| PVFS:/src/io/bmi/bmi_tcp/bmi-tcp.c |
Line History |
History |
latest revision Branch MAIN1.134 annotated / raw | Diffs: previous, other | Lines: 4019 ( +10, -1 ) Created: 2009-01-06 14:56:40 -0500 (17 hours ago) | Author: slang | Changeset: MAIN:slang:20090106195640 workaround for bmi tcp testcontext always returning immediately in shared client/server environments.
1.133 annotated / raw | Diffs: previous, other | Lines: 4010 ( +6, -12 ) Created: 2008-11-24 10:01:01 -0500 (43 days ago) | Author: pcarns | Changeset: MAIN:pcarns:20081124150101 Slight tweak to error handling in BMI tcp. Do not propigate accept() errorsall the way to the job level, just log and continue instead.
1.132 annotated / raw | Diffs: previous, other | Lines: 4016 ( +440, -370 ) Created: 2008-11-19 20:16:55 -0500 (48 days ago) | Author: slang | Changeset: MAIN:slang:20081120011639 merging hints/events code into trunk.
1.131 annotated / raw | Diffs: previous, other | Lines: 3946 ( +2, -1 ) Created: 2008-11-19 14:00:17 -0500 (48 days ago) | Author: slang | Changeset: MAIN:slang:20081119190003 bmi changes to allow separate bmi packaging
1.130 annotated / raw | Diffs: previous, other | Lines: 3945 ( +4, -4 ) Created: 2008-11-18 00:20:30 -0500 (50 days ago) | Author: pcarns | Changeset: MAIN:pcarns:20081118052030 committing patch from David Metheny to allow network specification in conffile to work with IP addresses that aren't precisely on a subnet boundary
1.129 annotated / raw | Diffs: previous, other | Lines: 3945 ( +1, -1 ) Created: 2008-11-12 12:08:50 -0500 (55 days ago) | Author: slang | Changeset: MAIN:slang:20081112170849 Move aligned allocation call to bmi-tcp. Remove padding stuff from flow.
1.128 annotated / raw | Diffs: previous, other | Lines: 3945 ( +4, -11 ) Created: 2008-11-11 14:25:53 -0500 (56 days ago) | Author: slang | Changeset: MAIN:slang:20081111192540 merge of direct IO code to PVFS HEAD.
1.127 annotated / raw | Diffs: previous, other | Lines: 3952 ( +4, -4 ) Created: 2008-01-08 16:18:41 -0500 (11 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20080108211841 test programs and fix for eager recv bug in bmi_tcpBranch point for: cu-security-branch directio-branch fuse-support he-branch rongrong small-file-branch Tags: small-file-branch-point small-file-branch$BP Root_cu-security-branch cu-security-branch$BP small-file-branch-point2 he-branch$BP he-branch-point merge_HEAD_NEW_FEATURE_ADD merge_HEAD_NEW_FEATURE_ADD2 from_HEAD_NEW_FEATURE_ADD2 from_MERGE_FROM_HEAD merge_MERGE_FROM_HEAD branch-rongrong rongrong$BP ... more hide 1.126 annotated / raw | Diffs: previous, other | Lines: 3952 ( +1, -1 ) Created: 2007-12-07 18:05:11 -0500 (13 months ago) | Author: slang | Changeset: MAIN:slang:20071207230511 merge nb recv EAGAIN fix
1.125 annotated / raw | Diffs: previous, other | Lines: 3952 ( +1, -3 ) Created: 2007-11-30 14:33:17 -0500 (13 months ago) | Author: pw | Changeset: MAIN:pw:20071130193315 Remove unused ops pointer from method addr structure.
1.124 annotated / raw | Diffs: previous, other | Lines: 3954 ( +7, -1 ) Created: 2007-11-08 16:48:21 -0500 (13 months ago) | Author: slang | Changeset: MAIN:slang:20071108214821 [on behalf of Phil]:pvfs2-aio-cancel.patch ---------------------- This patch fixes a bug in the I/O cleanup path on the server side. In cases where a flow needed to cancel pending I/O operations, the trove cancel function was calling aio_cancel() directly. This doesn't work correctly if the alt-aio implementation is used. pvfs2-root-squash-address.patch ------------------------------- This fixes a bug in the root squash checking on the server side. The routine that compares a client address against the root squash list was using getsockname() rather than getpeername(). The former retrieves the server's address rather than the client's. pvfs2-ls-rm.patch ----------------- This is an interim fix for the concurrent "rm -rf" and "ls" problem that was recently discussed on the mailing list. It sounds like the long term direction is to switch to using entry names as dirent tokens, but this patch fixes the majority of cases in the mean time without a protocol change. The problem in the case I was seeing was a cache conflict between the two clients (the ls was caching tokens in the pcache that caused rm to get the wrong position). The token is 64 bits wide, but only the first 32 bits are used (the START and END values are near the top of the 32 bit range). This patch takes advantage of the extra top 32 bits on the server side to set a unique identifier in the token for each "readdir session" so that their cache entries do not collide. The client is not aware of this change because it treats the token as an opaque value. A readdir session begins when a client requests the START position. pvfs2-client-buffer-logging.patch --------------------------------- I don't know if there is any interest in this, but this adds some debugging to the buffers used in the kernel module. On startup, pvfs2-client will print the buffer pointers (whether debugging is enabled or not). There are also new debugging messages that will show the first byte of each memory buffer passing through the kernel if enabled. These logging messages were added to help track down what ended up being a server side problem (see pvfs2-aio-cancel.patch), but we kept it in case it is useful in the future. Branch point for: cu-sandbox-branch pvfs-2-7-0 pvfs-2-7-branch pvfs2-2-7-branch Tags: pvfs-2-7-branch-point pvfs-2-7-branch$BP pvfs-2-7-0 cu-sandbox-branch-point cu-sandbox-branch$BP 1.123 annotated / raw | Diffs: previous, other | Lines: 3948 ( +1, -1 ) Created: 2007-11-07 11:44:00 -0500 (13 months ago) | Author: slang | Changeset: MAIN:slang:20071107164359 fix bad commit with endpoint parsing
1.122 annotated / raw | Diffs: previous, other | Lines: 3948 ( +85, -83 ) Created: 2007-11-06 18:08:36 -0500 (14 months ago) | Author: slang | Changeset: MAIN:slang:20071106230833 bmi cleanups.
