summaryrefslogtreecommitdiff
path: root/libdiskfs
AgeCommit message (Collapse)Author
2016-01-13Fix O_DIRECTORY lookup on trivial translatorsFlavio Cruz
* libdiskfs/dir-lookup.c (diskfs_S_dir_lookup): If mustbedir, make sure entry is a directory by retrying "/", or starting the translator and retrying "/". * libnetfs/dir-lookup.c (netfs_S_dir_lookup): Likewise.
2015-12-29fix compiler warnings in hurd/libdiskfsFlavio Cruz
2015-12-01libdiskfs: use ihash for the node cacheJustus Winter
Replace the hand-written hash table in the node cache with libihash. Libihash is a self-tuning hash table, whereas the previous code used a fixed number of buckets. * libdiskfs/Makefile (HURDLIBS): Link to `ihash'. * libdiskfs/diskfs.h (struct node): Remove bucket list, add slot pointer. * libdiskfs/node-cache.c (nodecache): New ihash table replacing the old `nodehash'. (lookup): Drop function. (diskfs_cached_lookup_context): Adapt accordingly. (diskfs_cached_ifind): Likewise. (diskfs_try_dropping_softrefs): Likewise. (diskfs_node_iterate): Likewise.
2015-11-29libihash: provide a general purpose hash algorithmJustus Winter
* libdiskfs/name-cache.c: Move the Murmur3 algorithm... * libihash/murmur3.c: ... here, and properly attribute the code. * libihash/ihash.h (hurd_ihash_hash32): New prototype. * libihash/Makefile (SRCS): Add new file.
2015-10-05libdiskfs: fflush stdout when pausingJustus Winter
* libdiskfs/boot-start.c (diskfs_start_bootstrap): fflush stdout when pausing for /hurd/startup.
2015-09-27libdiskfs: fflush stdout when pausingJustus Winter
* libdiskfs/boot-start.c (start_execserver): fflush stdout when pausing.
2015-09-27libdiskfs: improve error handlingJustus Winter
* libdiskfs/boot-start.c (start_execserver): Improve error handling.
2015-09-06Fix race condition in ext2fs when remountingJames Clarke
On some systems, ext2fs.static would regularly hang at startup, as a race condition meant it would process paging requests while remounting. To fix this, libpager has been altered to allow inhibiting and resuming its worker threads, and ext2fs uses this to inhibit paging while remounting. * console/pager.c (pager_requests): New variable. (user_pager_init): Updated call to pager_start_workers to use new pager_requests variable. * daemons/runsystem.sh: Removed artificial delay working around the race condition. * ext2fs/ext2fs.c (diskfs_reload_global_state): Call new inhibit_ext2_pager and resume_ext2_pager functions, and leave sblock as non-NULL so it will be munmapped. * ext2fs/ext2fs.h (inhibit_ext2_pager,resume_ext2_pager): New functions. * ext2fs/pager.c (file_pager_requests): New variable. (create_disk_pager): Updated call to pager_start_workers to use new file_pager_requests variable. (inhibit_ext2_pager,resume_ext2_pager): New functions. * fatfs/fatfs.h (inhibit_fat_pager,resume_fat_pager): New functions. * fatfs/pager.c (file_pager_requests): New variable. (create_fat_pager): Updated call to pager_start_workers to use new file_pager_requests variable. (inhibit_fat_pager,resume_fat_pager): New functions. * libdiskfs/disk-pager.c (diskfs_disk_pager_requests): New variable. (diskfs_start_disk_pager): Updated call to pager_start_workers to use new diskfs_disk_pager_requests variable. * libdiskfs/diskfs-pager.h (diskfs_disk_pager_requests): New variable. * libpager/demuxer.c (struct pager_requests): Renamed struct requests to struct pager_requests. Replaced queue with queue_in and queue_out pointers. Added inhibit_wakeup field. (pager_demuxer): Updated to use new queue_in/queue_out pointers. Only wake up workers if not inhibited. (worker_func): Updated to use new queue_in/queue_out pointers. Final worker thread to sleep notifies the inhibit_wakeup condition variable. (pager_start_workers): Added out parameter for the requests instance. Allocate heap space shared by both queues. Initialise new inhibit_wakeup condition. (pager_inhibit_workers,pager_resume_workers): New functions. * libpager/pager.h (struct pager_requests): Public forward definition. (pager_start_workers): Added out parameter for the requests instance. (pager_inhibit_workers,pager_resume_workers): New functions. * libpager/queue.h (queue_empty): New function. * storeio/pager.c (pager_requests): New variable. (init_dev_paging): Updated call to pager_start_workers to use new pager_requests variable.
