IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
More of these tests now pass against Windows. They still don't quite all
pass, but that's something to fix for another day.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Added delete protected test to known fail as Samba doesn't seem to enforce this yet.
Signed-off-by: Rob van der Linde <rob@catalyst.net.nz>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Remove knownfail on SMB1-DFS-SEARCH-PATHS, as we now
pass it with the new SMB1 remove DFS paths before pathname processing
changes.
Note, we still fail:
smb1.SMB1-DFS-PATHS.smbtorture\(fileserver_smb1\)
smb1.SMB1-DFS-OPERATIONS.smbtorture\(fileserver_smb1\)
even with the new SMB1 remove DFS paths before pathname
processing as those tests test *very* specific Windows behaviors. We now
pass many more of the individual internal tests, but
in order to pass them all completely I need to add
specific --with-sambaserver checks to avoid some
of the Windows DFS SMB1 insanity (error messages).
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Fri Mar 31 06:07:01 UTC 2023 on atb-devel-224
In smb2, smb1req->flags2 now never uses FLAGS2_DFS_PATHNAMES,
ucf_flags never has UCF_DFS_PATHNAME, and all calls to check_path_syntax_smb2()
pass "false" in this is_dfs parameter.
Remove all knownfails for smb2.SMB2-DFS* tests.
Now I can clean up check_path_syntax_smb2() and add
an assertion into filename_convert_dirfsp_nosymlink() that
UCF_DFS_PATHNAME is *NEVER* set in the ucf_flags for an
SMB2 connection.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Even if the client claims it's a DFS pathname. Matches what Windows does if it gets
a DFS pathname on a non-DFS share.
Remove samba3.smbtorture_s3.smb2.SMB2-NON-DFS-SHARE.smbtorture\(fileserver\)
test knownfail.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
There's nothing we can do to such a server (this
now matches the behavior for SMB1).
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15306
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Wed Mar 29 18:58:33 UTC 2023 on atb-devel-224
smbclient crashes when smbd has "smb2 max read = 0"
in the [global] section of smb.conf.
We should fail the protocol negotiation with
NT_STATUS_INVALID_NETWORK_RESPONSE in this case.
Adds knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15306
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Not a comprehensive test for all possible combinations, but it shows
the basic functionality, and it found a bug in the initial
implementation :-)
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Backlinks which are not allowed by the schema are hidden by default,
so we already set DSDB_RMD_FLAG_HIDDEN_BL on store, so we have a cheap
way to hide the backlinks.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12967
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Thu Mar 23 08:19:20 UTC 2023 on atb-devel-224
msDS-KeyCredentialLink/msDS-KeyCredentialLink-BL are defined as linked attribute pair,
but msDS-KeyCredentialLink-BL is not defined as allowed on any object class definition,
still it's possible to create msDS-KeyCredentialLink attributes.
msDS-KeyPrincipal/msDS-KeyPrincipalBL are also defined as linked attribute pair
and msDS-KeyPrincipalBL is only allowed on object class 'user', but it's possible
to create msDS-KeyPrincipal values pointing to non 'user' objects.
The result is that 'user' objects have a visible msDS-KeyPrincipalBL, but
the others don't have msDS-KeyPrincipalBL visible, by default.
The backlinks are always visible if the backlink attributes are
explicitly requested.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12967
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This allows us to pass the new tests we just added.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15338
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
These demonstrate how inherited aces are constructed and applies
per objectclass, with and without the NO_PROPAGATE_INHERIT flag.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15338
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: David Mulder <dmulder@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Mon Mar 20 20:20:41 UTC 2023 on atb-devel-224
This works as long as both tickets are issued by the same RODC.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
"samba-tool domain join" uses the replication API in a strange way, perhaps no longer
required, except that we often still have folks upgrading from very old Samba versions.
When deferring the writing out to the DB of link replication to the very end, there
is a greater opportunity for the deletion of an object to have been sent with the
other objects, and have the link applied later.