1.121 annotated / raw | Diffs: previous, other | Lines: 3946 ( +8, -3 ) Created: 2007-10-20 11:40:13 -0400 (14 months ago) | Author: slang | Changeset: MAIN:slang:20071020154013 commit of Phil's patches for the halloween bug fixes. See the pvfs2-dev for details.
Tags: mgmt-branch-point 1.120 annotated / raw | Diffs: previous, other | Lines: 3941 ( +2, -0 ) Created: 2007-10-16 18:53:44 -0400 (14 months ago) | Author: slang | Changeset: MAIN:slang:20071016225344 fix for mutex race between ref list mutex and tcp interface mutex
1.119 annotated / raw | Diffs: previous, other | Lines: 3939 ( +8, -0 ) Created: 2007-10-16 16:54:41 -0400 (14 months ago) | Author: slang | Changeset: MAIN:slang:20071016205441 switching from the blocking receive (where we had just done a peek) to the non-blocking receive didn't work perfectly since new connections weren't getting set to nonblocking after accept (it looks like this is why blocking receive was used in the first place). The fix here sets the new socket to nonblocking and adds some asserts to make sure that sockets passed to nbpeek and nbrecv are nonblocking already. Thanks to Kevin for helping debug this one.
1.118 annotated / raw | Diffs: previous, other | Lines: 3931 ( +5, -6 ) Created: 2007-10-09 17:58:31 -0400 (14 months ago) | Author: slang | Changeset: MAIN:slang:20071009215829 my fix for the halloween bug.
1.117 annotated / raw | Diffs: previous, other | Lines: 3932 ( +3, -3 ) Created: 2007-08-22 12:12:47 -0400 (16 months ago) | Author: slang | Changeset: MAIN:slang:20070822161244 Replace id_gen_safe with id_gen_fast. id_gen_safe is only needed for ids that are getting handed back to us through the system interfaces. If our own internal ids (pointers) are getting corrupted or freed...Hello valgrind!
1.116 annotated / raw | Diffs: previous, other | Lines: 3932 ( +7, -2 ) Created: 2007-07-18 17:32:58 -0400 (17 months ago) | Author: pw | Changeset: MAIN:pw:20070718213257 Move private method parameters structures to their respective methods.Branch point for: twod-branch Tags: twod-branch$BP twod-branch-point 1.115 annotated / raw | Diffs: previous, other | Lines: 3927 ( +0, -6 ) Created: 2007-07-18 17:04:00 -0400 (17 months ago) | Author: pw | Changeset: MAIN:pw:20070718210400 Remove unused method_addr field.
1.114 annotated / raw | Diffs: previous, other | Lines: 3933 ( +20, -9 ) Created: 2007-03-26 13:26:35 -0500 (21 months ago) | Author: slang | Changeset: MAIN:slang:20070326182634 [phil]: This patch corrects a variety of error code problems.See: http://www.beowulf-underground.org/pipermail/pvfs2-developers/2007-March/003253.html
Tags: pvfs-2-6-3-tag2 kunkel-tas-branch-2007-05-23 kunkel-migration-hint-branch-point-2007-06-17 1.113 annotated / raw | Diffs: previous, other | Lines: 3922 ( +4, -4 ) Created: 2006-10-23 18:49:43 -0400 (2 years 2 months ago) | Author: robl | Changeset: MAIN:robl:20061023224943 missed another not-c90 spotBranch point for: dbpf-methods-branch pvfs-2-6-3 pvfs-2-6-branch pvfs2-kunkel-tas-branch Tags: dbpf-methods-point pvfs-2-6-branch$BP pvfs2-kunkel-tas-branch$BP pvfs-2-6-0 pvfs-2-6-point2 pvfs-2-6-point3 dbpf-methods-branch$BP pvfs-2-6-point4 kunkel-hint-branch-point-26 pvfs-2-6-1 pvfs-2-6-2 kunkel-hint-branch-point-2007-02-17 kunkel-migration-hint-branch-point-2007-03-15 pvfs-2-6-3$BP ... more hide 1.112 annotated / raw | Diffs: previous, other | Lines: 3922 ( +2, -1 ) Created: 2006-10-23 17:28:43 -0400 (2 years 2 months ago) | Author: robl | Changeset: MAIN:robl:20061023212843 do things up c90-style (declarations before commands)
1.111 annotated / raw | Diffs: previous, other | Lines: 3921 ( +17, -2 ) Created: 2006-10-10 14:44:24 -0400 (2 years 2 months ago) | Author: slang | Changeset: MAIN:slang:20061010184422 [phil]: added ACCESS_DEBUG option to print hostnames instead of IP addresses in debug output.[slang]: fix to previous commit of debug statement in remove.sm
Tags: WALT3-point 1.110 annotated / raw | Diffs: previous, other | Lines: 3906 ( +1, -1 ) Created: 2006-09-17 15:09:55 -0400 (2 years 3 months ago) | Author: pw | Changeset: MAIN:pw:20060917190955 minor const fixBranch point for: 1.110.4 tau-sandbox Tags: tau-sandbox$BP tau-sandbox-point pvfs-2-6-point 1.109 annotated / raw | Diffs: previous, other | Lines: 3906 ( +1, -1 ) Created: 2006-09-11 20:49:31 -0400 (2 years 3 months ago) | Author: vilayann | Changeset: MAIN:vilayann:20060912004929 a) Use calloc() instead of malloc() for bmi-tcp's memalloc() callbackb) memset() the events structure used by epoll_ctl Both the above help prevent valgrind from spewing out uninitialized bytes warnings.. Rework the config file integrity checks into the fs-add state machine based on Sam and RobL's suggestions. The reason we do this is so that we keep the status quo of hiding config files from the system interface users. isys_fs_add continues to be the system interface whose state machine internally verifies that all servers are using the same fs.conf file by checking their file sizes and their cryptographic check sums and see if they match. pvfs2-ping.c goes back to its original form without the call to PINT_fetch_config_list() which is now deprecated and unused (it is still left in case we find a use for it someplace else). Also fixes the breakages caused by the recent export options merge in the test sub directory and a few gcc 4 warnings with void* arithmetic..