2015-08-25libdiskfs: rename `error' variableJustus Winter
* libdiskfs/dir-lookup.c (diskfs_S_dir_lookup): Rename `error' to `err'. * libdiskfs/file-getcontrol.c (diskfs_S_file_getcontrol): Likewise. * libdiskfs/file-syncfs.c (diskfs_S_file_syncfs): Likewise.
2015-08-25libdiskfs: add missing declarationsJustus Winter
* libdiskfs/diskfs.h (diskfs_node_disknode): Add missing declaration. (diskfs_disknode_node): Likewise.
2015-08-25libdiskfs: fix typoJustus Winter
* libdiskfs/diskfs.h: Fix typo.
2015-08-20libdiskfs: fix parent lookup in the name cacheJustus Winter
* libdiskfs/name-cache.c (diskfs_check_lookup_cache): Drop stray negation.
2015-07-04Do not keep mutex locked while waiting for authenticate loopSamuel Thibault
* libdiskfs/io-reauthenticate.c (diskfs_S_io_reauthenticate): Release the node lock while blocking on the auth server and client. * libnetfs/io-reauthenticate.c (netfs_S_io_reauthenticate): Likewise. * pfinet/io-ops.c (S_io_reauthenticate): Likewise.
2015-06-05libdiskfs: do not print a message on ENOMEMJustus Winter
* libdiskfs/node-cache.c (diskfs_node_iterate): Do not print a message on ENOMEM.
2015-05-19libdiskfs: fix renaming of directoriesJustus Winter
Previously, file permissions and ownership of the target directory were ignored when renaming a directory: % mkdir a b % chmod 555 a % mv b a % ls a b * libdiskfs/dir-renamed.c (diskfs_rename_dir): Fix error handling.
2015-04-17libdiskfs: implement a node cacheJustus Winter
Previously, all users of libdiskfs implemented a node cache on their own. Move the node cache from ext2fs into libdiskfs. We preserve the previous API by marking all functions that we pull from ext2fs as weak, so that users like tmpfs can still implement their own node cache. * ext2fs/dir.c (diskfs_lookup_hard): Adjust accordingly. * ext2fs/ext2fs.c (main): Don't call `inode_init'. * ext2fs/ext2fs.h (struct disknode): Drop `hnext', `hprevp'. * ext2fs/inode.c: Move the node cache into diskfs. (diskfs_user_make_node): New function. (diskfs_try_dropping_softrefs): Rename to `diskfs_user_try_dropping_softrefs'. (read_node): Rename to `diskfs_user_read_node'. Also move a chunk of code dealing with generations from `diskfs_cached_lookup' here. * libdiskfs/Makefile (OTHERSRCS): Add `node-cache.c'. * libdiskfs/diskfs.h (struct node): Add `hnext', `hprevp'. Amend existing comments, add forward declarations. * libdiskfs/node-cache.c: New file.
2015-04-17libdiskfs: declare all inline functions as `extern inline'Justus Winter
* libdiskfs/diskfs.h (diskfs_node_disknode, diskfs_disknode_node): Declare functions as `extern inline' so that we can use them in other functions declared as `extern inline'.
2015-04-17libdiskfs: fix node initializationJustus Winter
* libdiskfs/node-make.c (init_node): Initialize flag `author_tracks_uid'.
2015-04-17libdiskfs: drop unused fields from struct nodeJustus Winter
* libdiskfs/diskfs.h (struct node): Drop unused fields from struct node.
