Now CPU or DMA can access all dmabufs locked in step 3.

5. Unlock all dmabufs added in a sync object after DMA or CPU access
   to these dmabufs is completed:
dmabuf_sync_unlock(sync);

   And call the following functions to release all resources,
dmabuf_sync_put_all(sync);
dmabuf_sync_fini(sync);

You can refer to actual example codes:
"drm/exynos: add dmabuf sync support for g2d driver" and
"drm/exynos: add dmabuf sync support for kms framework" from
https://git.kernel.org/cgit/linux/kernel/git/daeinki/
drm-exynos.git/log/?h=dmabuf-sync

And this framework includes fcntl system call[3] as interfaces exported
to user. As you know, user sees a buffer object as a dma-buf file descriptor.
So fcntl() call with the file descriptor means to lock some buffer region being
managed by the dma-buf object.

The below is how to use interfaces for user application:
struct flock filelock;

1. Lock a dma buf:
filelock.l_type = F_WRLCK or F_RDLCK;

/* lock entire region to the dma buf. */
filelock.lwhence = SEEK_CUR;
filelock.l_start = 0;
filelock.l_len = 0;

fcntl(dmabuf fd, F_SETLKW or F_SETLK, &filelock);
...
CPU access to the dma buf

2. Unlock a dma buf:
filelock.l_type = F_UNLCK;

fcntl(dmabuf fd, F_SETLKW or F_SETLK, &filelock);

close(dmabuf fd) call would also unlock the dma buf. And for more
detail, please refer to [3]

References:
[1] http://lwn.net/Articles/470339/
[2] https://patchwork.kernel.org/patch/2625361/
[3] http://linux.die.net/man/2/fcntl

Signed-off-by: Inki Dae <inki.dae@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

mutex: Add w/w mutex slowpath debugging

Injects EDEADLK conditions at pseudo-random interval, with
exponential backoff up to UINT_MAX (to ensure that every lock
operation still completes in a reasonable time).

This way we can test the wound slowpath even for ww mutex users
where contention is never expected, and the ww deadlock
avoidance algorithm is only needed for correctness against
malicious userspace. An example would be protecting kernel
modesetting properties, which thanks to single-threaded X isn't
really expected to contend, ever.

I've looked into using the CONFIG_FAULT_INJECTION
infrastructure, but decided against it for two reasons:

- EDEADLK handling is mandatory for ww mutex users and should
  never affect the outcome of a syscall. This is in contrast to -ENOMEM
  injection. So fine configurability isn't required.

- The fault injection framework only allows to set a simple
  probability for failure. Now the probability that a ww mutex acquire
  stage with N locks will never complete (due to too many injected
  EDEADLK backoffs) is zero. But the expected number of ww_mutex_lock
  operations for the completely uncontended case would be O(exp(N)).
  The per-acuiqire ctx exponential backoff solution choosen here only
  results in O(log N) overhead due to injection and so O(log N * N)
  lock operations. This way we can fail with high probability (and so
  have good test coverage even for fancy backoff and lock acquisition
  paths) without running into patalogical cases.

Note that EDEADLK will only ever be injected when we managed to
acquire the lock. This prevents any behaviour changes for users
which rely on the EALREADY semantics.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: dri-devel@lists.freedesktop.org
Cc: linaro-mm-sig@lists.linaro.org
Cc: rostedt@goodmis.org
Cc: daniel@ffwll.ch
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20130620113117.4001.21681.stgit@patser
Signed-off-by: Ingo Molnar <mingo@kernel.org>

mutex: Add support for wound/wait style locks

Wound/wait mutexes are used when other multiple lock
acquisitions of a similar type can be done in an arbitrary
order. The deadlock handling used here is called wait/wound in
the RDBMS literature: The older tasks waits until it can acquire
the contended lock. The younger tasks needs to back off and drop
all the locks it is currently holding, i.e. the younger task is
wounded.

For full documentation please read Documentation/ww-mutex-design.txt.

References: https://lwn.net/Articles/548909/
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Rob Clark <robdclark@gmail.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: dri-devel@lists.freedesktop.org
Cc: linaro-mm-sig@lists.linaro.org
Cc: rostedt@goodmis.org
Cc: daniel@ffwll.ch
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/51C8038C.9000106@canonical.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>

arch: Make __mutex_fastpath_lock_retval return whether fastpath succeeded or not

This will allow me to call functions that have multiple
arguments if fastpath fails. This is required to support ticket
mutexes, because they need to be able to pass an extra argument
to the fail function.

Originally I duplicated the functions, by adding
__mutex_fastpath_lock_retval_arg. This ended up being just a
duplication of the existing function, so a way to test if
fastpath was called ended up being better.

This also cleaned up the reservation mutex patch some by being
able to call an atomic_set instead of atomic_xchg, and making it
easier to detect if the wrong unlock function was previously
used.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: dri-devel@lists.freedesktop.org
Cc: linaro-mm-sig@lists.linaro.org
Cc: robclark@gmail.com
Cc: rostedt@goodmis.org
Cc: daniel@ffwll.ch
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20130620113105.4001.83929.stgit@patser
Signed-off-by: Ingo Molnar <mingo@kernel.org>

ARM: DTS: update device tree for for charger_manager.