Tags: posix-extensions-point7 1.108 annotated / raw | Diffs: previous, other | Lines: 3906 ( +260, -35 ) Created: 2006-09-11 16:22:04 -0400 (2 years 3 months ago) | Author: vilayann | Changeset: MAIN:vilayann:20060911202200 Implement root squash, read only and all squash export options. All of these are server-sidesettings and set in the fs.conf files.. Perhaps the docs need to be pushed out to the pvfs.org website after this check in? Wild cards and mask-based addresses are allowed as part of the fs.conf export settings.. Also added a mount time option to honor setuid bit in files, mount -o suid... Removed the -c option from pvfs2-ping and made the config file checks default. Branch point for: locking-branch version-lock-actual-branch Tags: version-lock-actual-branch$BP version-lock-branch locking-branch$BP 1.107 annotated / raw | Diffs: previous, other | Lines: 3681 ( +14, -1 ) Created: 2006-09-11 11:42:39 -0400 (2 years 3 months ago) | Author: vilayann | Changeset: MAIN:vilayann:20060911154237 [pcarns]:tacl-xattr-homedir.patch: ------------------------- This makes tacl-xattr.sh slightly more portable. Some Linux distributions have adduser utilities that do not create the home directory for you. This patch explicitly does a mkdir -p and chmod after adduser to make sure that required home directories exist. tacl-xattr-symlink.patch: ------------------------- This is an important fix to the test script. It was using the -L argument to getfattr to traverse symbolic links when dumping the ACLs from the test directory. This leads to unpredictable results because there was no way to tell whether getfattr would traverse a real directory or its symbolic link first (it depends on the dirent order) and the other would always be left out. ACLs are not supported on symbolic links anyway, so it was just adding noise to the test script. xattr-symlink.patch: ------------------------- This is the only change to PVFS2 itself. Like most Linux file systems, PVFS2 does not support xattrs on symbolic links (despite what is implied by man pages). This is due to the fact that symbolic links have 777 permissions by default that would allow anyone on the system to store xattrs in any symbolic link. This patch updates PVFS2 semantics slightly, however, to be more in line with how other file systems implement this. In particular listxattr() is now allowed (it just returns that there are zero entries), and setxattr() is implemented to return EPERM rather than EOPNOTSUPP. pread-pwrite.patch: ------------------- This fixes a bug in a patch that I submitted earlier to provide a simple alternate AIO implementation. It defines _GNU_SOURCE in a limited area for dbpf so that we can get proper definitions of pread() and pwrite() on Linux. I tried using _XOPEN_SOURCE=500, but it will break any .c file that includes dbpf.h due to incompatibilities with Berkeley DB. zero-dfile.patch: ------------------- This fixes a bug in the getattr handling on pvfs2-server if it happens to find an attribute structure with the dfile array zeroed out. In this case, it needs to set the attr flag appropriately to prevent the response encoder from segfaulting while processing the array in the response structure. This condition is very hard to trigger, but the server should be able to gracefully report the error rather than crashing. bind-specific.patch: -------------------- This patch adds a new config file option (TCPBindSpecific) that if enabled tells the server to bind only to its specific IP address rather than using INADDR_ANY. This is particularly helpful in failover scenarios where you would like one physical machine to assume two ip addresses (and two servers) to run simultaneously when another server crashes. Without this patch you would need to select a different port on each server to prevent collisions.
1.106 annotated / raw | Diffs: previous, other | Lines: 3668 ( +39, -22 ) Created: 2006-08-22 11:41:14 -0400 (2 years 4 months ago) | Author: vilayann | Changeset: MAIN:vilayann:20060822154113 Added bmi unexpected free callbacks to tcp, gm and ib to indicate when unexpected buffersare done. All they do right now in these methods is to free these buffers. Move the call to bmi_unexpected_free() earlier (right after the decode is done) Cleaned up test programs and server to use the new api.. Also cleaned up tcp and gm bmi structures to init structures the c standard way.. Branch point for: kunkel-alternativeIO-branch kunkel-hint-branch kunkel-migration-branch threaded-aio-branch Tags: kunkel-hint-branch$BP posix-extensions-point6 threaded-aio-branch$BP kunkel-alternativeIO-branch$BP kunkel-branches-point threaded-aio-point kunkel-migration-branch$BP 1.105 annotated / raw | Diffs: previous, other | Lines: 3651 ( +15, -16 ) Created: 2006-08-11 15:18:05 -0400 (2 years 4 months ago) | Author: slang | Changeset: MAIN:slang:20060811191804 [phil]: alternative lio_listio implementation that spawns a thread and does pwrite/pread.