This tells the repl_meta_data code to behave as if GET_TGT had been sent at the
time the link was returned, allowing a link to a deleted object to be silently
discarded.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15329
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
"samba-tool domain join" uses the replication API in a strange way, perhaps no longer
required, except that we often still have folks upgrading from very old Samba versions.
By deferring the writing out to the DB of link replication to the very end, we have a
better chance that all the objects required are present, however the situation may
have changed during the cycle, and a link could still be sent, pointing to a deleted
object.
We currently fail in this situation.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15329
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Functions like `add_lock_to_json` and `add_profile_item_to_json` pass
some values to `json_add_int` with `intmax_t` types. This may cause
arithmetic overflow when the value grows very fast, such as the
read_bytes profiling data.
Use `json_add_int` instead of `int` to avoid the overflow.
RN: Make json output show intmax_t value properly
Signed-off-by: Li Yuxuan <liyuxuan.darfux@bytedance.com>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Thu Mar 9 21:33:43 UTC 2023 on atb-devel-224
Show that `json_add_int` can't handle value larger than int32 due to
overflow.
Add knownfail.
Signed-off-by: Li Yuxuan <liyuxuan.darfux@bytedance.com>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
When we modify a GPO, we must increment the
version number in the GPT.INI, otherwise client
machines won't process the update.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15327
Signed-off-by: David Mulder <dmulder@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
They have been made superfluous by newer declarative tests in
claims_tests.py and device_tests.py.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
These test the interaction between claims and groups in the PAC.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
These tests verify that the groups in the device info structure in the
PAC are exactly as expected under various scenarios.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Create a claim large enough to cause it to be compressed.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Windows erroneously shifts integer syntax claim values four bytes to the
right, resulting in incorrect values (if only one claim is present) or
corrupt claims data that cannot be unpacked (if other claims are
present). There's no reason to emulate such broken behaviour.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
When open_stream_pathref_fsp() returns
NT_STATUS_OBJECT_NAME_NOT_FOUND, smb_fname_rel->fsp
has been set to NULL, so we must free base_fsp separately
to prevent fd-leaks when opening a stream that doesn't
exist.
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15314
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Fri Mar 3 16:37:27 UTC 2023 on atb-devel-224
Shows we are leaking an fsp/fd if we request a non-existent stream on a file.
This then causes rename of a directory containing the file to be denied, as
it thinks we have an existing open file below it.
Add knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15314
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
The bug we're testing for needs "strict rename = no" (the default),
but the existing streams_xattr share uses "strict rename = yes" from
the [global] section.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15314
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This has the lower-case share, used in the next commit
Bug: https://bugzilla.samba.org/show_bug.cgi?id=15313
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Signed-off-by: Rob van der Linde <rob@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
The 'etype' field in a Kerberos request is ordered. Make this fact
clearer by using a tuple or an array to represent etypes rather than a
set.
get_default_enctypes() now returns encryption types in order of
strength. As a consequence, the encryption type chosen by the MIT KDC
matches up with that chosen by Windows, and more tests begin to pass.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Levels NetlogonValidationSamInfo2 and NetlogonValidationSamInfo4 behave
as might be expected, so we pass those tests. NetlogonValidationSamInfo
returns no resource groups and doesn't set the NETLOGON_EXTRA_SIDS flag,
and we fail its test.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Windows also disallows this. Note that changing a primary group to a
domain-local group is allowed by both Windows and Samba.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Heretofore we have treated the primary group SID specially, storing it
in a fixed position as the second element of the user_info_dc->sids
array, and filtering out other copies in the PAC_LOGON_INFO base
structure. This filtering has made it difficult to distinguish between
the case where the primary group is a universal or global group, located
in the base RIDs, and the case where it is a domain-local group, missing
from the base RIDs; especially since the attributes of a domain-local
primary group are lost by being stored in the PAC. Domain-local primary
groups are normally disallowed by Windows, but are allowed by Samba, and
so it is reasonable to support them with at least some measure of
consistency.