2015-04-17libdiskfs: make struct node more compactJustus Winter
* libdiskfs/diskfs.h (struct node): Turn flags into a bit field.
2015-04-17libdiskfs: lock-less reference counting of nodesJustus Winter
* libdiskfs/diskfs.h (struct node): Use refcounts_t for reference counting. (diskfs_node_refcnt_lock): Remove. (diskfs_node_norefs,diskfs_drop_node): Change comments accordingly. * libdiskfs/init-init.c: Adjust accordingly. * libdiskfs/node-drop.c: Likewise. * libdiskfs/node-make.c: Likewise. * libdiskfs/node-nput.c: Likewise. * libdiskfs/node-nputl.c: Likewise. * libdiskfs/node-nref.c: Likewise. * libdiskfs/node-nrefl.c: Likewise. * libdiskfs/node-nrele.c: Likewise. * libdiskfs/node-nrelel.c: Likewise. * ext2fs/inode.c: Likewise. * fatfs/inode.c: Likewise. * isofs/inode.c: Likewise. * tmpfs/node.c: Likewise. * doc/hurd.texi: Likewise.
2015-03-05Fix space styleSamuel Thibault
2015-01-17libdiskfs: fix port leakJustus Winter
* libdiskfs/boot-start.c (get_console): Deallocate `device_master'.
2014-12-19Avoid ./ components in mtab outputSamuel Thibault
* libdiskfs/dir-lookup.c (diskfs_S_dir_lookup): Match "." path as being the root directory. * libnetfs/dir-lookup.c (netfs_S_dir_lookup): Likewise.
2014-12-10Replace `bcopy' with `memcpy' or `memmove' as appropriateJustus Winter
* ext2fs/inode.c: Replace `bcopy' with `memcpy' or `memmove' as appropriate. * ext2fs/pager.c: Likewise. * isofs/lookup.c: Likewise. * isofs/main.c: Likewise. * isofs/rr.c: Likewise. * libdiskfs/file-get-trans.c: Likewise. * libiohelp/return-buffer.c: Likewise. * libpager/pagemap.c: Likewise. * libpipe/pq.c: Likewise. * libpipe/pq.h: Likewise. * libstore/unzipstore.c: Likewise. * mach-defpager/default_pager.c: Likewise. * pfinet/ethernet.c: Likewise. * pfinet/tunnel.c: Likewise. * storeio/dev.c: Likewise.
2014-12-10Replace `bzero' with `memset'Justus Winter
For reference, this patch was created using the following semantic patch, and then manually applying the change in all functions containing nested functions, as those are not supported by Coccinelle. @@ expression A, B; @@ - bzero (A, B) + memset (A, 0, B) * auth/auth.c: Replace `bzero' with `memset'. * boot/boot.c: Likewise. * defpager/defpager.c: Likewise. * exec/exec.c: Likewise. Also, drop `safe_bzero' and just use `hurd_safe_memset' directly. * ext2fs/ext2fs.c: Likewise. * ext2fs/getblk.c: Likewise. * ext2fs/pager.c: Likewise. * fatfs/pager.c: Likewise. * ftpfs/dir.c: Likewise. * ftpfs/netfs.c: Likewise. * isofs/inode.c: Likewise. * isofs/pager.c: Likewise. * libdiskfs/file-getfh.c: Likewise. * libdiskfs/file-statfs.c: Likewise. * libfshelp/fetch-root.c: Likewise. * libfshelp/start-translator.c: Likewise. * libftpconn/create.c: Likewise. * libftpconn/open.c: Likewise. * libftpconn/unix.c: Likewise. * libpipe/pipe.c: Likewise. * libps/procstat.c: Likewise. * libps/spec.c: Likewise. * libshouldbeinlibc/cacheq.c: Likewise. * libshouldbeinlibc/idvec.c: Likewise. * libshouldbeinlibc/ugids.c: Likewise. * libstore/argp.c: Likewise. * libstore/enc.c: Likewise. * libstore/kids.c: Likewise. * libthreads/alpha/thread.c: Likewise. * libtreefs/fsys.c: Likewise. * libtrivfs/file-statfs.c: Likewise. * mach-defpager/default_pager.c: Likewise. * pfinet/glue-include/asm/uaccess.h: Likewise. * pfinet/io-ops.c: Likewise. * pfinet/options.c: Likewise. * pfinet/socket.c: Likewise. * pfinet/timer-emul.c: Likewise. * pflocal/io.c: Likewise. * startup/startup.c: Likewise. * storeio/storeio.c: Likewise. * sutils/fstab.c: Likewise. * usermux/usermux.c: Likewise. * utils/fakeauth.c: Likewise. * utils/frobauth.c: Likewise. * utils/login.c: Likewise. * utils/x.c: Likewise.