[phil]: bmi-socket-close: This fixes a bug in the new BMI_set_info(... BMI_TCP_CLOSE_SOCKET ...) mechanism, which is used to reconnect the socket to the initial configuration server if new socket buffer sizes are specified in the config file. I didn't follow the code path find the exact problem, but at a high level it wasn't being thorough enough in cleaning out the old socket. This showed up when using epoll and specifying socket buffer sizes in the server configuration- in this case the client will often fail to mount with a cryptic "not a directory" error and leave some epoll() errors in the pvfs2-client.log file. I think a stale (or possibly reused) file descriptor was being left in the epoll fd set. At any rate, the fix is to use a different set of functions for tearing down the entire address etc. so that it is reconnected on the next BMI addr lookup. This path is already used by the server to discard old BMI addresses after critical errors on addresses that cannot be reconnected. It is triggered from bmi.c without entering the bmi_tcp module, so this patch also adds a check to make sure we don't bother for non-tcp methods. [phil]: bmi-test-overflow: One of the bmi bandwidth test programs was using types that might overflow if testing large enough transfers. The fix is to convert to doubles and drop in several type casts to be cautious when performing the computation that was causing trouble. [phil]: cancel-bugs: The biggest fix here is a change to the job timer code. It was performing some pointer operations in the wrong order, which could lead to job timers failing to trigger in some cases. This would prevent some operations from ever timing out. A secondary fix is a minor cleanup in BMI to catch potential race conditions in cancellation where a lock wasn't being held while checking to see if the target operation is complete. [phil]: flow-post-error: This patch adds checks in the client side I/O state machine to test for failure at post time for flow operations. This type of error is uncommon unless the flow parameters are faulty, but it should have checked anyway to be safe. [phil]: dfile-config-check: This is a safety test. The problem here is that there was no bounds checking for the DefaultNumDFiles option in the config file. This made it possible to select -1 (which in PVFS1 meant "use the default number"). In PVFS2 this number gets passed verbatim to the client and would cause malloc failures and various other odd results when used. The patch just checks at parse time to make sure the value isn't negative. [phil]: bmi-flow-logging: This patch just adds a little bit more gossip logging output to BMI and the flow protocol. In particular, it adds tag information to messages about BMI operations, and adds pointer information to the flow protocol to help match posts and completions of flows in log files. [phil]: server-freopen: This patch just removes duplicate calls to fdreopen(stdin/stdout/stderr) that were being performed on server startup when the server is run in the background. The first set should be sufficient.
1.104 annotated / raw | Diffs: previous, other | Lines: 3652 ( +13, -0 ) Created: 2006-07-14 16:46:00 -0400 (2 years 5 months ago) | Author: slang | Changeset: MAIN:slang:20060714204559 [pete] add a connection close set_info option to bmi-tcp so that we can reset the buffer sizes for the connection to the mountpoint server.[slang] set the defaults for the buffer sizes to 0 in the config file so they don't get set on new connections unless the config file requests it. Branch point for: kunkel-branch Tags: kunkel-branch$BP posix-extensions-point4 before-style-changes 1.103 annotated / raw | Diffs: previous, other | Lines: 3639 ( +1, -1 ) Created: 2006-05-30 00:53:26 -0400 (2 years 7 months ago) | Author: vilayann | Changeset: MAIN:vilayann:20060530045326 Consolidates I/O paths through VFS for read, write and readv, writev.Move duplicated get_object_type function to pvfs2-util.h Changed the type of a bmi local variable from int to bmi_size_t. Since new dspace layout makes use of comparision function, update the db-dspace "hack utility" as well Branch point for: WALT3 pvfs2-1-5-branch Tags: pvfs2-1-5-1-rc3 pvfs2-1-5-1-rc2 pvfs2-1-5-1-rc1 WALT3$BP pvfs2-1_5_1_for_bgl lockd-point2 posix-extensions-point2 pvfs2-1-5-0-rc1 pvfs2-1-5-0 pvfs2-1-5-point2 pvfs2-1-5-0-rc2 pvfs2-1-5-point pvfs2-1_5_1_for_bgl2 pvfs2-1-5-branch$BP ... more hide 1.102 annotated / raw | Diffs: previous, other | Lines: 3639 ( +8, -7 ) Created: 2006-05-25 22:53:42 -0400 (2 years 7 months ago) | Author: slang | Changeset: MAIN:slang:20060526025342 fix the BMI_set_info calls for send/recv buffer to pass in an actual pointer. Fix the unused variable warning tcp_set_info.
1.101 annotated / raw | Diffs: previous, other | Lines: 3638 ( +47, -2 ) Created: 2006-05-25 18:17:18 -0400 (2 years 7 months ago) | Author: slang | Changeset: MAIN:slang:20060525221716 * [Phil] removed db stat left over from keyval-per-db days* [Phil] added stranded bstream checking * [Phil] added comparison function for dspace db * [Phil] added db cache size and type config options * [Julian] added tcp send/recv buffer size options
1.100 annotated / raw | Diffs: previous, other | Lines: 3593 ( +0, -1 ) Created: 2006-03-31 17:49:51 -0500 (2 years 9 months ago) | Author: slang | Changeset: MAIN:slang:20060331224951 add #ifndef to turn off small io path if desiredBranch point for: dbpf-coalesce-branch lockd-branch posix-extensions-branch Tags: posix-extensions-branch$BP posix-extensions-point before-dbpf-onekeyval dbpf-coalesce-branch$BP lockd-branch$BP dbpf-coalesce-point lockd-point 1.99 annotated / raw | Diffs: previous, other | Lines: 3594 ( +1, -1 ) Created: 2005-12-20 19:23:07 -0500 (3 years ago) | Author: robl | Changeset: MAIN:robl:20051221002306 fix some warningsBranch point for: dbpf-onekeyval-branch pvfs2-1-4-0-branch Tags: pvfs2-1-4-0 dbpf-onekeyval-point dbpf-slowskip pvfs2-1-4-0-branch$BP dbpf-onekeyval-branch$BP pvfs2-1-4-0-point 1.98 annotated / raw | Diffs: previous, other | Lines: 3594 ( +7, -1 ) Created: 2005-12-14 16:50:21 -0500 (3 years ago) | Author: slang | Changeset: MAIN:slang:20051214215017 merging small I/O changes to trunk. The majority of this commit includes thenew small I/O state machines on client and server, and the modifications to the client IO state machine. It also includes mods to the zero-fill code to perform zero-fill of holes only if they exist instead of zero-ing