The second element of user_info_dc->sids is still reserved for the
primary group's SID, but we no longer filter out any other copies in the
array. The first two elements are no more than the SIDs of the user and
the primary group respectively; and the remaining SIDs are as if taken
without modification from arrays of SIDs in the PAC. user_info_dc->sids
should therefore become a more faithful representation of the SIDs in
the PAC. After adding resource SIDs to it with
dsdb_expand_resource_groups(), we should have a result that more closely
and in more cases matches that of Windows.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Primary groups are handled differently from normal groups of which a
user is simply a member. Of particular note is the case where a
domain-local group is made a primary group; a case normally disallowed
by Windows, but not by Samba. Therefore we want tests for it.
Our testing framework must be able to set the user's primary group, and
to clean up afterwards; to set the primary group RID of a PAC; and to
check that the primary group RID is as expected in the PAC returned to
us.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
The domain-local groups that are added to the PAC of a service ticket
are now, if the service doesn't disclaim support for SID compression,
placed into the resource groups structure in PAC_LOGON_INFO.
In a TGS exchange directed to a KDC, rather than to a service, the
resource groups structure is simply copied into the updated PAC without
any processing being done.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Besides the NETLOGON_GUEST bit indicating whether the user has been
authenticated, we now carry all of the other bits as well. This lets us
match Windows' behaviour of simply passing these bits through to an
updated PAC when processing a TGS-REQ.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This lets us test what happens in TGS-REQ exchanges to the KDC, when the
flags and resource groups are simply passed through into the new PAC,
regardless of what value the flags hold.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Crucially, in these tests the user's domain and its SID are different
from our domain and its SID. These tests will assert that in such a case
resource groups are added to the PAC and handled correctly.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Change some of the attributes in TGS-REQ PACs to ensure they are handled
correctly.
Add a test of a PAC containing resource SIDs for a service without
resource SID compression support, ensuring the SIDs are correctly
removed.
Add a Samba 4.17 compatibility test for a service with resource SID
compression support.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
As part of mitigations of CVE-2022-37966, we no longer regard the
msDS-SupportedEncryptionTypes attribute of the KDC when determining its
supported encryption types. Thus, these tests that try to disable SID
compression support by setting this attribute run to no purpose.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Resource group SIDs should only be placed into a service ticket, but we
were including them in all tickets. Now that we have access to the group
attributes, we'll filter out any groups with SE_GROUP_RESOURCE set if
we're creating a TGT.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Group expansion, performed in dsdb_expand_nested_groups(), now
incorporates a check of the type of each group. Those that are resource
groups receive the SE_GROUP_RESOURCE bit in the attributes which are now
carried alongside each group SID.
Whereas before, in auth_convert_user_info_dc_sambaseinfo() and
auth_convert_user_info_dc_saminfo6(), we invariantly used the flag
combination SE_GROUP_MANDATORY | SE_GROUP_ENABLED_BY_DEFAULT |
SE_GROUP_ENABLED to set attributes in the PAC, we now take the correct
attributes from user_info_dc.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=10635
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Tue Jan 31 13:43:54 UTC 2023 on atb-devel-224
This allows our new tests to pass as these need to be checked first.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=10635
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
We want to totally ignore the string DN if there is a GUID,
as clients like "Microsoft Azure AD connect cloud sync" will
set a literal "DummyDN" string.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=10635
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This make this funciton the gatekeeper between the wire format and the
internal struct ldb_dn, checking if the DN exists and which NC
it belongs to along the way, and presenting only a DB-returned
DN for internal processing.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=10635
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This allows lookup of a DN with a GUID only or GUID and string,
possibly not yet in the database, yet still getting the correct result.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=10635
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15287
Signed-off-by: Pavel Filipenský <pfilipensky@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Thu Jan 26 15:07:57 UTC 2023 on atb-devel-224
vfs_virusfilter expects a non-NULL fsp->fsp_name to use for printing debugs
(it always indirects fsp->fsp_name). vfs_fruit also does the same, so would
also crash in fruit_close() with 'debug level = 10' and vfs_default:VFS_OPEN_HOW_RESOLVE_NO_SYMLINKS = no
set (we don't test with that which is why we haven't noticed
this before).
