Ronny Chevalier [Sat, 13 Dec 2014 14:14:48 +0000 (15:14 +0100)]
systemctl: handle correctly template units for edit verb
Previously, if we provided getty@.service to systemctl edit it would
have failed when using the bus because it is an invalid unit name.
But it would have succeeded when searching in the filesystem.
Now, we check if we have a template, if we do we search in the
filesystem, if we don't have a templae and we can use the bus, we do.
Furthermore, if we provided getty@tty1.service it would not have worked
when searching the filesystem, but it would have worked with the bus.
So now, when using the filesystem we use the template name and not the
unit name, and the same when logging errors.
Michal Schmidt [Fri, 12 Dec 2014 21:51:24 +0000 (22:51 +0100)]
journal: consistently use OBJECT_<type> names instead of numbers
Note that numbers 0 and -1 are both replaced with OBJECT_UNUSED,
because they are treated the same everywhere (e.g. type_to_context()
translates them both to 0).
Michal Schmidt [Wed, 10 Dec 2014 15:37:29 +0000 (16:37 +0100)]
journal: consistently allow type==0 to mean "any type"
If type==0 and a non-NULL object were given as arguments to
journal_file_hmac_put_object(), its object type check would fail and it
would return -EBADMSG.
All existing callers use either a positive type or -1. Still, for
behavior consistency with journal_file_move_to_object() let's allow
type 0 to pass.
The only user is sd_journal_enumerate_unique() and, as explained in
the previous commit (fed67c38e3 "journal: map objects to context set by
caller, not by actual object type"), the use of them there is now
superfluous. Let's remove them.
This reverts major parts of commits: ae97089d49 journal: fix access to munmapped memory in
sd_journal_enumerate_unique 06cc69d44c sd-journal: fix sd_journal_enumerate_unique skipping values
Tested with an "--enable-debug" build and "journalctl --list-boots".
It gives the expected number of results. Additionally, if I then revert
the previous commit ("journal: map objects to context set by caller, not
to actual object type"), it crashes with SIGSEGV, as expected.
Michal Schmidt [Fri, 12 Dec 2014 17:21:55 +0000 (18:21 +0100)]
journal: map objects to context set by caller, not by actual object type
When the caller of journal_file_move_to_object() specifies type==0,
the object header is at first mapped in context 0. Then after the header
is checked, the whole object is mapped in a context determined by
the actual object type (which is not even range-checked using
type_to_context()). This looks wrong. It should map in the
caller-specified context.
An old comment in sd_journal_enumerate_unique() supports this view:
/* We do not use the type context here, but 0 instead,
* so that we can look at this data object at the same
* time as one on another file */
Clearly the expectation was that the data object will remain mapped
in context 0 without being pushed away by mapping other objects in
context OBJECT_DATA.
I suspect that this was the real bug that got fixed by ae97089d49
"journal: fix access to munmapped memory in sd_journal_enumerate_unique".
In other words, journal_file_object_keep/release are superfluous after
applying this patch.
Michal Schmidt [Fri, 12 Dec 2014 16:57:22 +0000 (17:57 +0100)]
journal: add debug mode for mmap-cache (--enable-debug=mmap-cache)
This is useful for exposing unsafe access to mmapped objects after
the context that they were mapped in was already moved.
For example:
journal_file_move_to_object(f1, OBJECT_DATA, p1, &o1);
journal_file_move_to_object(f2, OBJECT_DATA, p2, &o2);
t = o1->object.type; /* this usually works, but is unsafe */
This way entries from the same brand with the same dpi and frequency
can be coalesced. It is also visually easier to find the right DPI
than order hexadecimal identifiers.
nspawn: when booting in ephemeral mode, append random token to machine name
Also, when booting up an ephemeral container of / use the system
hostname as default machine name.
This way specifiyng -M is unnecessary when booting up an ephemeral
container, while allowing any number of ephemeral containers to run from
the same tree.
nspawn: allow spawning ephemeral nspawn containers based on the root file system of the OS
This works now:
# systemd-nspawn -xb -D / -M foobar
Which boots up an ephemeral container, based on the host's root file
system. Or in other words: you can now run the very same host OS you
booted your system with also in a container, on top of it, without
having it interfere. Great for testing whether the init system you are
hacking on still boots without reboot the system!
util: when using basename() for creating temporary files, verify the resulting name is actually valid
Also, rename filename_is_safe() to filename_is_valid(), since it
actually does a full validation for what the kernel will accept as file
name, it's not just a heuristic.