the entire memory region initially. Also, some debug stuff has been added to the state machine processing code to print enter/exit of state machines if sm debug is enabled. The genconfig now includes ioports/metaports for running multiple servers on the same host for testing. Branch point for: dbpf-changes-branch Tags: dbpf-changes dbpf-changes-branch$BP 1.97 annotated / raw | Diffs: previous, other | Lines: 3588 ( +8, -3 ) Created: 2005-11-09 20:27:03 -0500 (3 years 1 month ago) | Author: slang | Changeset: MAIN:slang:20051110012652 * malloc.h is obsolete in freebsd, check for it* No error.h in freebsd * linux allows Lu,Lx, and Ld to be interchangeable with llu,llx, and lld. Other platforms (solaris, bsd) don't. Replaced all the L* with ll*. * bsd doesn't have mntent.h. use fstab.h and our own funcs. * bsd doesn't know about ssize_t. Use size_t instead. * bsd doesn't have asm errnos. Check for each of those errno macros and if not defined define to -1 * check for db3/db.h db4/db.h header if db path is specified using --with-db * darwin has aio support builtin, doesn't need -lrt * remove kernel stuff in pvfs2-types.h that sets WORDSIZE. gcc defines __LP64__ on archs with 64bit longs so we use that instead. * darwin's open doesn't accept O_LARGEFILE. off_t is an int64 by default. define O_LARGEFILE if it hasn't been. * need to run ranlib on darwin after libpvfs2.a gets installed to regenerate table of contents Branch point for: pvfs2-1-3-1-branch Tags: pvfs2-1-3-2 pvfs2-1-3-1-branch$BP pvfs2-1-3-1 1.96 annotated / raw | Diffs: previous, other | Lines: 3583 ( +254, -8 ) Created: 2005-10-25 14:00:58 -0400 (3 years 2 months ago) | Author: vilayann | Changeset: MAIN:vilayann:20051025180054 Attached patches implements trusted ports and networks similar towhat PVFS1 implemented. It adds a new <Security> </Security> tag to the fs.conf config file and parse utilities accordingly. We can edit these later if there is a strong need. Within these tags, you have to specify TrustedPorts <Range of ports> and TrustedNetwork tcp://<network> tcp://<mask>. I dont know how to support this cleanly on non tcp/ip bmi architectures, so currently this only enables it for the bmi_tcp module alone. (i guess we could do it for bmi_ib module as well but not done until someone says that is the right thing to do...) Also adds support for a --trusted argument to the pvfs2-genconfig script that will prompt for the trusted port ranges and the network address/masks and write it out to the fs.conf and server.conf file. Configure now supports a --enable-trusted to enable all these features. (NOTE that the fs.conf tags (or lack of one) still allow you to not select these features even if configure was asked to compile with this option). this patch also cleans up all the warnings that arises with gcc (3.4) and --enable-strict build mode. Branch point for: slang-sio-branch Tags: slang-sio-point slang-sio-branch$BP 1.95 annotated / raw | Diffs: previous, other | Lines: 3337 ( +60, -11 ) Created: 2005-09-29 16:15:43 -0400 (3 years 3 months ago) | Author: robl | Changeset: MAIN:robl:20050929201543 [pcarns]: reverse name lookup patch. --enable-rev-hostname if you wanthostnames instead of ip addresses
Tags: pvfs2-1-3-0 1.94 annotated / raw | Diffs: previous, other | Lines: 3288 ( +1, -1 ) Created: 2005-08-02 15:54:11 -0400 (3 years 5 months ago) | Author: pw | Changeset: MAIN:pw:20050802195411 make it compile in both casesBranch point for: slang-reqsched-branch Tags: slang-reqsched-branch$BP slang-reqsched-point 1.93 annotated / raw | Diffs: previous, other | Lines: 3288 ( +45, -0 ) Created: 2005-08-02 13:56:13 -0400 (3 years 5 months ago) | Author: robl | Changeset: MAIN:robl:20050802175610 [pcarns]: add a new bmi method that can resolve connections from unexpectedrequests into hostnames
1.92 annotated / raw | Diffs: previous, other | Lines: 3243 ( +8, -7 ) Created: 2005-07-13 11:45:56 -0400 (3 years 5 months ago) | Author: walt | Changeset: MAIN:walt:20050713154552 fixed error reporting for gethostbyname in BMI-TCPalso fixed simple omissions and errors in related code Branch point for: slang-attr-cache-fixups-branch Tags: pvfs2-1-2-0 slang-attr-cache-fixups-point slang-attr-cache-fixups-branch$BP 1.91 annotated / raw | Diffs: previous, other | Lines: 3242 ( +19, -20 ) Created: 2005-07-07 17:47:48 -0400 (3 years 6 months ago) | Author: robl | Changeset: MAIN:robl:20050707214748 [pcarns]: rework some of bmi-tcp's error messages. use _err instead of _lerr
1.90 annotated / raw | Diffs: previous, other | Lines: 3243 ( +8, -3 ) Created: 2005-03-28 13:33:43 -0500 (3 years 9 months ago) | Author: pw | Changeset: MAIN:pw:20050328183343 report restartable transfers as debug messages, not errors; make up an errno when one is not reported (other side closed)Branch point for: WALT1 WALT2 pvfs2-bugfix-67 pvfs2-tag-bugfix-67 slang-event-changes-branch Tags: pvfs2-1-1-0 pvfs2-tag-bugfix-67$BP slang-event-changes-branch$BP WALT1$BP WALT2$BP pvfs2-bugfix-67$BP slang-event-changes-point 1.89 annotated / raw | Diffs: previous, other | Lines: 3238 ( +0, -1 ) Created: 2004-09-29 15:16:06 -0400 (4 years 3 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040929191604 stub in just a tiny bit of the bmi gm cancel function and fix a few oldcomments Branch point for: pvfs2-bugfix-39 Tags: pvfs2-snapshot-Alejandro pvfs2-0-9-0 pvfs2-1-0-0 pvfs2-0-7-0 pvfs2-bugfix-39$BP pvfs2-1-0-1 snapshot-bluegene-20050216 pvfs2-0-8-0 1.88 annotated / raw | Diffs: previous, other | Lines: 3239 ( +5, -1 ) Created: 2004-08-19 11:09:58 -0400 (4 years 4 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040819150956 - added an optional socket management component that uses epoll() instead ofpoll() for its work cycle (epoll is a linux 2.6 specific alternative to poll that should result in much lower latency and better scalability) - added a --enable-epoll configure option to enable it - fixed a bug in configure.in; AC_PROG_CC was being called too late, which made it possible for several earlier tests to (if triggered) accidentally discard the default CFLAGS that we would normally expect - epoll implementation needs more testing (both for correctness and performance)
