Add rwlock to open file

[davies]

There is a race accessing cached chunks, so we should add a rwlock to fix that.

[davies]

logging before bus error

504055: TRUNCATE UP     from 0x8a0c (35340) to 0x17264 (94820)
504056: MAPWRITE        0x1097c (67964) thru 0x17263 (94819)	(0x68e8 (26856) bytes)
504057: MAPWRITE        0xe24 (3620) thru 0x7e13 (32275)	(0x6ff0 (28656) bytes)
504058: WRITE           0x768c (30348) thru 0xd85b (55387)	(0x61d0 (25040) bytes)
504059: MAPREAD         0x4b24 (19236) thru 0xaab7 (43703)	(0x5f94 (24468) bytes)
504060: MAPREAD         0xeb28 (60200) thru 0x17263 (94819)	(0x873c (34620) bytes)
504061: WRITE           0xb478 (46200) thru 0x135eb (79339)	(0x8174 (33140) bytes)
504062: MAPREAD         0x16760 (92000) thru 0x16ca7 (93351)	(0x548 (1352) bytes)
504063: TRUNCATE DOWN   from 0x17264 (94820) to 0x77c4 (30660)
504064: TRUNCATE UP     from 0x77c4 (30660) to 0x123f4 (74740)
504065: READ            0x13c (316) thru 0x58af (22703)	(0x5774 (22388) bytes)
504066: WRITE           0x7ba8 (31656) thru 0xdcd3 (56531)	(0x612c (24876) bytes)
504067: READ            0x3550 (13648) thru 0xfbdb (64475)	(0xc68c (50828) bytes)
504068: TRUNCATE DOWN   from 0x123f4 (74740) to 0x3d10 (15632)
extend file size from 0x3d10 (15632) to 0x186a0 (100000)
504069: TRUNCATE UP     from 0x3d10 (15632) to 0x186a0 (100000)
504070: MAPWRITE        0xb0d0 (45264) thru 0x1869f (99999)	(0xd5d0 (54736) bytes)
Bus error

acesslog

2021.07.23 23:44:53.050964 [uid:0,gid:0,pid:25327] setattr (62162,0x248,[size=74740;]): OK (62162,[-rw-r--r--:0100644,1,0,0,1627054445,1627055093,1627055093,74740]) <0.000278>
2021.07.23 23:44:53.051332 [uid:0,gid:0,pid:25327] read (62162,65536,0): OK (65536) <0.000255>
2021.07.23 23:44:53.051547 [uid:0,gid:0,pid:25327] getattr (62162): OK (62162,[-rw-r--r--:0100644,1,0,0,1627054445,1627055093,1627055093,74740]) <0.000120>
2021.07.23 23:44:53.051668 [uid:0,gid:0,pid:25327] write (62162,24876,31656): OK <0.000014>
2021.07.23 23:44:53.051805 [uid:0,gid:0,pid:25327] getattr (62162): OK (62162,[-rw-r--r--:0100644,1,0,0,1627054445,1627055093,1627055093,74740]) <0.000079>
2021.07.23 23:44:53.053039 [uid:0,gid:0,pid:25327] setattr (62162,0x248,[size=15632;]): OK (62162,[-rw-r--r--:0100644,1,0,0,1627054445,1627055093,1627055093,15632]) <0.000723>
2021.07.23 23:44:53.053213 [uid:0,gid:0,pid:25327] read (62162,12288,65536): OK (0) <0.001266>
2021.07.23 23:44:53.053418 [uid:0,gid:0,pid:25327] getattr (62162): OK (62162,[-rw-r--r--:0100644,1,0,0,1627054445,1627055093,1627055093,15632]) <0.000099>
2021.07.23 23:44:53.053860 [uid:0,gid:0,pid:25327] setattr (62162,0x248,[size=100000;]): OK (62162,[-rw-r--r--:0100644,1,0,0,1627054445,1627055093,1627055093,100000]) <0.000339>
2021.07.23 23:44:53.054237 [uid:0,gid:0,pid:25327] read (62162,102400,0): OK (67232) <0.000232>
2021.07.23 23:44:53.054513 [uid:0,gid:0,pid:0] write (62162,22176,45056): OK <0.000017>
2021.07.23 23:44:53.055282 [uid:0,gid:0,pid:25327] flush (62162): OK <0.000608>
2021.07.23 23:44:53.055851 [uid:0,gid:0,pid:0] release (62162): OK <0.000012>

It could be case that read() returned few number of bytes (67232) than expected (100000)