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15283
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Fri Jan 13 08:33:47 UTC 2023 on sn-devel-184
Modify check_infected_read() test to use a 2-level deep
directory.
We must have vfs_default:VFS_OPEN_HOW_RESOLVE_NO_SYMLINKS = no
set on the virusscanner share as otherwise the openat flag
shortcut defeats the test.
Add knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15283
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
After s3 and s4 rpc servers merge we can avoid duplicated code.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240
Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Instead of using the generic deprecated option use the specific
server require schannel:COMPUTERACCOUNT = no in order to allow
legacy tests for pass.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240
Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
MacOS clients send SMB2 DFS pathnames as \server\share\file\name.
Ensure smbd can cope with this by stipping any leading '\\'
characters from an SMB2 packet with the DFS flag set.
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15277
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Wed Jan 4 07:46:06 UTC 2023 on sn-devel-184
Shows that we fail to cope with MacOSX clients that send a
(or more than one) leading '\\' character for an SMB2 DFS pathname.
I missed this in earlier tests as Windows, Linux, and
libsmbclient clients do NOT send a leading backslash
for SMB2 DFS paths. Only MacOSX (sigh:-).
Passes against Windows. Adds a knownfail for smbd.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15277
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
This fixes a regression in commit f03665bb7e
The use of reload_services() has a lot of side effects, e.g. reopen of
log files and other things, which are only useful in smbd, but not in rpcd_classic.
It was also unloading the user and registry shares we loaded a few lines
above.
We need to do all (re-)loading as root, otherwise we won't be able
to read root only smb.conf files, access registry shares, ...
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15243
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15266
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Walker <awalker@ixsystems.com>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Thu Dec 29 21:14:02 UTC 2022 on sn-devel-184
Currently ad_member_s3_join is only used for testing samba-tool join and that'll
work just fine being joined to vampire_dc instead of ad_dc.
vampire_dc is an additional DC in the SAMBADOMAIN "started" by ad_dc_ntvfs, so
by joining ad_member_s3_join to the SAMBADOMAIN, it is member of a domain with
more then one DC.
Subsequently I'll add a test that needs such an environment.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Currently applied files which are manually
removed do not get re-applied.
Signed-off-by: David Mulder <dmulder@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Don't do the get_real_filename() retry if we're in posix context of if
the connection is case sensitive.
The whole concept of case sensivity blows my brain. In SMB1 without
posix extensions it's a per-request thing. In SMB2 without posix
extensions this should just depend on "case sensitive = yes/no", and
in future SMB2 posix extensions this will become a per-request thing
again, depending on the existence of the posix create context.
Then there are other semantics that are attached to posix-ness, which
have nothing to do with case sensivity. See for example merge request
2819 and bug 8776, or commit f0e1137425. Also see
check_path_syntax_internal().
This patch uses the same flags as openat_pathref_fsp_case_insensitive()
does, but I am 100% certain this is wrong in a subtle way.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Thu Dec 15 11:30:04 UTC 2022 on sn-devel-184
We need to take the value from the msDS-SupportedEncryptionTypes
attribute and only take the default if there's no value or
if the value is 0.
For krbtgt and DC accounts we need to force support for
ARCFOUR-HMAC-MD5 and AES encryption types and add the related bits
in addtition. (Note for krbtgt msDS-SupportedEncryptionTypes is
completely ignored the hardcoded value is the default, so there's
no AES256-SK for krbtgt).
For UF_USE_DES_KEY_ONLY on the account we reset
the value to 0, these accounts are in fact disabled completely,
as they always result in KRB5KDC_ERR_ETYPE_NOSUPP.