2014-12-07libdiskfs: add payload-aware intrans functionsJustus Winter
* libdiskfs/diskfs.h (diskfs_begin_using_protid_payload): New function. (diskfs_begin_using_control_payload): Likewise. (diskfs_begin_using_bootinfo_payload): Likewise. * libdiskfs/fsmutations.h: Add mutators.
2014-11-23libdiskfs: fix fabrication of protidJustus Winter
* libdiskfs/dir-init.c (diskfs_init_dir): Fix fabrication of protid.
2014-11-21startup: bind the startup server to /servers/startupJustus Winter
Previously, the Hurd (ab)used the fact that the startup server speaks all protocols on its message port. Any server that wished to register for shutdown notifications would use proc_getmsgport to get a port to the startup server. This hardcodes the PID of /hurd/startup, and does not allow one to point a server to ones own startup server (e.g. using remap). Bind the startup server to /servers/startup instead. Use this to contact the startup server. * exec/main.c (S_exec_init): Use /servers/startup. Fall back to the old method so that the system still boots when the node /servers/startup is missing. * hurd/paths.h (_SERVERS_STARTUP): New macro. * libdiskfs/boot-start.c (diskfs_S_fsys_init): Use /servers/startup. * libdiskfs/init-startup.c (_diskfs_init_completed): Likewise. * pfinet/main.c (arrange_shutdown_notification): Likewise. * startup/Makefile (OBJS): Add fsysServer.o. * startup/startup.c (demuxer): Handle the fsys protocol. (main): Bind to /servers/startup. (S_fsys_getroot): Implement fsys_getroot. Stub out the rest.
2014-11-21startup: rename /hurd/init to /hurd/startupJustus Winter
This patch series splits /hurd/init into two programs. As a first step, this patch renames /hurd/init to /hurd/startup. It is called startup because it speaks the startup protocol. * startup: Rename init to startup. Adjust accordingly. * Makefile (prog-subdirs): Likewise. * doc/hurd.texi (Server Bootstrap): Likewise. * hurd/paths.h (_HURD_STARTUP): Likewise. * libdiskfs/boot-start.c (diskfs_boot_init_program): Likewise. * libdiskfs/opts-std-startup.c (startup_options): Likewise.
2014-11-21Fix recording complete path of passive relative translatorsSamuel Thibault
* libdiskfs/dir-lookup.c (diskfs_S_dir_lookup): Prepend current path to the relative path before recording the active translator.
2014-11-09Deal with odd kernel behaviorSamuel Thibault
Some versions of gnumach actually take address as a mapping hint, and would fail if the hint is bogus. Make sure to pass 0 for those versions. * console/pager.c (user_pager_create): Make sure to set *user to 0 before calling vm_map. * libdiskfs/disk-pager.c (diskfs_start_disk_pager): Make sure to set *image to 0 before calling vm_map. * libpager/pager-memcpy.c (pager_memcpy): Set window to 0 before calling vm_map. * tmpfs/node.c (diskfs_get_filemap): Make sure to set np->dn->u.reg.memref to 0 before calling vm_map.
