* master:
  bigfile: Note that ramh_alloc_page does not add the page to any lists
  bigfile: Unstub ram_close
  bigfile/test/test_ram: Don't forget to free allocated Page structs
  bigfile: RAM must be explicitly free'ed after close
  bigfile/test: Don't forget to close opened RAM
  tests: TSAN no longer fails on test_virtmem
  tox: Don't duplicate setup.py on which for-tests dependencies we need
  3rdparty/ccan: Update

It was already said there that allocated page is not associated with
fileh. However the code is doing more - it does not add the page to
ram->lru_list and (obviously) to fileh->dirty_pages lists.

Document that explicitly.

It should release resources associated with RAM. Make it cal .ram_close
from RAM ops. Add corresponding .ram_close to ram_shmfs. This fixes
SHMFS_RAM->prefix leak.

test_ram is low-level test that tests RAM pages allocation/mmapping.
As allocated pages are not integrated with virtmem (not added to any
file mapping and RAM->lru_list) the Page structs have to be explicitly
freed. Fixes e.g.

	Direct leak of 80 byte(s) in 1 object(s) allocated from:
	    #0 0x7ff29af46518 in calloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9518)
	    #1 0x56131dc22289 in zalloc include/wendelin/utils.h:67
	    #2 0x56131dc225d6 in ramh_alloc_page bigfile/tests/../ram.c:41
	    #3 0x56131dc2a19e in main bigfile/tests/test_ram.c:130
	    #4 0x7ff29ac9f09a in __libc_start_main ../csu/libc-start.c:308

Else on-heap allocated RAM object is leaked. Fixes e.g. the following
error on ASAN:

	Direct leak of 56 byte(s) in 1 object(s) allocated from:
	    #0 0x7fc9ef390518 in calloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9518)
	    #1 0x555ca792f309 in zalloc include/wendelin/utils.h:67
	    #2 0x555ca7935f9a in ram_limited_new bigfile/tests/../../t/t_utils.c:35
	    #3 0x555ca793a0ba in test_file_access_synthetic bigfile/tests/test_virtmem.c:292
	    #4 0x555ca7967bc4 in main bigfile/tests/test_virtmem.c:1121
	    #5 0x7fc9ef0e909a in __libc_start_main ../csu/libc-start.c:308

Only one place that was using ram_new was missing to call ram_close in
the end.

For failing case compiler-rt support was added in 2014 - 5 years ago
(see links in removed code).

-> Use .[test] to refer to them.
https://stackoverflow.com/a/41398850/9456786

Just an update to latest CCAN - there is actually no changes to modules
that we use (tap, array_size, minmax, bitmap, build_assert)

... and is still stuck in kernel waiting for FUSE request it sent to
self (mkrevfile) to complete. For example:

	==================
	WARNING: DATA RACE
	Write at 0x00c00009a2a0 by goroutine 17:
	  github.com/hanwen/go-fuse/fuse/nodefs.(*portableHandleMap).Register()
	      /home/kirr/src/neo/src/github.com/hanwen/go-fuse/fuse/nodefs/handle.go:87 +0x130
	  github.com/hanwen/go-fuse/fuse/nodefs.(*FileSystemConnector).lookupUpdate()
	      /home/kirr/src/neo/src/github.com/hanwen/go-fuse/fuse/nodefs/fsconnector.go:126 +0x65
	  github.com/hanwen/go-fuse/fuse/nodefs.(*rawBridge).Lookup()
	      /home/kirr/src/neo/src/github.com/hanwen/go-fuse/fuse/nodefs/fsops.go:119 +0x203
	  github.com/hanwen/go-fuse/fuse.doLookup()
	      /home/kirr/src/neo/src/github.com/hanwen/go-fuse/fuse/opcode.go:335 +0x16e
	  github.com/hanwen/go-fuse/fuse.(*Server).handleRequest()
	      /home/kirr/src/neo/src/github.com/hanwen/go-fuse/fuse/server.go:470 +0x868
	  github.com/hanwen/go-fuse/fuse.(*Server).loop()
	      /home/kirr/src/neo/src/github.com/hanwen/go-fuse/fuse/server.go:437 +0x1de

	Previous read at 0x00c00009a2a0 by goroutine 44:
	  main.(*Root).mkrevfile()
	      /home/kirr/src/neo/src/github.com/hanwen/go-fuse/fuse/nodefs/inode.go:66 +0x476
	  main.(*BigFile).invalidateBlk.func1()
	      /home/kirr/src/wendelin/wendelin.core/wcfs/wcfs.go:980 +0x1b0
	  main.(*BigFile).invalidateBlk()
	      /home/kirr/src/wendelin/wendelin.core/wcfs/wcfs.go:990 +0x33f
	  main.(*Root).handleδZ.func3()
	      /home/kirr/src/wendelin/wendelin.core/wcfs/wcfs.go:811 +0x8a
	  golang.org/x/sync/errgroup.(*Group).Go.func1()
	      /home/kirr/src/tools/go/g.env/src/golang.org/x/sync/errgroup/errgroup.go:57 +0x64