Set supply regulator for charger manager.

Signed-off-by: Jonghwa Lee <jonghwa3.lee@samsung.com>

ARM: config: Update tizen_defconfig.

Enable CONFIG_CHARGER_MANAGER, CONFIG_EXTCON

Signed-off-by: Jonghwa Lee <jonghwa3.lee@samsung.com>

power: charger-manager: Fix a bug when it unregisters notifier block of extcon.

This patch prevents NULL pointer error cauesed by unregistering unregistered
exton notifier block. At the probing time of charger manager, it tries to
remove extcon notifier block when it fails to initialize them. It has to be
applied for only registered one. Otherwise, it'd make kernel panic. To make it
work right, it checks extcon_specific_cable_nb's extcon_dev node. If extcon
cable notifier block was registered successfully, it has proper extcon_dev
pointer if not so it has NULL pointer.

Signed-off-by: Jonghwa Lee <jonghwa3.lee@samsung.com>

power: charger-manager: regulator_get() never returns NULL.

This patch fixes return value checking of regulator_get() in charger-manager
driver. The API, regulator_get(), returns ERR_PTR() when it fails to get
regulator with given name, not NULL.

Signed-off-by: Jonghwa Lee <jonghwa3.lee@samsung.com>

charger_manager: Workaround for supporting platform data without DT parsing.

All data will be move into DT later.

Signed-off-by: Jonghwa Lee <jonghwa3.lee@samsung.com>

ARM: EXYNOS: firmware: Surround l2x0 ops with an ifdef

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: EXYNOS: pm: Add support for firmware-based L2X0 resume

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: EXYNOS: Move L2X0 cache resume to SoC PM code

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: EXYNOS: firmware: Implement l2x0_resume operation

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: firmware: Add L2X0 resume operation

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

tty: serial: samsung: Remove useless checks in suspend/resume callbacks

Since uart_port struct exists through whole driver lifetime, there is no
need to check if it is non-NULL in suspend/resume callbacks.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

clocksource: exynos_mct: Register sched_clock

This patch adds sched_clock registration to Exynos MCT driver to
improve scheduler precision by providing a high resolution clock source
for scheduling purposes.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>
Conflicts:
drivers/clocksource/exynos_mct.c

ARM: dts: exynos4412-slp_pq: Add regulator states in mem mode

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

regulator: max77686: Fix suspend-mode settings of regulators

LDO20, LDO21, LDO22, BUCK8 and BUCK9 regulators provide external pins
for enable control in addition to I2C control, which is usually used to
control regulator states in suspend.

BUCK5-BUCK7 do not provide any suspend-mode control.

This patch modifies the driver to account for these factors, making it
program sleep mode states of regulators correctly.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>
Conflicts:
drivers/regulator/max77686.c

regulator: max77686: Add set_suspend_enable callback for buck regulators

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

regulator: of: Parse regulator state in mem mode from DT

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

regulator: Handle suspend and resume of regulators

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: dts: exynos4412-slp_pq: Add sleep mode pin configuration

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

pinctrl: samsung: Parse pin groups before calling pinctrl_register()

Calling pinctrl_register() means that the driver is fully initialized
and might accept pinmux/pinconf requests, so pin groups must be parsed
before.

This patch fixes this problem by moving device tree parsing
before call to pinctrl_register().

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: EXYNOS: Add support for firmware-assisted suspend/resume

This patch adds firmware ops related to system suspend/resume that
allows suspend/resume of systems with secure firmware.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: EXYNOS: pm: Move call to flush_cache_all() before outer_flush_all()

Flushing L1 cache might cause data to be stored to L2 cache, so it is
more reasonable to flush L1 cache first and only then L2 cache.

This patch moves call to flush_cache_all() in exynos_cpu_suspend()
before outer_flush_all() to make sure that all data are flushed
correctly. In addition flushing is enabled for all Exynos SoCs to
prevent any data loss.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

clk: samsung: exynos4: Do not disable ISP bus clocks

Some ISP bus clocks must be kept enabled when entering sleep mode,
otherwise the system hangs. This patch adds CLK_IGNORE_UNUSED flags to
those clocks to keep them ungated when FIMC IS driver is not present.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

tizen_defconfig update

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>

ARM: dts: exynos4412-redwood: Add device nodes for modem_if

Add device nodes for modem_if, also set HSIC regulator to always on.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: Add modem drivers to Kconfig and Makefile

Add modem_if to Kconfig and Makefile to enable compilation.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: Change net device name choice mechanism

CONFIG_SLP is no longer used hence new method of net device names was
implemented.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: ifdef wake_lock support in the modem driver

Wake_locks are not supported outside of Android so it is essential to
make their use optional.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: Move code from board-m0-modems.c to the xmm6262 driver