Then we try to get all encryption keys marked in
supported_enctypes, and the available_enctypes
is a reduced set depending on what keys are
actually stored in the database.
We select the supported session key enctypes by the available
keys and in addition based on AES256-SK as well as the
"kdc force enable rc4 weak session keys" option.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13135
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15237
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This tests work out the difference between
- msDS-SupportedEncryptionTypes value or it's default
- software defined extra flags for DC accounts
- accounts with only an nt hash being stored
- the resulting value in the KRB5_PADATA_SUPPORTED_ETYPES announcement
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13135
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15237
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
ENC_HMAC_SHA1_96_AES256_SK is a flag introduced for by Microsoft in this
CVE to indicate that additionally, AES session keys are available. We
set the etypes available for session keys depending on the encryption
types that are supported by the principal.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15219
Pair-Programmed-With: Joseph Sutton <josephsutton@catalyst.net.nz>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This shows that changes around RC4 encryption types do not break older
functional levels where only RC4 keys are available.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15237
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
ENC_HMAC_SHA1_96_AES256_SK is a flag introduced for by Microsoft in this CVE
to indicate that additionally, AES session keys are available.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15237
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Extend the RC4 Protected Users tests to use different preauth etypes.
This helps test the nuances of the new expected behaviour and allows the
tests to continue passing.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15237
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15231
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
We need to select server, not client, to compare client etypes against.
(It is not useful to compare the client-supplied encryption types with
the client's own long-term keys.)
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15237
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
The KDC should leave the choice of ticket encryption type up to the
target service, and admit no influence from the client.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15237
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
By default we'll now require schannel connections with
privacy/sealing/encryption.
But we allow exceptions for specific computer/trust accounts.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15260
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Instead of using the generic deprecated option use the specific
allow nt4 crypto:COMPUTERACCOUNT = yes and
server reject md5 schannel:COMPUTERACCOUNT = no
in order to allow legacy tests for pass.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
AES is supported by Windows Server >= 2008R2, Windows (Client) >= 7 and Samba >= 4.0,
so there's no reason to allow md5 clients by default.
However some third party domain members may need it.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Instead of using the generic deprecated option use the specific
server require schannel:COMPUTERACCOUNT = no in order to allow
legacy tests for pass.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
By compiling the knownfail lists ahead of time we change a 20min test
into a 90sec test.
This could be improved further by combining this into a single regular expression,
but this is enough for now. The 'reason' is thankfully not used.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15258
Pair-programmed-with: Joseph Sutton <josephsutton@catalyst.net.nz>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Mainly so I can go
make bin/test_lzxpress_plain && bin/test_lzxpress_plain
valgrind bin/test_lzxpress_plain
rr bin/test_lzxpress_plain
rr replay
in a tight loop.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
st/summary is useless. If you'll find anything, it'll be in st/subunit.
However, in case *something* useful ever ends up there we still mention it.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
This format is described in [MS-XCA] 2.1 and 2.2, with exegesis in
many posts on the cifs-protocol list[1].
The two public functions are:
ssize_t lzxpress_huffman_decompress(const uint8_t *input,
size_t input_size,
uint8_t *output,
size_t output_size);
uint8_t *lzxpress_huffman_decompress_talloc(TALLOC_CTX *mem_ctx,
const uint8_t *input_bytes,
size_t input_size,
size_t output_size);
In both cases the caller needs to know the *exact* decompressed size,
which is essential for decompression. The _talloc version allocates
the buffer for you, and uses the talloc context to allocate a 128k
working buffer. THe non-talloc function will allocate the working
buffer on the stack.
This compression format gives better compression for messages of
several kilobytes than the "plain" LXZPRESS compression, but is
probably a bit slower to decompress and is certainly worse for very
short messages, having a fixed 256 byte overhead for the first Huffman
table.
Experiments show decompression rates between 20 and 500 MB per second,
depending on the compression ratio and data size, on an i5-1135G7 with
no compiler optimisations.
