summaryrefslogtreecommitdiff
path: root/source3/smbd
AgeCommit message (Collapse)AuthorFilesLines
2007-10-10r974: Fix open code to pass more torture tests. We now pass the Samba4Jeremy Allison1-28/+87
oplock test. We do this be changing the algorithm when breaking oplocks slightly. Previously we broke an oplock, then re-loaded the share modes and re-iterated. Now we break all oplocks, then re-load the share modes and check the share details - then iterate. This seems to match the way Win2k3 does it. Jeremy. (This used to be commit 5ec4f4e4e6596ea0f52aca5c9e1a75bf35612e37)
2007-10-10r971: Auto remove store dos attributes if underlying filesystemJeremy Allison2-11/+26
doesn't support EA's. Jeremy. (This used to be commit 9de6b25c9e9abe982e21b0229df520802cafbfd8)
2007-10-10r957: Fixup moribund unix extensions mknod code path. Don't add inJeremy Allison1-2/+3
configure test yet. Also allow unix domain socket create. Jeremy. (This used to be commit 2db0c4ca7b23cdd1dbf83d4de58acbf4dc118df2)
2007-10-10r945: "Correct" (as in, more correct) way to handle stat opens. Doesn't regressJeremy Allison1-22/+25
the torture tester. Passes OPEN tests in Samba3 and Samba4 and oplock tests in Samba4. Last thing to check, can an "attribute only" open actually create a file. I think it can.... Jeremy. (This used to be commit d8177a42d4a2a2b6df5d0593f2f92bddc29ffe94)
2007-10-10r942: If using DOS error codes, we need to return ERRDOS/ERRbadpath for ↵Volker Lendecke1-1/+1
chkpth even if only the last component failed. I'm not sure if all the other cases of NT_STATUS_OBJECT_NAME_NOT_FOUND also need to be fixed, this at least helps with 'copy test1.txt test2.txt' from DOS within a subdirectory. Yes, I do have someone who needs this :-) Jeremy, could you take a look at this? Thanks, Volker (This used to be commit 3093ab100653782bd8e029170d315c68b7f271af)
2007-10-10r933: When using widelinks = no, use realpath to canonicalize theJeremy Allison3-2/+41
connection path on connection create for the user. We'll be checking all symlinked paths are below this directory. Jeremy. (This used to be commit b562fe9fbca4971059b913959bbaca02af42c1a4)
2007-10-10r931: Ensure we push 16 bytes (including null termination)Jeremy Allison1-2/+2
not 15. (This used to be commit 1d3fd1a58e0a8b08fe028ceee03618180c7d4b97)
2007-10-10r925: add changes frpm trunk (r841 and r842) -- enable background queue ↵Gerald Carter1-6/+7
update process and allow printers to have different sharenames from printernames (This used to be commit 066b9c4276a968788a03709a00d4f672ac032df7)
2007-10-10r814: conn is 0 during ioctl (at least during smbtorture IOCTL test)Herb Lewis2-2/+4
fix smbd panic (This used to be commit 61da0bb1f6ea22df78ea7bd22a740c5868f62591)
2007-10-10r799: BUG 1259 -- add 'printcap cache time' patch from LarsGerald Carter2-1/+51
(This used to be commit fac90741139b953d0e88d050dd457657f0b9c9f3)
2007-10-10r786: Memory leak fixes in (mostly) error code paths fromJeremy Allison2-2/+10
kawasa_r@itg.hitachi.co.jp. A couple of mem leak fixes in mainline code paths though :-). Jeremy. (This used to be commit 4695cc95fe576b6da0d0cb0686f208fc306b2646)
2007-10-10r779: Fix specific case of open that doesn't cause oplock break, or share ↵Jeremy Allison2-7/+22
mode check. Test case provided by Volker will be added later. There may be other tests needed. Jeremy. (This used to be commit 1f8e7946edbb2930ba5f9738688dc843bc368fbf)
2007-10-10r762: Fix for #1319 when security > share.Jeremy Allison1-2/+8
Jeremy. (This used to be commit 9fe2240d6b68a2f8a495df585d69ae20c9825d77)
2007-10-10r742: Volker discovered that we can truncate a locked file. Well I never ! :-).Jeremy Allison1-22/+2
Jeremy (This used to be commit 1099d60113c2d7b2e5a27a10666cdf885c668778)
2007-10-10r740: Fix Bug #1301. Return NT_STATUS_SHARING_VIOLATION when share mode lockingRichard Sharpe1-0/+14