2014-11-03libpager: use a fixed number of threadsJustus Winter
Previously, libpager used an unbounded number of threads to receive messages from the pager bucket. It used sequence barriers to execute the requests to order requests to each object. The sequence barriers are implemented in seqnos.c. A server function uses _pager_wait_for_seqno to wait for its sequence number and _pager_release_seqno to release it, or it uses _pager_update_seqno to do both operations in one step. These sequence barriers divide each server function in three parts: A, B, and C. A_i happens "before" the sequence barrier i, B_i happens "in order", C_i happens "after" the sequence barrier. This partial order < has the following properties: * This order is *per object*. Requests to different objects are not ordered. * A_i < B_i, B_i < C_i (due to the structure of the code) * B_i < B_{i+1} (this is due to the sequence barriers) * Note that only the B parts are ordered by the sequence numbers, we are free to execute C_i and C_{i+1} in any possible order. The same argument applies to the A parts. The sequence barriers are implemented using a very simple ticket algorithm. Every request, even the invalid ones, is processed by a thread, and waits until the ticket count reaches its seqno, does some work in-order, then increments the ticket and awakes all threads that have piled up up to this moment. All of them except one will then discover that it's not their turn yet and go to sleep again. Creating one thread per request has proven to be problematic as memory_object requests often arrive in large batches. This patch does two things: * Use a single thread to receive messages from the port bucket. All incoming request are put into a queue. * Use a fixed-number of threads (though even one is actually enough) to execute the the server functions. If multiple threads are used, a work-delegation mechanism ensures that the per object order < is preserved. For reference, I used the following command to create workloads that highlight the problem this patch is addressing: % settrans t .../ext2fs --sync=30 /dev/sd2s1 ... % /usr/bin/time zsh -c 'for ((i=0; i < 1500; i++)); do dd if=/dev/zero of=t/src/$i bs=4k count=290 2>/dev/null echo -n . if ((i % 100 == 0)) ; then echo -n $i; fi done' * libpager/queue.h: New file. * libpager/demuxer.c: Manage a queue of requests received from the port bucket. (pager_demuxer): Just decode the server function and enqueue the request. (worker_func): New function that consumes and executes the requests from the queue. (service_paging_requests): New function. (pager_start_workers): Likewise. * libpager/data-request.c: Remove the seqno barriers. * libpager/data-return.c: Likewise. * libpager/data-unlock.c: Likewise. * libpager/chg-compl.c: Likewise. * libpager/lock-completed.c: Likewise. * libpager/no-senders.c: Likewise. * libpager/notify-stubs.c: Likewise. * libpager/object-init.c: Likewise. * libpager/object-terminate.c: Likewise. * libpager/seqnos.c: Remove file. * libpager/stubs.c: Likewise. * libpager/pager.h (pager_demuxer): Drop declaration. (pager_start_workers): New declaration. * libpager/priv.h: Remove the _pager_seqno declarations. * libpager/Makefile (SRCS): Drop seqnos.c. * console/pager.c (user_pager_init): Call pager_start_workers. * libdiskfs/disk-pager.c: Likewise. * storeio/pager.c: Likewise. * ext2fs/pager.c (service_paging_requests): Remove function. (create_disk_pager): Start separate file pager using `pager_start_workers'. * fatfs/pager.c (service_paging_requests): Remove function. (create_fat_pager): Start separate file pager using `pager_start_workers'.
2014-11-03Fix optional definitions in static binariesSamuel Thibault
* libcons/extra-version.c (cons_extra_version): Add weak attribute. * libcons/vcons-add.c (cons_vcons_add): Likewise. * libcons/vcons-remove.c (cons_vcons_remove): Likewise. * libdiskfs/extra-version.c (diskfs_extra_version): Likewise. * libdiskfs/get-source.c (diskfs_get_source): Likewise. * libdiskfs/readonly-changed.c (diskfs_readonly_changed): Likewise. * libdiskfs/sync-default.c (diskfs_default_sync_interval): Likewise. * libdiskfs/validate-author.c (diskfs_validate_author_change): Likewise. * libdiskfs/validate-flags.c (diskfs_validate_flags_change): Likewise. * libdiskfs/validate-group.c (diskfs_validate_group_change): Likewise. * libdiskfs/validate-mode.c (diskfs_validate_mode_change): Likewise. * libdiskfs/validate-owner.c (diskfs_validate_owner_change): Likewise. * libdiskfs/validate-rdev.c (diskfs_validate_rdev_change): Likewise. * libnetfs/file-get-storage-info-default.c (netfs_file_get_storage_info): Likewise. * libnetfs/get-source.c (netfs_get_source): Likewise. * libnetfs/set-get-trans.c (netfs_set_translator, netfs_get_translator): Likewise. * libtrivfs/get-source.c (trivfs_get_source): Likewise.