Tags: pvfs2-0-6-5 pvfs2-0-6-3 pvfs2-0-6-4 1.87 annotated / raw | Diffs: previous, other | Lines: 3235 ( +9, -0 ) Created: 2004-08-10 13:30:39 -0400 (4 years 4 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040810173039 comment out an optimization that seems to be causing a deadlock under heavyload
Tags: pvfs2-0-6-2 1.86 annotated / raw | Diffs: previous, other | Lines: 3226 ( +26, -10 ) Created: 2004-08-01 19:44:27 -0400 (4 years 5 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040801234427 response time optimizations for bmi_tcp; helps to stabilize metadataoperation latency (but there is still more work to do on this front) - avoid poll() system call in testcontext and testunexpected if there are ops already completed - also avoid poll() system call in testcontext if there are unexpected ops ready but no context ops ready (on the assumption that a testunexpected call will occur soon to retrieve them, so we shouldn't sleep unecessarily).
1.85 annotated / raw | Diffs: previous, other | Lines: 3210 ( +1, -1 ) Created: 2004-07-28 10:32:40 -0400 (4 years 5 months ago) | Author: neill | Changeset: MAIN:neill:20040728143227 - applied Nathan's expandtab patch (expandtab-2.patch.gz)first referenced at: http://www.beowulf-underground.org/pipermail/pvfs2-developers/2004-July/000745.html
Tags: pvfs2-0-6-1 1.84 annotated / raw | Diffs: previous, other | Lines: 3210 ( +3, -1 ) Created: 2004-07-12 17:12:41 -0400 (4 years 5 months ago) | Author: neill | Changeset: MAIN:neill:20040712211239 - add some error handling; null out ptrs that are no longer in use(fixes several crash bugs found in the job expiration case); add some sanity checks, etc - make server continue processing, rather than abort on BMI_testcontext errors (these are recoverable)
Tags: pvfs2-0-6-0 1.83 annotated / raw | Diffs: previous, other | Lines: 3208 ( +28, -3 ) Created: 2004-07-08 12:17:09 -0400 (4 years 5 months ago) | Author: neill | Changeset: MAIN:neill:20040708161703 - merging in the pvfs2-nm-nb-branch with the main treesee ChangeLog for details, or browse the cvs history of the branch for full details
1.82 annotated / raw | Diffs: previous, other | Lines: 3183 ( +1, -7 ) Created: 2004-05-13 08:10:33 -0400 (4 years 7 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040513121033 cleaned out some old cruft, pointed out by Christoph KlingBranch point for: pvfs2-nm-nb-branch Tags: pvfs2-0-6-0pre1 pvfs2-0-5-1 pvfs2-nm-nb-branch$BP 1.81 annotated / raw | Diffs: previous, other | Lines: 3189 ( +8, -0 ) Created: 2004-04-29 11:23:22 -0400 (4 years 8 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040429152322 bmi_tcp now safely destroys failed address references which cannot bereconnected
Tags: pvfs2-0-5-0 1.80 annotated / raw | Diffs: previous, other | Lines: 3181 ( +25, -1 ) Created: 2004-04-29 09:59:58 -0400 (4 years 8 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040429135958 - started hacking on a way to let servers release bmi resources for addressesthat cannot be reconnected; commented out for now until I have more time to work on it - also a bug fix on socket cleanup path, making sure not to call close on a socket that is already set to -1
1.79 annotated / raw | Diffs: previous, other | Lines: 3157 ( +0, -11 ) Created: 2004-04-28 19:14:40 -0400 (4 years 8 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040428231440 on second thought, handle reference counting at the bmi level rather thanhanding it off to individual modules
1.78 annotated / raw | Diffs: previous, other | Lines: 3168 ( +1, -0 ) Created: 2004-04-28 17:22:52 -0400 (4 years 8 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040428212252 server now reference counts bmi addresses
1.77 annotated / raw | Diffs: previous, other | Lines: 3167 ( +10, -1 ) Created: 2004-04-28 17:10:58 -0400 (4 years 8 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040428211058 added bmi hint to reference count addresses; not doing anything with it yet
1.76 annotated / raw | Diffs: previous, other | Lines: 3158 ( +2, -2 ) Created: 2004-04-28 14:08:15 -0400 (4 years 8 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040428180815 convert bmi and flow cancel cases to use new non-errno error code,PVFS_ECANCEL, instead of the confusing PVFS_EINTR. Will result in error messages like this: PVFS_sys_lookup: Operation canceled (possibly due to timeout)
1.75 annotated / raw | Diffs: previous, other | Lines: 3158 ( +12, -10 ) Created: 2004-04-27 16:54:00 -0400 (4 years 8 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040427205400 bug fix: allow recvs to get queued on client side even if addr is in failuremode (I didn't think this case was reachable but apparently it is :)
1.74 annotated / raw | Diffs: previous, other | Lines: 3156 ( +1, -2 ) Created: 2004-04-27 12:37:32 -0400 (4 years 8 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040427163732 removed a write only variable :)
1.73 annotated / raw | Diffs: previous, other | Lines: 3157 ( +48, -26 ) Created: 2004-04-27 10:49:44 -0400 (4 years 8 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040427144944 some changes so that _client_ side bmi_tcp addresses have a chance toreconnect; probably needs more testing. a few other little cleanups along the way.