requests fail. (This used to be commit 5082eb30d0d7c3a4c79b7578457682523a4e80a3)
2007-10-10r716: Fix for bugid #1345 (Macromedia Homesite cannot connect anymore after ↵Jeremy Allison1-0/+4
upgrade to 3.0.4). Cause was premature optimization in unix_convert(). My fault, sorry. Jeremy. (This used to be commit e5438f0a841e8ea57b515544cbf638c38405ed55)
2007-10-10r710: Fix smbclient symlink command when widelinks = no.Jeremy Allison1-70/+7
Jeremy. (This used to be commit ed699a73f899b3212185ecf52d1c4cd59cc8a7ea)
2007-10-10r706: Fix erroneous commit... I've not finished testing theJeremy Allison1-9/+3
symlink stuff yet. Noticed by Stefan (metze). Jeremy. (This used to be commit 389ad1a05b2197edb20eda270fbd5c7eed7bb2f3)
2007-10-10r698: Now wb pipe is non-blocking remember to read in non-blocking mode...Jeremy Allison1-2/+8
Jeremy. (This used to be commit 3399727864f3aa8981f022254dfed622fcb50c49)
2007-10-10r658: Oops - make smbclient work again with widelinks = no :-).Jeremy Allison2-6/+8
Jeremy. (This used to be commit a6979417063e0f0067004dc6207e8e78aa41a58f)
2007-10-10r656: Make widelinks use realpath(). Tidy up cases where we need to become a ↵Jeremy Allison5-148/+105
service. Jeremy. (This used to be commit a03b6a05e02ec8415efc0e8ceade102e06f8fffe)
2007-10-10r654: There is no point in having two lines of:Richard Sharpe1-1/+1
sbuf.st_mtime &= ~1; So I had one refer to st_ctime. (This used to be commit cf7a64c05dbc281802a2dc97a182b530fb925db9)
2007-10-10r653: Ok, so tridge convinced me. Allow full UNIX path symlinksJeremy Allison1-26/+25
to anywhere on the server filesystem so long as widelinks is set to true.... :-). Jeremy. (This used to be commit ba9809fcd493a53b33c3a24c9f91b0c1d9bd9726)
2007-10-10r645: Patch from kawasa_r@itg.hitachi.co.jp to correctly enable core dumps.Jeremy Allison1-0/+4
Jeremy. (This used to be commit ea41d694270264557f740cd40ccc69b4acaa57e9)
2007-10-10r640: Make cifsvfs symlinks work with paths within and external to theJeremy Allison1-7/+41
share. Store external paths prefixed with smbln:. Jeremy. (This used to be commit 14a9997b2eb394174ccb36d2a10a755279740cba)
2007-10-10r570: Remove lots of globals to handle case issues - move themJeremy Allison9-100/+70
to connection struct entries (as they should have been from the start). Jerry, once you've cut over to 3.0.4 release branch I'll add this to 3.0 also. - Jerry cut over :-). Jeremy. (This used to be commit 578a508509d21226ad3332fc54c3ab54cd8ae452)
2007-10-10r519: fix bug in authorise_login() that broke security = shareGerald Carter1-0/+2
(This used to be commit b27192d8e7a4a1fd0cd737b4fd9190bcecfbb507)
2007-10-10r486: BUG 1309: fix seg fault caused by trying to strdup() a NULL pointerGerald Carter1-1/+5
(This used to be commit 0e1a6b34f2690da32bf568330575651988d0c965)
2007-10-10r483: Fix typo of appending '\\' instead of '/' now we're usingJeremy Allison1-2/+2
check_path_syntax. Jeremy. (This used to be commit 50ce579322ae394cb7c7a27e06f4ff08499b77b3)
2007-10-10r481: Fix up assert caught by check_path_syntax.Jeremy Allison1-2/+3
Jeremy. (This used to be commit 8a58bd39da596cac68abc0143cdb17981642daf6)
2007-10-10r478: Added Volkers fix to be able to force DOS errors when needed.Jeremy Allison2-5/+5
Jeremy. (This used to be commit a9d1738ebab42ab9bab73f18341d79e086e290b3)
2007-10-10r428: add acls debug classHerb Lewis1-0/+3
(This used to be commit b7703799f8899affda205eacb0bf79cf8e2b9362)
2007-10-10r422: Compile fix for OpenBSD (ENOTSUP not supported - ha ha).Tim Potter1-0/+2
From Eric Mertens. (This used to be commit d98a36b4596d1de979adab6c476022eb95474c48)
2007-10-10r421: Fix debug message. Closes bugzilla #1305.Tim Potter1-1/+1
(This used to be commit 6654b03cbd08e8bbf3f9cae117b84a820c8857b2)
2007-10-10r410: merge tpot's changeset 353 from trunk:Stefan Metzmacher1-0/+1019
Move msdfs.c into the source/smbd directory and remove source/msdfs. metze (This used to be commit 88e6e6d29ca14e0cddbc1df611051a96568dc0c9)