This compression format is used in AD claims and in SMB, but that
doesn't happen with this commit.
I will not try to describe LZ77 or Huffman encoding here. Don't expect
an answer in MS-XCA either; instead read the code and/or Wikipedia.
[1] Much of that starts here:
https://lists.samba.org/archive/cifs-protocol/2022-October/
but there's more earlier, particularly in June/July 2020, when
Aurélien Aptel was working on an implementation that ended up in
Wireshark.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Pair-programmed-with: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Sometimes (e.g. in lzxpress Huffman encoding, and in some of our
tests: c.f. https://lists.samba.org/archive/samba-technical/2018-March/126010.html)
we want a stable sort algorithm (meaning one that retains the previous
order of items that compare equal).
The GNU libc qsort() is *usually* stable, in that it first tries to
use a mergesort but reverts to quicksort if the necessary allocations
fail. That has led Samba developers to unthinkingly assume qsort() is
stable which is not the case on many platforms, and might not always
be on GNU/Linuxes either.
This adds four functions. stable_sort() sorts an array, and requires
an auxiliary working array of the same size. stable_sort_talloc()
takes a talloc context so it ca create a working array and call
stable_sort(). stable_sort_r() takes an opaque context blob that gets
passed to the compare function, like qsort_r() and ldb_qsort(). And
stable_sort_talloc_r() rounds out the quadrant.
These are LGPL so that the can be used in ldb, which has problems with
unstable sort.
The tests are borrowed and extended from test_ldb_qsort.c.
When sorting non-trivial structs this is roughly as fast as GNU qsort,
but GNU qsort has optimisations for small items, using direct
assignments of rather than memcpy where the size allows the item to be
cast as some kind of int.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Remove knownfail.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Thu Dec 1 16:04:07 UTC 2022 on sn-devel-184
Cf MS-FSA 2.1.5.14.2
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15252
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Mon Nov 28 10:14:12 UTC 2022 on sn-devel-184
For now we allow build warnings and only do some basic testing.
We also ignore timestamp related problems, as well as some charset
failures.
Over time we should try to address the situation by not allowing warnings
and verify if expected failures are harmless or not.
But it's already much better then having no 32bit testing at all!
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Thu Nov 24 12:05:26 UTC 2022 on sn-devel-184
We need to pass --mitkrb5 to selftest.pl in all cases we use
system mit kerberos not only when we also test the kdc.
We can't use a replay cache in selftest verifies the stat.st_uid
against getuid().
BTW: while debugging this on ubuntu 22.04 I exported
KRB5_TRACE="/dev/stderr", which means we get tracing into
the servers log file and into selftest_prefix/subunit for the client...
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This still all fails, but if you run them against Windows they work.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Nov 22 19:25:34 UTC 2022 on sn-devel-184
The next commits will create symlinks via posix extensions to test the
smb2 symlink error return. Creating posix symlinks is not allowed with
follow symlinks = no, but it's currently our only way to create
symlinks over SMB. This could go away once we can create symlinks via
reparse points.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This still all fails, but if you run them against Windows they work.
How to run:
PYTHONPATH=bin/python \
LOCAL_PATH=/tmp \
SMB1_SHARE=share \
SMB2_SHARE=share \
SHARENAME=share \
SERVER_IP=<server-ip> \
DOMAIN=<your-domain> \
USERNAME=Administrator \
PASSWORD=<your-password> \
SMB_CONF_PATH=/usr/local/samba/etc/smb.conf \
SERVERCONFFILE="$SMB_CONF_PATH" \
python3 -m samba.subunit.run samba.tests.reparsepoints
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Ensure `samba-tool gpo manage sudoers remove` is
backward compatible with the GPME sudo rules.
Signed-off-by: David Mulder <dmulder@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The file format for storing the sudo rules
changed in samba-tool, but these can still be
added via the GPME. We should still include them
here.