2014-09-29libdiskfs: remove code counting cache missesJustus Winter
* libdiskfs/lookup.c (cache_misses): Remove. (cm_lock): Likewise. (diskfs_lookup): Do not count cache misses.
2014-09-02libdiskfs: fix /servers/exec lookupJustus Winter
* libdiskfs/boot-start.c (diskfs_start_bootstrap): Pass retry_name to dir_lookup, which is later checked to be the empty string.
2014-08-31hurd: fix semantic of file_get_childrenJustus Winter
When first introduced as fsys_get_children, it made sense to return the list of children using paths relative to the root of the filesystem that was queried. Making the get_children method part of the fsys protocol was a mistake that has since been corrected in 9366d6b2. Instead of returning paths relative to the root of the translator, return paths relative to the path of the receiving node. This fixes a problem with the mtab translator. Previously, the mtab translator invoked on a target that was not the root directory of a translator would compute invalid paths, e.g.: /hurd/mtab: /any/path/servers/socket/26 No such file or directory * hurd/fs.defs (file_get_children): Update comment. * libfshelp/translator-list.c (fshelp_get_active_translators): Add argument PREFIX. Filter entries not beginning with PREFIX if non-NULL, and omit PREFIX from the returned paths. * libfshelp/fshelp.h (fshelp_get_active_translators): Update comment accordingly. Also clarify that both FILTER and PREFIX can be NULL. * libdiskfs/file-get-children.c (diskfs_S_file_get_children): Update comment, pass prefix to fshelp_get_active_translators. * libnetfs/file-get-children.c (netfs_S_file_get_children): Likewise.
2014-08-31libdiskfs: fix reference counting of peropen objectsJustus Winter
Previously, peropen objects were created with a reference count of zero. Therefore, if diskfs_create_protid fails, passing such an object to diskfs_release_peropen would lead to a reference count underflow. * libdiskfs/peropen-make.c (diskfs_peropen_make): Initialize reference count to one. * libdiskfs/protid-make.c (diskfs_start_protid): And consume this reference on success. Update comment. (diskfs_create_protid): Update comment. * libdiskfs/diskfs.h: Update comments. * libdiskfs/io-duplicate.c (diskfs_S_io_duplicate): Adjust reference counting accordingly. * libdiskfs/io-reauthenticate.c (diskfs_S_io_reauthenticate): Likewise. * libdiskfs/io-restrict-auth.c (diskfs_S_io_restrict_auth): Likewise. * doc/hurd.texi (Diskfs Internals): Update accordingly.
2014-06-23libdiskfs: fix deadlockJustus Winter
Previously, libdiskfs would deadlock on contention on renamedirlock due to the lock being taken spuriously. Found using fsstress from the Linux Test Project. * libdiskfs/dir-rename.c (diskfs_S_dir_rename): Remove spurious pthread_mutex_lock.
2014-06-22libdiskfs: expose the ST_NOATIME flagPino Toscano
Expose ST_NOATIME as flag (if available in glibc) if diskfs is set in noatime mode. * libdiskfs/file-statfs.c (diskfs_S_file_statfs): Set ST_NOATIME if _DISKFS_NOATIME is set.
2014-06-18libdiskfs: add permission check to file_chflagsJustus Winter
Only root is allowed to change the high 16 bits. The TODO entry says otherwise, but that must be a mistake. For reference, see the glibc sources, sysdeps/mach/hurd/bits/stat.h. * libdiskfs/file-chflags.c (diskfs_S_file_chflags): Add permission check. * TODO (libdiskfs): Remove entry.
2014-06-06libdiskfs: avoid implicit integer conversionJustus Winter
* libdiskfs/boot-start.c (diskfs_boot_start): Avoid implicit integer conversion.
