Commit graph

31 commits

Author SHA1 Message Date
Joey Hess
b6ac443b60
fix build warnings under ghc 7.10
Caused by AMP.. Since I've finally upgraded my dev laptop to 7.10,
I may start missing imports that are not needed with it but are with older
versions..
2015-12-19 17:42:45 -04:00
Joey Hess
c670a0642c
fix warning 2015-11-16 15:37:27 -04:00
Joey Hess
e2b4861bff
store abspath to the lock file
Avoids problems if the program chdirs
2015-11-16 15:25:04 -04:00
Joey Hess
2e44da5c46
clean up side lock files when we're done with them
There's a potential race, but it's detected and just results in the other
process failing to take the side lock, so possibly retrying one second
later on. The race window is quite narrow so the extra delay is minor.

Left the side lock files mode 666 because an interruption can leave a side
lock file created by another user for a shared repository. When this
happens, the non-owning user can't delete it (+t) but can still lock it,
and so the code falls back to acting as it did before this commit.
2015-11-16 11:36:11 -04:00
Joey Hess
8efd3d71c8
starting to get a handle on how to detect that mad gleam in lustre's eye 2015-11-13 16:18:44 -04:00
Joey Hess
70bfe218f5
one more try to get sane behavior our of lustre 2015-11-13 15:51:45 -04:00
Joey Hess
389c6c7d37
fixed a fd double-close 2015-11-13 15:43:09 -04:00
Joey Hess
b0155d9093
also compare lock file contents to double-check link worked
And it closes the tmp file before this. I don't know if this will help
avoid lustre's craziness, but it can't hurt..
2015-11-13 15:20:52 -04:00
Joey Hess
1aba23ab4e
use /tmp for sidelock file when no /dev/shm 2015-11-13 14:49:30 -04:00
Joey Hess
60a9c7f5c6
require the side lock be held to take pidlock
This is less portable, since currently sidelocks rely on /dev/shm.
But, I've seen crazy lustre inconsistencies that make me not trust the
link() method at all, so what can you do.
2015-11-13 14:44:53 -04:00
Joey Hess
85345abe8b
avoid over-long filenames for side lock files 2015-11-13 14:04:29 -04:00
Joey Hess
c2cbe5619b
add stat check
I have a strace taken on a lustre filesystem on which link() returned 0,
but didn't actually succeed, since the file already existed.

One of the linux man pages recommended using link followed by checking like
this. I was reading it yesterday, but cannot find it now.
2015-11-13 13:22:45 -04:00
Joey Hess
88d94e674c
clean up temp file 2015-11-13 12:52:24 -04:00
Joey Hess
e31a51c5bb
better lock dropping order 2015-11-13 12:36:37 -04:00
Joey Hess
aa4192aea6
pid locking configuration and abstraction layer for git-annex
(not actually used anywhere yet)
2015-11-12 17:50:34 -04:00
Joey Hess
77b490bfba
add timeout for pid lock waiting 2015-11-12 17:12:54 -04:00
Joey Hess
0f25a7365a
module for PidLocks in LockPool 2015-11-12 16:31:34 -04:00
Joey Hess
710d1eeeac
module for pid lock files with atomic stale lock file takeover when possible 2015-11-12 15:39:49 -04:00
Joey Hess
c8fad345f2
add tryLockShared 2015-10-08 13:40:23 -04:00
Joey Hess
9461019e9a
open lock file ReadOnly when taking shared lock
It's only necessary to open a file for write when taking an exclusive lock.
2015-10-08 13:34:49 -04:00
Joey Hess
6c3cea7699 need more polymorphism 2015-05-22 13:50:37 -04:00
Joey Hess
9de5cd2966 fix crash in stale transfer lockfile cleanup code
Need to differentiate between the lockfile not being locked, and it not
existing.
2015-05-19 23:35:24 -04:00
Joey Hess
1312e721ed convert lockContent to use new LockPools
Also cleaned up the code, avoiding creating a lock file if we're going to
open it for create later anyway.

And, if there's an exception while preparing to lock the file, but not at
the point of actually taking the lock, throw an exception, instead of
silently not locking and pretending to succeed.

And, on Windows, always use lock file, even if the repo somehow got into
indirect mode (maybe with cygwin git..)
2015-05-19 14:12:23 -04:00
Joey Hess
6915b71c57 lock pools to work around non-concurrency/composition safety of POSIX fcntl 2015-05-18 15:57:17 -04:00
Joey Hess
e9172263e5 comment typos 2015-05-17 14:22:14 -04:00
Joey Hess
8c2dd7d8ee Fix an unlikely race that could result in two transfers of the same key running at once.
As discussed in bug report.
2015-05-12 19:39:28 -04:00
Joey Hess
afc5153157 update my email address and homepage url 2015-01-21 12:50:09 -04:00
Joey Hess
59eae904b1 final scary locking refactoring (for now)
Note that while before checkTransfer this called getLock with WriteLock,
getLockStatus's use of ReadLock will also notice any exclusive locks.
Since transfer info files are only locked exclusively, never shared,
there is no behavior change.

Also, fixes checkLocked to actually return Just False when the file
exists, but is not locked.
2014-08-20 19:30:40 -04:00
Joey Hess
1994771215 more lock file refactoring
Also fixes a test suite failures introduced in recent commits, where
inAnnexSafe failed in indirect mode, since it tried to open the lock file
ReadWrite. This is why the new checkLocked opens it ReadOnly.

This commit was sponsored by Chad Horohoe.
2014-08-20 18:58:14 -04:00
Joey Hess
ec7dd0446a more lock file refactoring 2014-08-20 17:03:04 -04:00
Joey Hess
d279180266 reorganize and refactor lock code
Added a convenience Utility.LockFile that is not a windows/posix
portability shim, but still manages to cut down on the boilerplate around
locking.

This commit was sponsored by Johan Herland.
2014-08-20 16:45:58 -04:00