2007-10-10r408: - replace (conn->admin_user != True) with (current_user.uid != 0)Stefan Metzmacher3-4/+4
because someone changed it in all other places too - fix quotas support from windows explorer we now got the unix file name of a fake_file metze (This used to be commit 87e97d7723674e3835578ef080ce554d9c5537ac)
2007-10-10r370: The 'it does never happen -- error on close()' does happen when you exceedVolker Lendecke1-1/+1
your quota on an AFS file system. The specific errno was thrown away by close_normal_file(). Thus we returned NT_STATUS_UNSUCCESSFUL and not NT_STATUS_DISK_FULL as we should. Fix that. (Not that this gives more sane Windows app behaviour.... :-( ) Jerry, jra, could you please look over this one, it's been quite a while since I touch file server code. Volker (This used to be commit 588351d1b37a900eb90eddd80882170c201e9162)
2007-10-10r283: removing --with-cracklib after discussion with abartklet @ sambaXPGerald Carter1-47/+0
(This used to be commit df94b0471eb0628aa27f534134d60b62ed123688)
2007-10-10r247: Fixup a couple of debug messages.Jeremy Allison1-2/+2
Jeremy. (This used to be commit c664a277b428c036510966734ba71e43112c1cac)
2007-10-10r225: Patch from Pat.Hayward@propero.net to make the session_users list dynamic.Jeremy Allison2-9/+29
I restricted it to 128k max to prevent DOS attacks. Jeremy. (This used to be commit 70fb2a196d83c4bde11d27608da27f956f3f19b8)
2007-10-10r150: Make 3.x pass the Samba 4.x RAW-SEARCH tests - except for theJeremy Allison1-75/+175
UNIX info levels, and the short case preserve names. Tested with NT - needs more testing. Will work on this more on Monday. Jeremy. (This used to be commit 88a9bca9aade3ee0ce9713857becfe0a13bbfd07)
2007-10-10r96: Stupid f&%'n UNIX extensions.... SETPATHINFOJeremy Allison2-36/+39
normally takes as it's param entry the filename to be acted upon.... Unless it's UNIX extensions create hardlink, or UNIX extensions create symlink. Then it's param -> newfile name data -> oldfile name. This caused me to stuff them up in 3.0.2 (and the client commands link and symlink). Fixed them, everything is now called oldname and newname - thus specifying which name should already exist (hint - the old one...) and which will be created (newname). Jeremy. (This used to be commit 21cc6ab7e8a41160a3e2970623ade7445b5214d6)
2007-10-10r81: Make EA's case independent - fix smbtorture to test for this.Jeremy Allison1-1/+24
Jeremy. (This used to be commit 0661357484a33329c75457527416667feaf57ef9)
2007-10-10r69: Global rename of 'nt_session_key' -> 'user_session_key'. The session ↵Andrew Bartlett1-2/+2
key could be anything, and may not be based on anything 'NT'. This is also what microsoft calls it. (This used to be commit 724e8d3f33719543146280062435c69a835c491e)
2007-10-10r59: revert session key problemAndrew Tridgell1-2/+2
(This used to be commit 40b5794ae0919c6c6f1b8a451871dcc95bbab5cc)
2007-10-10r50: Fix bug 1139 as per fix suggested by jdev@panix.com,Jeremy Allison1-3/+3
swap lookups for user and group - group will do an algorithmic lookup if it fails, user won't. Jeremy. (This used to be commit a205c56a75c93c82796fd68687e8c0db26459073)
2007-10-10r49: Support SMB signing on connections using only theAndrew Bartlett1-8/+4
LANMAN password. This also corrects the 'session key' for these connections. (This used to be commit 26d8791ddedb7964c219067480cf4a7d61877765)
2007-10-10r39: * importing .cvsignore filesGerald Carter1-3/+0
* updateing WHATSNEW with vl's change (This used to be commit a7e2730ec4389e0c249886a8bfe1ee14c5abac41)
2004-04-03Fix wildcard src with wildcard dest renames.Jeremy Allison1-1/+7
Jeremy. (This used to be commit 58b39942444e62c9d0f98c2ba6f434d74c244a2b)
2004-04-03Fix wildcard unlink.Jeremy Allison1-2/+3
Jeremy. (This used to be commit 6e9b4c57ccb385acb9feaf9f7412778474231a92)