2014-06-06libdiskfs: use correct type for disk offsetsJustus Winter
* libdiskfs/diskfs.h (struct peropen): Use off_t for filepointer.
2014-05-29libdiskfs: use a hash table for the name cacheJustus Winter
Previously, name cache lookup operation completed in O(n) time. This means that making the cache too large would decrease the performance. Therefore it was required to tune the size. Implement the name cache using a hash table. We use buckets of a fixed size. We approximate the least-frequently used cache algorithm by counting the number of lookups using saturating arithmetic in the two lowest bits of the pointer to the name. Using this strategy we achieve a constant worst-case lookup and insertion time. Since we are not bound by the size of the cache anymore, increase its size from 200 to 1024. * libdiskfs/name-cache.c: Implement the name cache using a hash table. (diskfs_enter_lookup_cache): Change accordingly. (diskfs_purge_lookup_cache): Likewise. (diskfs_check_lookup_cache): Likewise. Also, hard code a cache miss for the parent of the root directory and merge unlocking and releasing of node references.
2014-05-29libdiskfs: remove the statistics code from the name cacheJustus Winter
The current name cache lookup operation completes in O(n) time. This means that making the cache too large would decrease the performance. Therefore it was required to tune the size, hence the need for statistics. We will use a data structure with worst case constant lookup times in the future, removing the need to fine tune the cache size. * libdiskfs/name-cache.c: Remove the statistics code from the name cache.
2014-05-28libdiskfs: add diskfs_make_node_alloc to allocate fat nodesJustus Winter
libdiskfs has two kind of nodes, struct node and struct netnode. struct node is used to store libdiskfs specific data, while struct netnode contains user supplied data. Previously, both objects were allocated separatly, and a pointer from the node to the netnode provided a mapping from the former to the latter. Provide a function diskfs_make_node_alloc that allocates both nodes in a contiguous region. This reduces the memory allocation overhead when creating nodes. It also makes the relation between node and netnode a simple offset calculation. Provide two functions to compute the netnode address from the node address and vice-versa. Most notably, this makes implementing a cache on top of libdiskfs easier. Auxiliary data for the cache can be stored in the user-defined netnode, and the fat node can be used as the value. * libdiskfs/node-make.c (init_node): Move initialization here. (diskfs_make_node): Use init_node. (diskfs_make_node_alloc): New function to allocate fat nodes. * libdiskfs/diskfs.h (diskfs_make_node_alloc): New declaration. (_diskfs_sizeof_struct_node): Likewise. (diskfs_node_disknode): Compute disknode address from node address. (diskfs_disknode_node): And vice-versa. * libdiskfs/init-init.c (_diskfs_sizeof_struct_node): New variable.
2014-05-26libdiskfs: fix node leak in the name cacheJustus Winter
* libdiskfs/name-cache.c (diskfs_check_lookup_cache): Release node reference in a special case of lookup failure.
2014-05-26libdiskfs: fix type of dir_cache_id, node_cache_idJustus Winter
* libdiskfs/name-cache.c (struct lookup_cache): Fix type of dir_cache_id, node_cache_id.
2014-05-26libdiskfs: lock-less reference counting for peropen objectsJustus Winter
* libdiskfs/diskfs.h (struct peropen): Use refcount_t for field refcnt. * libdiskfs/peropen-make.c (diskfs_make_peropen): Initialize refcnt. * libdiskfs/peropen-rele.c (diskfs_release_peropen): Adjust accordingly. * libdiskfs/protid-make.c (diskfs_start_protid): Likewise. Also, the node must no longer be locked, adjust comment accordingly. (diskfs_create_protid): Likewise.
2014-04-29libdiskfs: set the default sync interval to 30 secondsJustus Winter
The default sync interval has been changed in 9e55fdd7 from 30 to 5 seconds. This change was not reflected in the documentation. At least for current hardware, using 30 seconds instead of just 5 alleviates the thread-storm problem. Make 30 seconds the default again. * libdiskfs/priv.h (DEFAULT_SYNC_INTERVAL): Set to 30 seconds.