summaryrefslogtreecommitdiff
path: root/source3
AgeCommit message (Collapse)AuthorFilesLines
2000-05-10Using a structure for a tdb key can lead to insideous, hardHerb Lewis3-20/+25
to find bugs. On 64 bit IRIX, structure packing means that a struct { SMB_DEV_T dev /* 4 bytes */ SMB_INO_T ino /* 8 bytes */ } has 4 bytes of padding between the two members. If you don't null the memory before using it as a tdb key, you randomly can't find keys depending on what is in the padding. This caused me immense pain and was hard to track down.... :-) Jeremy. (This used to be commit f2a5ba3f0939f59097f0ef6a25f1cf9b5574f157)
2000-05-09Fix for check_kernel_oplocks looking for the var directoryHerb Lewis1-2/+2
before it may need to be created. Jeremy. (This used to be commit 6bf74da8b84a22aaddb0fbb160f02621abba01ba)
2000-05-09This is reall Jeremy commiting as Herb at SGI labs. FixHerb Lewis1-1/+1
for "socket operation on non-socket" error in log.smb on HEAD branch startup (server_fd not initialized to -1). Jeremy. (This used to be commit ebb4f76550062b0ea2005a2b8842641a2c91f528)
2000-05-09- use smb_gwtpwnam() in another couple of placesAndrew Tridgell2-8/+10
- don't call add/del user if the scripts are empty (This used to be commit 43860215d4d16cb1bacdc77f1c46c54e4c54abd7)
2000-05-09fixed some winbind cache bugsAndrew Tridgell1-16/+17
(This used to be commit 3cac3ccf047ce9a5c28916f9a2b1b3d38741e373)
2000-05-09the beginnings of a new scheme I've working on to allow an easierAndrew Tridgell11-300/+2018
head/tng merge. It goes something like this: - headers from tng get copied over one at a time - the old headers get renamed to *_old.h - server side code that used the old headers gets a #define OLD_NTDOMAIN 1 #undef OLD_NTDOMAIN at the start and end of the code - mkproto.awk recognises these special defines and does magic stuff so that each .c file sees the right headers - we start moving the rpc client libraries from tng to head. if this goes OK then, in theory, we should be able to move the client side rpc code from tng to head without disturbing the existing head server side code. Then when that works we can consider merging the server side. it remains to be seen if this scheme will work. So far I've moved rpc_samr.h and don't seem to have broken anything. Note this this is still a very delicate operation, as at every step of the way I want to keep head fully functional. Please don't take part unless you discuss it with me first. (This used to be commit f76c037255a6a79d11bec65e863e009a41a4f0fd)
2000-05-09don't attempt to build rpcclient in the head branchAndrew Tridgell1-1/+1
(This used to be commit 1b6d01e1152d1ac64e3b1f721227d06061982112)
2000-05-09brought across some rpc header files from tngAndrew Tridgell2-0/+112
(This used to be commit db241a0b09fc38916db3b1a30aa493910d22567b)
2000-05-09brought the winbindd code into headAndrew Tridgell14-0/+4802
this does not yet compile, but I'm working on that. (This used to be commit 3fb862531a4e78dca13d16d958517b16e5bdd4e2)
2000-05-09added some rules for winbindd and pam_winbindAndrew Tridgell1-1/+35
(This used to be commit 6c5d139844334191f78257fbc9f01d0fa0c2bcf0)
2000-05-09Added tdb_store_by_string() and tdb_fetch_by_string() functions to storeTim Potter1-0/+27
data with null terminated string keys. (This used to be commit d58146321b6fe50e1cc1a73da80c3d2e8c3412dc)
2000-05-08Fix for VMS platforms from "John E. Malmberg" <wb8tyw@qsl.net>Jeremy Allison1-0/+2
Jeremy. (This used to be commit eb281324fa409296bb3f29c9b7c59b2337fadc0d)
2000-05-08Someone :-) forgot to add secrets.c to HEAD.Jeremy Allison2-1/+91
Jeremy. (This used to be commit ac40971f30e00a93fc5e138bfd1afa05cd8dea4b)