Signed-off-by: David Mulder <dmulder@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Ensure `samba-tool gpo manage sudoers list` is
backward compatible with the GPME sudo rules.
Signed-off-by: David Mulder <dmulder@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The file format for storing the sudo rules
changed in samba-tool, but these can still be
added via the GPME. We should still include them
here.
Signed-off-by: David Mulder <dmulder@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15243
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri Nov 18 19:17:31 UTC 2022 on sn-devel-184
Async read and write go synchronous in the same case,
so do the same here.
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15172
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Thu Nov 17 05:55:42 UTC 2022 on sn-devel-184
Shows we fail sending an SMB2_OP_FLUSH + SMB2_OP_FLUSH
compound if we immediately close the file afterward.
Internally the flushes go async and we free the req, then
we process the close. When the flushes complete they try to access
already freed data.
Extra test which will allow me to test when the final
component (flush) of the compound goes async and returns
NT_STATUS_PENDING.
Add knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15172
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Shows we fail sending an SMB2_OP_FLUSH + SMB2_OP_CLOSE
compound. Internally the flush goes async and
we free the req, then we process the close.
When the flush completes it tries to access
already freed data.
Found using the Apple MacOSX client at SNIA SDC 2022.
Add knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15172
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Currently all fsctls we implement need the base fsp, not
an alternate data stream fsp. We may revisit this later
if we implement fsctls that operate on an ADS.
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15236
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andrew Walker <awalker@ixsystems.com>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Mon Nov 14 18:13:31 UTC 2022 on sn-devel-184
Thanks to Metze for the hint that all file servers already listen on 2
addressess -- V4 and V6 :-)
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Thu Nov 10 08:23:14 UTC 2022 on sn-devel-184
In which we make AS and TGS requests and verify the SIDs we expect are
returned in the PAC.
Example command to test against Windows Server 2019 functional level
2016 with FAST enabled:
ADMIN_USERNAME=Administrator ADMIN_PASSWORD=locDCpass1 \
CLAIMS_SUPPORT=1 COMPOUND_ID_SUPPORT=1 DC_SERVER=ADDC.EXAMPLE.COM \
DOMAIN=EXAMPLE EXPECT_PAC=1 FAST_SUPPORT=1 KRB5_CONFIG=krb5.conf \
PYTHONPATH=bin/python REALM=EXAMPLE.COM SERVER=ADDC.EXAMPLE.COM \
SKIP_INVALID=1 SMB_CONF_PATH=smb.conf STRICT_CHECKING=1 \
TKT_SIG_SUPPORT=1 python3 python/samba/tests/krb5/group_tests.py
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Tue Nov 8 03:37:37 UTC 2022 on sn-devel-184
The intention of this option was to hide *files*. Before this patch we
also hide directories where new files are dropped.
This is a change in behaviour, but I think this option is niche enough
to justify not adding another parameter that we then need to test. If
workflows break with this change and people depend on directories also
to be hidden, we can still add the additional option value required.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Mon Nov 7 22:58:33 UTC 2022 on sn-devel-184
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14808
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: David Mulder <dmulder@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Nov 1 18:31:22 UTC 2022 on sn-devel-184
Add torture test to show smbc_getxattr() should return -1 on
failure, 0 on success.
Add knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14808
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: David Mulder <dmulder@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15212
Signed-off-by: David Mulder <dmulder@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): David Mulder <dmulder@samba.org>
Autobuild-Date(master): Tue Oct 25 15:21:08 UTC 2022 on sn-devel-184
Startup scripts were failing to execute when no
parameters were provided to the script.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15212
Signed-off-by: David Mulder <dmulder@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We later subtract 8 when calculating the length of the output message
buffer. If padlength is excessively high, this calculation can underflow
and result in a very large positive value.