1.72 annotated / raw | Diffs: previous, other | Lines: 3135 ( +3, -3 ) Created: 2004-04-26 13:58:35 -0400 (4 years 8 months ago) | Author: neill | Changeset: MAIN:neill:20040426175834 - make all of bmi use the id-safe-register methods instead of the fast onesto disassociate the ids from the data
1.71 annotated / raw | Diffs: previous, other | Lines: 3135 ( +67, -53 ) Created: 2004-03-09 12:30:04 -0500 (4 years 9 months ago) | Author: neill | Changeset: MAIN:neill:20040309173004 - cleanup the error codes in bmi by translating them and making themmore pvfs friendly. this only fixes the generic bmi and bmi-tcp parts, as the other modules will need some similar translation - bug fix that will not add an invalid socket if detected
Tags: pvfs2-0-1-3 pvfs2-0-1-2 1.70 annotated / raw | Diffs: previous, other | Lines: 3121 ( +0, -1 ) Created: 2004-03-03 16:36:00 -0500 (4 years 10 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040303213600 warning fix
Tags: pvfs2-0-1-1 1.69 annotated / raw | Diffs: previous, other | Lines: 3122 ( +34, -18 ) Created: 2004-03-03 16:32:16 -0500 (4 years 10 months ago) | Author: neill | Changeset: MAIN:neill:20040303213216 - try to not bungle up the zero read limit check so badly (take 2)- reset on valid data reads - handle the other case where we need to make this check
1.68 annotated / raw | Diffs: previous, other | Lines: 3106 ( +13, -0 ) Created: 2004-03-03 12:33:26 -0500 (4 years 10 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040303173326 Threw away old socket collection code, replaced with completely newimplementation. Quite a few improvements: - cleaner - better thread safety - O(1) updates of poll set - O(1) socket <-> bmi address mapping - no long queue searches Probably introduced some bugs along the way, so let me know if you see anything fishy...
1.67 annotated / raw | Diffs: previous, other | Lines: 3093 ( +27, -5 ) Created: 2004-03-02 16:16:47 -0500 (4 years 10 months ago) | Author: neill | Changeset: MAIN:neill:20040302211647 - give the socket a few chances before assuming it's a dropped connectionwhen read returns 0 after poll says we're ready to read
1.66 annotated / raw | Diffs: previous, other | Lines: 3071 ( +0, -1 ) Created: 2004-03-02 15:19:12 -0500 (4 years 10 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040302201912 fix typo
1.65 annotated / raw | Diffs: previous, other | Lines: 3072 ( +19, -1 ) Created: 2004-03-02 14:58:55 -0500 (4 years 10 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040302195855 handling yet another odd socket error condition- if you find that thereis no data available to read after poll() has told you that there is, then it should be interpretted as a closed socket. Fixes some server side hangs in which some write flows never completed because they did not detect the death of a client that was trying to send data
1.64 annotated / raw | Diffs: previous, other | Lines: 3054 ( +2, -1 ) Created: 2004-02-17 18:58:55 -0500 (4 years 10 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040217235855 fixed a list operation handling bug that snuck in last week
Tags: pvfs2-0-1-0 1.63 annotated / raw | Diffs: previous, other | Lines: 3053 ( +5, -5 ) Created: 2004-02-16 16:22:41 -0500 (4 years 10 months ago) | Author: rbross | Changeset: MAIN:rbross:20040216212238 combo PVFS_id_gen_t -> PVFS_BMI_addr_t, formatting, PVFS error code patch.Ugly. Sorry if my formatting pisses someone off; at least I'm not using > 80 columns any more :).
1.62 annotated / raw | Diffs: previous, other | Lines: 3053 ( +1, -1 ) Created: 2004-02-14 18:20:40 -0500 (4 years 10 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040214232040 demoted an error code
1.61 annotated / raw | Diffs: previous, other | Lines: 3053 ( +1, -1 ) Created: 2004-02-12 18:25:28 -0500 (4 years 10 months ago) | Author: rbross | Changeset: MAIN:rbross:20040212232528 Spelling fixes, misc. stuff.
1.60 annotated / raw | Diffs: previous, other | Lines: 3053 ( +17, -0 ) Created: 2004-02-12 17:50:12 -0500 (4 years 10 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040212225012 mark persistent errors on bmi_tcp addresses that have failed in the past
1.59 annotated / raw | Diffs: previous, other | Lines: 3036 ( +5, -1 ) Created: 2004-02-12 17:30:38 -0500 (4 years 10 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040212223038 tightening the bmi_tcp error handling a bit, still needs work
1.58 annotated / raw | Diffs: previous, other | Lines: 3032 ( +2, -1 ) Created: 2004-02-11 14:49:48 -0500 (4 years 10 months ago) | Author: neill | Changeset: MAIN:neill:20040211194948 - this nearly trivial fix is the culmination of probably 5 hours of pcarn'sand neillm's time >:\ - if you're running in an nptl configuration, be prepared for major performance boosts :-D - drop mutex while sleeping in case of test context where no work can be done
1.57 annotated / raw | Diffs: previous, other | Lines: 3031 ( +61, -15 ) Created: 2004-02-10 17:52:55 -0500 (4 years 10 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040210225255 important fix for bmi_tcp; detect when we have a big backlog of data onthe sockets being watched, and throttle back if we can't make any progress on that backlog. We now give up more cpu time in hopes of the rest of the system catching up. Should make a big difference on systems where the tcp/ip net is much faster than the disk (like on big writes to localhost)
1.56 annotated / raw | Diffs: previous, other | Lines: 2985 ( +71, -104 ) Created: 2004-02-10 10:19:18 -0500 (4 years 10 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040210151918 merged in two bmi_tcp optimizations to reduce number of system calls:send msg envelope along with data payload in the same writev() call, and assume default socket mode is nonblocking rather than querying/setting mode on every socket call.