2000-05-08added secrets.tdb and changed storage of trust account password to useAndrew Tridgell11-262/+58
it (This used to be commit 88ad00b82acc4636ab57dfe710af08ea85b82ff1)
2000-05-06added winbindd options in head branch, so it is possible to combineAndrew Tridgell1-0/+49
branches (This used to be commit cd5e2494279792f8516947a63bf313f8142a52c0)
2000-05-06fixed a vfs crash bugAndrew Tridgell1-6/+4
(This used to be commit 76da586396c85b6614579b9234eec34f57c12ffe)
2000-05-06signed/unsigned fixes so we can handle a lock base close to 2^32Andrew Tridgell1-31/+29
(This used to be commit 0ce2ca5ccdd0e2e7711dcbf66b72a183b2692cfe)
2000-05-05Fix for uninitialized memory read in brlock code. brl_locktest now needsJeremy Allison3-3/+4
to correctly set the fnum, as the brl_conflict code looks at it. Jeremy. (This used to be commit df87259a356b91989604a10b3691480dcd16cf3f)
2000-05-05Fix for dual unix_to_dos conversion in readdir from A.V.Shutko ↵Jeremy Allison1-1/+0
<AVShutko@mail.khstu.ru> Jeremy. (This used to be commit 1152f2ed008dbfa7e2b1cb9719d31575e7d05b89)
2000-05-05Fix for crash bug in OpenConfFile() if no smb.conf exists.Tim Potter1-1/+1
(This used to be commit b531ddb2657e9eac1c1e9f32344b67c3963ce30a)
2000-05-05Fix for read_file() returning -1.Jeremy Allison1-0/+2
Jeremy. (This used to be commit b899943658bee90994b416aa6caff0ee7cab5f28)
2000-05-05Two fixes. Added missong logic & case in lock split code.Jeremy Allison1-4/+12
Fixed range split into two, as DLIST_ADD has the wrong semantics... Jeremy. (This used to be commit 82681edda14dcc3d58bb303cfac5452072de67df)
2000-05-05make debug easier to readAndrew Tridgell1-6/+9
(This used to be commit c14a2616441b2f3357a081c6dbae7d5b677e5a12)
2000-05-05make debug easier to readAndrew Tridgell1-5/+6
(This used to be commit d64bb07f0ad0efa1a20015b620ae33fb082b1e44)
2000-05-05nasty hack to print posix locksAndrew Tridgell1-0/+17
(This used to be commit efc9752f1e00b38abe3b5d3a98627a9fe6122bc4)
2000-05-05example of broken posix lock behaviourAndrew Tridgell1-0/+8
(This used to be commit b3999f3b20d470b9c873b297e7aeb043da61588d)
2000-05-04rpc_server/srv_lsa.c: Bring into sync with 2.0.x.Jeremy Allison4-19/+41
rpc_server/srv_pipe_hnd.c: Bring into sync with 2.0.x. smbd/blocking.c: Improve blocking debug reporting. utils/torture.c: Added check for NT locking bug. Jeremy. (This used to be commit e8ff6d3fb5537c39611a5784bf7216ae812acd27)
2000-05-04Updated to fix overlapping problem.Jeremy Allison1-68/+142
Jeremy. (This used to be commit 9bdfe0f5023988962f8a8b4d847de7a0ee27f85c)
2000-05-04a minimal change to get appliance mode to work with winbinddAndrew Tridgell5-16/+70
we needed to accept usernames of the form DOMAIN/user, which means we needed to pass the domain to a getpwnam() like routine in certain critical spots. What I'd rather do is get rid of "char *user" everywhere and use the new userdom_struct, but that will have to wait a few days. (This used to be commit 8b7a10febead8be182e7d5b1d68259e31530b69c)
2000-05-04parameterize the lock timeoutAndrew Tridgell1-2/+3
(This used to be commit 98ad4095ccc8d0349d05e6e9223eaad057029626)
2000-05-04added a test for the NT byte range lock into smbtortureAndrew Tridgell1-1/+16
(This used to be commit 7cfd45cc56e01a20a122c7742dc7e6a4ed59415f)
2000-05-04an even simpler example of NT gettings its locking code wrong. ThisAndrew Tridgell1-20/+14
one doesn't even need two connections to the box, just two file handles. it is a very simple case actually, and one I think will happen quite a lot in real life. I wonder how they haven't noticed it? I checked and W2K has the same bug. (This used to be commit 0b335e415818028ac0daad5f99c2fd9086a2a656)