David Herrmann [Fri, 12 Dec 2014 08:43:54 +0000 (09:43 +0100)]
hwdb: fix dangling 'else' ambuguity
Imagine the following use of hwdb:
if (condition_A)
SD_HWDB_FOREACH_PROPERTY(hwdb, modalias, key, value)
operation_A(key, value);
else
log_error("...");
This should work just fine, but but definitely does not what you would
expect. Due to how SD_HWDB_FOREACH_PROPERTY is defined, the dangling
'else' is linked to the hidden 'if' statement in the macro instead of the
outer 'if (condition_A)'. This is unexpected and really annoying to debug.
Fix this by never leaving un-finished if-statements in
SD_HWDB_FOREACH_PROPERTY(). We simply inverse the if() statement and
explicitly add an 'else'-branch. This way, the statement is closed and all
ambuguities are resolved.
Tom Gundersen [Wed, 3 Dec 2014 16:02:34 +0000 (17:02 +0100)]
libsystemd: add sd-hwdb library
This is libudev-hwdb, but decoupled from libudev and in the libsystemd style.
The core code is unchanged, apart from the following minor changes:
- hwdb.bin located in /**/systemd/hwdb/ take preference over the ones located
in /**/udev/
- properties are stored internally in an OrderedHashmap, rather than a
linked list.
- a new API call allows individual properties to be queried directly, rather
than iterating over them all
- the iteration over properties have been moved inside the library, rather than
exposing a list directly
- the unused 'flags' parameter was dropped
David Herrmann [Thu, 11 Dec 2014 12:19:41 +0000 (13:19 +0100)]
bus: sync with kdbus.git
Sync up with recent kdbus changed:
* several ioctls gained .size and .items members (but still unused)
* CMD_SEND gained its own ioctl structure
* several members of kdbus_msg were dropped as they were only used during
SEND, not during RECV etc.
* CMD_RECV and CMD_SEND now share a kdbus_reply member which contains the
offset and size of the returned message.
David Herrmann [Thu, 11 Dec 2014 09:44:41 +0000 (10:44 +0100)]
bus: fix memfd-cache regarding memfd offsets
We must restore part->mmap_begin when poping memfds from the memfd-cache.
We rely on the memfds to be unsealed, so we can be sure that we own the
whole FD. Therefore, simply set part->mmap_begin to the same as
part->data.
Let's stick to generic sections that describe the general technology,
instead of specific per-object sections, unless we really have a reason
to do that otherwise.
Patrik Flykt [Wed, 10 Dec 2014 14:17:35 +0000 (16:17 +0200)]
networkd-dhcp6: Support ICMPv6 Other information
When ICMPv6 Other information is received, enable Information request
in DHCPv6. If the DHCPv6 client already exists, only update the client
if there is a transition from Other to Managed state.
Patrik Flykt [Wed, 10 Dec 2014 14:17:33 +0000 (16:17 +0200)]
test-dhcp6-client: Add a simple Information Request test case
Start the DHCPv6 test case by sending an Information Request, verifying
the reply and continuing at once with the normal address acquisition
procedure. Reuse the DHCPv6 Solicit Reply so that the client code is
verified to ignore any erroneously added IPv6 address information.
Patrik Flykt [Wed, 10 Dec 2014 14:17:32 +0000 (16:17 +0200)]
sd-dhcp6-client: Implement Information Request message
Implement Information Request message according to RFC 3315, section
18.1.5. with the excepion that the first message is not delayed by a
random amount. Instead systemd-networkd is supposed to take care of
desynchronizing between other clients.
Initialize the DHCPv6 client structure in sd_dhcp6_client_start()
as this allows toggling between information request and normal
DHCPv6 address aquisition modes.
Patrik Flykt [Wed, 10 Dec 2014 14:17:31 +0000 (16:17 +0200)]
sd-dhcp6-lease: Return only -EINVAL when a NULL lease is supplied
Suppyling a NULL lease is not a condition that needs to be asserted,
returning -EINVAL is informative enough to the caller. This simplifies
calling code and doesn't falsely indicate that something erroneous was
attempted.
Patrik Flykt [Wed, 10 Dec 2014 14:17:30 +0000 (16:17 +0200)]
sd-dhcp6-client: Make end of successfull option parsing explicit
When all DHCPv6 options have been parsed, dhcp6_option_parse() returns
-ENOMSG. Explicitely set the return value to indicate success so that
later code does not need to take this special value into account.