	Goroutine 17 (running) created at:
	  github.com/hanwen/go-fuse/fuse.(*Server).readRequest()
	      /home/kirr/src/neo/src/github.com/hanwen/go-fuse/fuse/server.go:304 +0x5d3
	  github.com/hanwen/go-fuse/fuse.(*Server).loop()
	      /home/kirr/src/neo/src/github.com/hanwen/go-fuse/fuse/server.go:415 +0x78
	  github.com/hanwen/go-fuse/fuse.(*Server).Serve()
	      /home/kirr/src/neo/src/github.com/hanwen/go-fuse/fuse/server.go:360 +0x7d
	  main._main.func2()
	      /home/kirr/src/wendelin/wendelin.core/wcfs/wcfs.go:2292 +0x52

	Goroutine 44 (running) created at:
	  golang.org/x/sync/errgroup.(*Group).Go()
	      /home/kirr/src/tools/go/g.env/src/golang.org/x/sync/errgroup/errgroup.go:54 +0x73
	  main.(*Root).handleδZ()
	      /home/kirr/src/wendelin/wendelin.core/wcfs/wcfs.go:810 +0x47d
	  main.(*Root).zwatcher()
	      /home/kirr/src/wendelin/wendelin.core/wcfs/wcfs.go:731 +0x32b
	  main._main()
	      /home/kirr/src/wendelin/wendelin.core/wcfs/wcfs.go:2303 +0x21d1
	  main.main()
	      /home/kirr/src/wendelin/wendelin.core/wcfs/wcfs.go:2114 +0x89
	==================

	C: test timed out after 7.0s -> aborting wcfs fuse connection to unblock
	Traceback (most recent call last):
	  File "/home/kirr/src/tools/go/pygolang/golang/__init__.py", line 122, in _
	    return f(*argv, **kw)
	  File "/home/kirr/src/wendelin/wendelin.core/wcfs/wcfs_test.py", line 289, in close
	    tf.close()
	  File "/home/kirr/src/wendelin/wendelin.core/wcfs/wcfs_test.py", line 505, in close
	    t.f.close()
	IOError: [Errno 107] Transport endpoint is not connected
	Traceback (most recent call last):
	  File "/home/kirr/src/tools/go/pygolang/golang/__init__.py", line 122, in _
	    return f(*argv, **kw)
	  File "/home/kirr/src/wendelin/wendelin.core/wcfs/wcfs_test.py", line 1284, in test_wcfs_basic
	    at2 = t.commit(zf, {2:'c2', 3:'d2', 5:'f2'})
	  File "/home/kirr/src/wendelin/wendelin.core/wcfs/wcfs_test.py", line 376, in commit
	    t._wcsync()
	  File "/home/kirr/src/wendelin/wendelin.core/wcfs/wcfs_test.py", line 382, in _wcsync
	    l = t._wc_zheadfh.readline()
	IOError: [Errno 103] Software caused connection abort

before the patch after data race reporting on wcfs.go side and wcfs.go
aborting (we set GORACE="halt_on_error") the test was still stuck and
one wcfs thread was still alive:

	├─py.test,58 /home/kirr/src/wendelin/venv/z-dev/bin/py.test -p no:cacheprovider -vsx -k test_wcfs_basic
	│   └─(wcfs,67)
	│       └─{wcfs},82

with 67 being already killed - zombie - and 82 received the same signal
but being alive waiting in kernel for answer to a FUSE request:

	# cat /proc/82/stack
	[<0>] request_wait_answer+0x206/0x23d
	[<0>] __fuse_request_send+0x5d/0x80
	[<0>] fuse_simple_request+0xc8/0x180
	[<0>] fuse_lookup_name+0xfa/0x200
	[<0>] fuse_lookup+0x4a/0x110
	[<0>] fuse_atomic_open+0xa8/0x100
	[<0>] path_openat+0xbe2/0x12a0
	[<0>] do_filp_open+0x8e/0x100
	[<0>] do_sys_open+0x181/0x220
	[<0>] do_syscall_64+0x59/0x35b
	[<0>] entry_SYSCALL_64_after_hwframe+0x44/0xa9
	[<0>] 0xffffffffffffffff

and manual `kill -9 82` not making any effect, because after
git.kernel.org/linus/a131de0a482a request_wait_answer even after,
receiving fatal signal, waits for request to be answered.

It was explicitly said in that kernel commit that the way to unblock
such deadlocks is /sys/fs/fuse/connection/<X>/abort - so we do it this
way on timeout to make sure a test is never stuck forever.