1.55 annotated / raw | Diffs: previous, other | Lines: 3018 ( +18, -13 ) Created: 2004-01-30 15:12:12 -0500 (4 years 11 months ago) | Author: neill | Changeset: MAIN:neill:20040130201209 - rename all gossip masks to GOSSIP_xxx- some misc cleanups
Tags: pvfs2-0-0-9 1.54 annotated / raw | Diffs: previous, other | Lines: 3013 ( +3, -3 ) Created: 2004-01-30 15:05:01 -0500 (4 years 11 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040130200500 demoted some gossip_lerr messages to gossip_err
1.53 annotated / raw | Diffs: previous, other | Lines: 3013 ( +58, -1 ) Created: 2004-01-21 18:52:49 -0500 (4 years 11 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040121235249 implemented cancel() for BMI TCP module, untested
1.52 annotated / raw | Diffs: previous, other | Lines: 2956 ( +2, -1 ) Created: 2004-01-20 20:34:57 -0500 (4 years 11 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20040121013456 hooks to start tinkering with a cancel operation function in BMI
1.51 annotated / raw | Diffs: previous, other | Lines: 2955 ( +2, -1 ) Created: 2003-12-09 16:11:41 -0500 (5 years ago) | Author: pcarns | Changeset: MAIN:pcarns:20031209211141 step two in nptl cleanup; this is a hack that prevents bmi-tcp fromholding global mutex while sleeping in poll(). Needs to be cleaned up and may even just be reverted later, depending on outcome of some more experiments. Next step is to fix similar (but easier to fix cleanly) problem in thread-mgr mutex locking...
Tags: pvfs2-0-0-8 pvfs2-0-0-7 1.50 annotated / raw | Diffs: previous, other | Lines: 2954 ( +59, -0 ) Created: 2003-12-09 14:24:40 -0500 (5 years ago) | Author: pcarns | Changeset: MAIN:pcarns:20031209192439 step one of some lock reorganization- get rid of "one big mutex" thatserializes all BMI calls, and put burden of thread safety on individual modules. for now this means that each module has its own big mutex to serialize things, but this can be optimized on a per module basis later. this is a necessary step towards making pvfs2 work under nptl; the fact that we hold mutex locks for too long seems to exacerbate starvation problems in that environment...
1.49 annotated / raw | Diffs: previous, other | Lines: 2895 ( +38, -24 ) Created: 2003-12-09 11:05:27 -0500 (5 years ago) | Author: pcarns | Changeset: MAIN:pcarns:20031209160527 took Pete's idea about how to get an errno out of a socket that hasexperienced a poll error, and extended it to propigate the error code to all pending messages for that socket (if more than one is queued)
1.48 annotated / raw | Diffs: previous, other | Lines: 2881 ( +10, -10 ) Created: 2003-11-04 11:03:10 -0500 (5 years 2 months ago) | Author: rbross | Changeset: MAIN:rbross:20031104160310 Prefixed external socket collection functions with BMI_.Branch point for: pvfs2-dbpf-attr-cache-branch Tags: pvfs2-0-0-6 pvfs2-dbpf-attr-cache-branch$BP 1.47 annotated / raw | Diffs: previous, other | Lines: 2881 ( +8, -8 ) Created: 2003-10-06 09:00:21 -0400 (5 years 3 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20031006130021 added event logging macros to read_list and write_list in dbpf (thoughthey do not record sizes yet), adjusted some other even macros here and there
Tags: snapshot-kraken-10-16-2003 pvfs2-0-0-5 pvfs2-0-0-4 snapshot-kraken-10-16-2003-B 1.46 annotated / raw | Diffs: previous, other | Lines: 2881 ( +67, -8 ) Created: 2003-10-02 12:10:25 -0400 (5 years 3 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20031002161024 instrumented bmi_tcp code
1.45 annotated / raw | Diffs: previous, other | Lines: 2822 ( +21, -21 ) Created: 2003-09-09 12:12:47 -0400 (5 years 3 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20030909161246 renamed some internal BMI functions to avoid collisions w/ pvfs1 name space
Tags: pvfs2-0-0-3 1.44 annotated / raw | Diffs: previous, other | Lines: 2822 ( +44, -39 ) Created: 2003-08-27 07:34:35 -0400 (5 years 4 months ago) | Author: pw | Changeset: MAIN:pw:20030827113434 Change BMI send functions to accept const buffer pointers.Change BMI list functions to promise not to modify the lists themselves. Some casts in BMI methods are to stick non-const pointers in const pointer storage to permit the use of generic functions for both send and recv.
Tags: pvfs2-0-0-2 1.43 annotated / raw | Diffs: previous, other | Lines: 2817 ( +1, -0 ) Created: 2003-08-13 10:22:51 -0400 (5 years 4 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20030813142251 fixed a subtle bug in BMI- was losing context_id in some corner cases, couldcause BMI to hang under the right conditions
Tags: pvfs2-0-0-1 1.42 annotated / raw | Diffs: previous, other | Lines: 2816 ( +4, -4 ) Created: 2003-08-06 13:40:40 -0400 (5 years 5 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20030806174040 made message size error codes more useful
1.41 annotated / raw | Diffs: previous, other | Lines: 2816 ( +17, -0 ) Created: 2003-08-06 13:34:37 -0400 (5 years 5 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20030806173437 added more gossip_err statements to help narrow down problems in the sendpath
1.40 annotated / raw | Diffs: previous, other | Lines: 2799 ( +4, -5 ) Created: 2003-07-23 17:14:43 -0400 (5 years 5 months ago) | Author: pcarns | Changeset: MAIN:pcarns:20030723211443 bugfix: switch case was falling through when it shouldn't have...
|