2000-05-04 - use full_name instead of real_nameAndrew Tridgell8-21/+9
- got rid of guest map code in lpq parser (This used to be commit 8e53f781d3cf6a7007764916a0d8e8f1abea1f66)
2000-05-04need LOCKING_OBJ in locktest nowAndrew Tridgell2-4/+8
(This used to be commit 89e71994450f479a0f3839f192d868924626fc84)
2000-05-04more locktest improvementsAndrew Tridgell1-31/+81
the current preset tests show a very interesting NT bug - it sometimes refuses a lock that it shouldn't. (This used to be commit 85b5dc64cf7cc80567f9069a48c4d20c10ccec01)
2000-05-04fixed message textAndrew Tridgell1-1/+1
(This used to be commit 99e522023eae7cf1a88e5de50f2b09df3ea318d3)
2000-05-04if the stat cache is off then don't initialise itAndrew Tridgell1-0/+2
(This used to be commit cd5dcd9e659a7255f03eb496c0202986375e73a6)
2000-05-04added -A analyze mode to locktest - it can now automatically pruneAndrew Tridgell1-139/+242
the list of locks that cause a failure (This used to be commit 0beee59432f17f90a1c4c2c68a7ea6022b5d0908)
2000-05-04found a much simpler case that kills the posix lockingAndrew Tridgell1-5/+5
(This used to be commit b703cecf81dc62234171eaaa4c0cae448f144e1f)
2000-05-04fixed a uninit memory read that insure foundAndrew Tridgell1-6/+3
(This used to be commit 43fe5b98b4514d59e1c4bbc54a628cfbd771fc4f)
2000-05-04added ability to present lock testsAndrew Tridgell1-16/+40
(This used to be commit 3fc6a467fc3dfd2a59d4509ec52172d2dd048c80)
2000-05-04Fix for debug statement crash.Jeremy Allison1-1/+1
Jeremy. (This used to be commit 57bf92d90147b207664152d44ce4bfb5235dc7d8)
2000-05-03Fix based on Elrond's code to reduce the store size on removing dead records.Jeremy Allison1-0/+1
Jeremy. (This used to be commit 09e92a61a966d09f543ba541ddb3240cc4718579)
2000-05-03Fixed the bug locktest.c found, it was an off-by-one error in the non-overlapJeremy Allison1-1/+1
case. Jeremy. (This used to be commit 65150c408a5254215de89c8c774c33c4d011c2c0)
2000-05-03Fixed silly bug where I wasn't checking for matching fd's in closing a file.Jeremy Allison1-7/+24
This caused smbd crashes on SIGKILL. Jeremy. (This used to be commit d4dcefd12d14df112f083c312acbea0196bc5c23)
2000-05-03Added range info to the panic message to track down the bug withJeremy Allison1-1/+5
locktest. Jeremy. (This used to be commit 91f038356b7efa04cf4bfa5e6afd8b144b6b4ad5)
2000-05-03improved the lock test programAndrew Tridgell1-15/+27
(This used to be commit a0a3931c6e5824121bf4a8d075ef24b9b70e8aa1)
2000-05-03- clear dead locks at startupAndrew Tridgell1-2/+47
- fixed a bug which caused lock records to not decrease in size (This used to be commit 51624a3a2f5a4ac55b4f26f572a2fc399f9c808e)
2000-05-03added a nasty lock testing programAndrew Tridgell2-1/+350
it opens 2 connections to each of 2 servers, and opens 2 fnums on the same file on each connection (a total of 8 file descriptors) then it does random lock/unlock/reopen requests in a 100 byte range on the file and compares the results from the 2 servers. strangely enough, NT fails this test against itself right now - I'm still trying to figure that out. (This used to be commit 2f14d7c9bca5e170b8d2c169801cb200021fd467)
2000-05-03Fix for stacking locks in brlock and POSIX. Windows only allows a read lockJeremy Allison2-92/+130
to overlay a write lock on the same fnum. When overlaying read locks onto a write lock, the number of locks is counted, and the first unlock removes the write lock and downgrades this to a read lock. Do the same when mapping to POSIX. Jeremy. (This used to be commit 74d42644e6e52808037975e909aa56c850838b76)