Moved code from the board file to the driver of the xmm6262 chip. Also
handling of sim card detection has been commented out.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

usb: ehci-s5p: Add s5p_ehci_configurate, needed by modem_if

s5p_ehci_configurate function configures the data transfer for HSIC modem.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: Comment out enable_wake_irq in modem drivers (HSIC, xmm6262)

Enabling the interrupt caused problem with the modem.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: Move code from board-m0-modems.c to the modem driver

Move modem related code from the board file to the modem driver.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: device_tree: Add modem_if to exynos4412-slp_pq.dts

Add modem_if to the device tree of slp_pq board.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: Modification of the modem driver to compile on the 3.8 kernel

The modem_if driver code was taken from 3.0 kernel and did not compile
on 3.8. These changes are necessary to make the driver work.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: Add modem.h to include/linux/platform_data

This file is needed by modem_if driver.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: Remove linux/cma.h use

Remove linux/cma.h from modem_modemctl_device_xmm6262.c
It wasn't used anyway and was included for linux/rbtree.h only.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

modem_if: Add modem_if driver files

All files were taken from exynos3.4 kernel.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

usb: ehci-s5p: Add support for HSIC to the Exynos 4x12 SoC family

Add support for HSIC to the Exynos 4x12 SoC family.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

usb: s3c-hsotg: Changing MAC interface to 8 bit

This patch changes UTMI+ PHY interface of udc core from 16bit to
8bit. Although the manual says that only 16 bit is supported. This
solves the problem of EHCI and Device not working together.

Signed-off-by: Praveen Paneri <p.paneri@samsung.com>

ARM: dts: Enable ehci for exynos4412-redwood.dts

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

usb: ehci-s5p: Fix device tree compatible string for ehci-s5p.c

Change device tree compatible string to a more generic one covering whole
Exynos SoC family.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

usb: ehci-s5p: Create EHCI & OHCI power control sysfs

This patch enable EHCI & OHCI power control function using sysfs

This control ehci port power and initialize kernel resources.
EHCI power off: echo 0 > /sys/devices/platform/s5p-ehci/ehci_power
EHCI power on: echo 1 > /sys/devices/platform/s5p-ehci/ehci_power

OHCI power off: echo 0 > /sys/devices/platform/s5p-ohci/ohci_power
OHCI power on: echo 1 > /sys/devices/platform/s5p-ohci/ohci_power

Change-Id: I51f99fa5759cd5041b1e70cf7309047aac2232e5
Signed-off-by: Yulgon Kim <yulgon.kim@samsung.com>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>

ARM: dts: exynos4210-origen: Use real regulators for USB OTG

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: dts: exynos4210-origen: Add device node for ehci controller

This patch adds device tree node for ehci controller used on
exynos4210-origen board.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: DTS: exynos4210-origen: Add MAX8997 node entry

The node entry is added as per documentation and legacy board file.

Signed-off-by: Tushar Behera <tushar.behera@linaro.org>
Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: dts: exynos4x12: Extend usbphy reg property to cover HSIC registers

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: dts: exynos4412-slp_pq: Enable EHCI controller

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

[HACK] ARM: dts: exynos4412-slp_pq: Keep HSIC regulators always on

This is needed because the s5p-ehci driver does not control voltage
regulators.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: dts: exynos4: Add node for EHCI

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

ARM: dts: exynos4210-origen: Add device nodes for USB gadget

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

exynos4-is: Ungate uart clocks on system suspend

Due to hardware requirements, several ISP clocks must be enabled when
suspending the system. This patch adds ungating to suspend and
gating to resume callbacks to keep such clocks ungated.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>

exynos4-is: Handle suspend/resume of fimc-is-i2c correctly

If the same callbacks are used for runtime and system suspend/resume,
clocks can get disabled twice, which can lead to negative reference
counts and kernel warnings.

This patch splits suspend/resume callbacks into separate runtime and
system-wide functions, so clock gating is done correctly.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>

mmc: sdhci-s3c: remove the duplicated calling sequence

Didn't need to call the sdhci_s3c_notify_change in
setup_card_detect_gpio()

Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>

ARM: dts: Remove redundant sysreg node from exynos4.dtsi

Remove the incorrect sysreg node, which is likely in this file
due to wrong merge conflict resolution.

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>

mmc: sdhci-s3c: revert "fix missing clock for gpio card-detect"

It's produced the clk-disable counting warning.
(the pair is not right)

Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>

ARM: dts: Add missing I2C controller node alias for GP2AP020A00F sensor

Without this alias probing of the I2C0 adapter was failing as the
'i2c-gpio-2' i2c-gpio adapter was getting I2C bus id 0 assigned to it.

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>

clk: honor CLK_GET_RATE_NOCACHE in clk_set_rate

clk_set_rate() uses clk->rate directly. This causes problems if the clock
is marked as CLK_GET_RATE_NOCACHE. Hence call clk_get_rate() to get the
current rate.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>

tizen_defconfig update

Enable MIPI CSI-2/DSIM PHY driver and geenric PHY framework.

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>

video: exynos_dsi: Use generic PHY driver

Use the generic PHY API instead of the platform callback to control
the MIPI DSIM DPHY.

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>