Now we properly constrain the value of padlength so underflow shouldn't
be possible.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15134
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
If len_len is equal to total_len - 1 (i.e. the input consists only of a
0x60 byte and a length), the expression 'total_len - 1 - len_len - 1',
used as the 'len' parameter to der_get_length(), will overflow to
SIZE_MAX. Then der_get_length() will proceed to read, unconstrained,
whatever data follows in memory. Add a check to ensure that doesn't
happen.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15134
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We should make sure that the result of 'total_len - mech_len' won't
overflow, and that we don't memcmp() past the end of the buffer.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15134
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
There were some reports that strace output an LDAP server socket is in
CLOSE_WAIT state, returning EAGAIN for writev over and over (after a call to
epoll() each time).
In the tstream_bsd code the problem happens when we have a pending
writev_send, while there's no readv_send pending. In that case
we still ask for TEVENT_FD_READ in order to notice connection errors
early, so we try to call writev even if the socket doesn't report TEVENT_FD_WRITE.
And there are situations where we do that over and over again.
It happens like this with a Linux kernel:
tcp_fin() has this:
struct tcp_sock *tp = tcp_sk(sk);
inet_csk_schedule_ack(sk);
sk->sk_shutdown |= RCV_SHUTDOWN;
sock_set_flag(sk, SOCK_DONE);
switch (sk->sk_state) {
case TCP_SYN_RECV:
case TCP_ESTABLISHED:
/* Move to CLOSE_WAIT */
tcp_set_state(sk, TCP_CLOSE_WAIT);
inet_csk_enter_pingpong_mode(sk);
break;
It means RCV_SHUTDOWN gets set as well as TCP_CLOSE_WAIT, but
sk->sk_err is not changed to indicate an error.
tcp_sendmsg_locked has this:
...
err = -EPIPE;
if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
goto do_error;
while (msg_data_left(msg)) {
int copy = 0;
skb = tcp_write_queue_tail(sk);
if (skb)
copy = size_goal - skb->len;
if (copy <= 0 || !tcp_skb_can_collapse_to(skb)) {
bool first_skb;
new_segment:
if (!sk_stream_memory_free(sk))
goto wait_for_space;
...
wait_for_space:
set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
if (copied)
tcp_push(sk, flags & ~MSG_MORE, mss_now,
TCP_NAGLE_PUSH, size_goal);
err = sk_stream_wait_memory(sk, &timeo);
if (err != 0)
goto do_error;
It means if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN)) doesn't
hit as we only have RCV_SHUTDOWN and sk_stream_wait_memory returns
-EAGAIN.
tcp_poll has this:
if (sk->sk_shutdown & RCV_SHUTDOWN)
mask |= EPOLLIN | EPOLLRDNORM | EPOLLRDHUP;
So we'll get EPOLLIN | EPOLLRDNORM | EPOLLRDHUP triggering
TEVENT_FD_READ and writev/sendmsg keeps getting EAGAIN.
So we need to always clear TEVENT_FD_READ if we don't
have readable handler in order to avoid burning cpu.
But we turn it on again after a timeout of 1 second
in order to monitor the error state of the connection.
And now that our tsocket_bsd_error() helper checks for POLLRDHUP,
we can check if the socket is in an error state before calling the
writable handler when TEVENT_FD_READ was reported.
Only on error we'll call the writable handler, which will pick
the error without calling writev().
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15202
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15202
Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15195
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Pavel Filipenský <pfilipensky@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Oct 19 00:13:56 UTC 2022 on sn-devel-184
For type == ADOUBLE_META, fio->fake_fd is true so
writes are already synchronous, just call tevent_req_post().
For type == ADOUBLE_RSRC we know we are configured
with FRUIT_RSRC_ADFILE (because fruit_must_handle_aio_stream()
returned true), so we can just call SMB_VFS_NEXT_FSYNC_SEND()
after replacing fsp with fio->ad_fsp.
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15182
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Böhme <slow@samba.org>
This shows we currently hang when sending an SMB2_OP_FLUSH on
an AFP_Resource fork.
Adds knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15182
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Böhme <slow@samba.org>
