| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
can_set_delete_on_close() is correctly called before any setting
of the disposition bit (clean up the do_unlink() call).
Jeremy.
|
|
files
from Windows 7. Original patch by me fixed up with the correct open files number
by jmaggard10@hotmail.com.
Jeremy.
|
|
is used). Allows make test to pass with acl_xattr.so prepended to the vfs modules.
Jeremy.
|
|
for paths
This allows smbd to pass the freshly updated RAW-SFILEINFO-END-OF-FILE
torture test.
|
|
|
|
|
|
Step 0 to restore it as a per-share paramter
|
|
Guenther
|
|
samba.
Guenther
|
|
I know Volker will look at this closely so here's the explaination :-).
Originally on a case-sensitive share we simply did a stat (or lstat)
call and returned success of fail based on the result. However this
failed to take account of incoming paths with a wildcard (which must
always fail, and with different error messages depending on whether
the wildcard is the last component or in the path). Also it failed
to take account of a stat fail with ENOENT due to a missing component
of the path as the last component (which is ok as it could be a new
file) or if the ENOENT was due to the missing component within
the path (not the last component) - which must return the correct
error. What this means is that with "case sensitive = yes" we do
one more talloc call (to get the parent directory) and one more
stat call (on the parent directory) in the case where the stat
call fails. I think this is an acceptable overhead to enable
case sensitive shares to return the correct error messages for
applications. Volker please examine carefully :-).
Jeremy.
|
|
Jeremy.
|
|
This matches what is outlined here:
http://msdn.microsoft.com/en-us/library/ms804363.aspx
This is also inline with how winXP/win7 handle this. See
RAW-SFILEINFO-END-OF-FILE* in smbtorture4.
|
|
RAW-CHKPATH when case sensitive = yes, but isn't the correct way to do
it. I'm testing a larger patch to smbd/filename.c that should fix this
correctly, and will add a torture test to ensure RAW-CHKPATH is run against
a case sensitive share once this is done.
Jeremy.
|
|
wide links = no.
Jeremy.
|
|
set. Doh !
Jeremy.
|
|
it worked in 3.3.x and 3.2.x. UCF_POSIX_PATHNAMES may no longer be needed,
as lp_posix_pathnames() being set isn't the same as case handling from
NtCreateX (lp_posix_pathnames() changes the STAT calls to LSTAT, not
the case semantics).
Jeremy.
|
|
Jeremy.
|
|
out by Volker.Restores the pathname handling for FILE_FLAG_POSIX_SEMANTICS but still prevents the O_APPEND problems. Jeremy.
|
|
Clients using a Windows open call should not be able to set
FILE_FLAG_POSIX_SEMANTICS in an open.
Jeremy.
|
|
This reverts commit e23d8a3d1f558a7e98ef2afd71e1d15c5b3a71bc.
|
|
This reverts commit f7b4151a64d8c6851e62255a7139fd00a5fc63a3.
|
|
explicit"
This reverts commit 6a149022976fe6a5579ec9afc7a4d2dcb44dc8af.
|
|
This reverts commit c85a4c9ba4a7de65a7850f6f9708df66bd24deea.
|
|
|
|
Guenther
|
|
|
|
|
|
|
|
|
|
This looks innocent, but it is visible in a netbench run. Due to boolean
short-circuiting we don't have to execute the conditions on the right-hand side
of the &&. So putting the less likely condition left gains a bit.
|
|
done on both Windows and POSIX mkdirs instead of
only on Windows mkdir (as intended). The variable
"file_attributes" had already had FILE_FLAG_POSIX_SEMANTICS
removed above in the function if it had already been set.
Jeremy.
|
|
In very hot codepaths like the statcache copy_smb_filename and the subsequent
recursive talloc_free is noticable in the CPU load.
|
|
This is a hot codepath (called from the stat cache)
|
|
the directory enumeration code (which needs it).
Jeremy.
|
|
in the "user.DOSATTRIB" EA. From the docs:
In Samba 3.5.0 and above the "user.DOSATTRIB" extended attribute has been extended to store
the create time for a file as well as the DOS attributes. This is done in a backwards compatible
way so files created by Samba 3.5.0 and above can still have the DOS attribute read from this
extended attribute by earlier versions of Samba, but they will not be able to read the create
time stored there. Storing the create time separately from the normal filesystem meta-data
allows Samba to faithfully reproduce NTFS semantics on top of a POSIX filesystem.
Passes make test but will need more testing.
Jeremy.
|
|
variable.
Jeremy.
|
|
with Bad File Descriptor.
Jeremy.
|
|
|
|
Based on a patch submitted by Tsukasa Hamano <hamano@osstech.co.jp>,
this is a change in the POSIX ACL mapping to deal with the lossy
mapping for directory ACE entries:
We have a lossy mapping: directory ACE entries
CREATOR_OWNER ------\
(map to) +---> SMB_ACL_USER_OBJ
owning sid ------/
CREATOR_GROUP ------\
(map to) +---> SMB_ACL_GROUP_OBJ
primary group sid --/
on set. And on read of a directory ACL
SMB_ACL_USER_OBJ ----> CREATOR_OWNER
SMB_ACL_GROUP_OBJ ---> CREATOR_GROUP.
Deal with this on set by duplicating
owning sid and primary group sid ACE
entries into the directory ACL.
Jeremy.
|
|
Guenther
|
|
The code to read the new V2 SAMBA_PAI entries had
two errors.
Jeremy.
|
|
BASE-DELAYWRITE and also RAW-CLOSE.
Jeremy.
|
|
the logic. This was incorrect (I'll revisit this tomorrow).
Jeremy.
|
|
set_close_write_time().
We were treating a file time set on close as a sticky write time set, and I don't
think it is. I will add a torture test later to RAW-CLOSE to confirm this.
Jeremy.
|
|
Jeremy.
|
|
"Normal" non truncate writes always cause the timestamp to
be set on close. Once a close is done on a handle this can
reset the sticky write time to current time also.
Updated smbtorture4 confirms this.
Jeremy.
|
|
|
|
We want to free the record early, not when talloc_tos() is free'ed.
|
|
Jeremy.
|
|
using older protocols (LANMAN2 or below).
Jeremy.
|
