Commit Graph

7338 Commits

Author SHA1 Message Date
Stefan Eßer
8310a2b88c Remove duplicated parapgraph.
MFC after:	3 days
2013-07-24 08:36:29 +00:00
Hiroki Sato
d2034d1b22 Fix a bug in cp += SA_SIZE() in RTA_* loop. This could prevent
RTA_IFP from displaying correctly in route get subcommand.

Spotted by:	dim
2013-07-24 04:05:48 +00:00
Luiz Otavio O Souza
cc320e372e Add a new flag (ETHERSWITCH_VID_VALID) to say what vlangroups are in use.
This fix the case when etherswitch is printing the information of port 0
vlan group (in port based vlan mode) with no member ports.

Add the ETHERSWITCH_VID_VALID support to ip17x driver.

Add the ETHERSWITCH_VID_VALID support to rt8366 driver.

arswitch doesn't need to be updated as it doesn't support vlans management
yet.

Approved by:	adrian (mentor)
2013-07-23 13:56:38 +00:00
Luiz Otavio O Souza
da2a0dcb7a Fix the usage error message. The valid range is up to max. vlan - 1 since vlangroups starts at 0.
Approved by:	adrian (mentor)
2013-07-23 13:40:26 +00:00
Hiroki Sato
9ae3241587 - Use getnameinfo() for both of AF_INET and AF_INET6 in routename().
- Add missing "static".
2013-07-21 14:27:07 +00:00
Hiroki Sato
1addf6e47a - Fix nflag in routename().
- Display a AF_LINK address in #linkN when sdl_{nlen,alen,slen) == 0 and
  sdl_index != 0.
- Reduce unnecessary loop in pmsg_addrs().
- Remove iso_ntoa().  This is not used.
2013-07-21 11:59:41 +00:00
Hiroki Sato
247ba4776c - Simplify getaddr() and print_getmsg() by using RTAX_* instead of RTA_*
as the argument.
- Reduce unnecessary loop in print_getmsg().
2013-07-20 16:46:51 +00:00
Hiroki Sato
b3e0158e59 Show "default" for the zero-filled address consistently when nflag == 0. 2013-07-20 15:58:43 +00:00
Hiroki Sato
74ff9accb4 Add cast to (void *) to the following cases to suppress warnings by
-Wcast-align.  These do not increase the alignment requirement:

- rtm = (struct rt_msghdr *)(rtm + rtm->rtm_msglen)
- struct sockaddr *sa = &sa0; sX = (struct sockaddr_X *)sa
2013-07-20 15:23:42 +00:00
Jim Harris
38441bd9a9 Add message when nvd disks are attached and detached.
As part of this commit, add an nvme_strvis() function which borrows
heavily from cam_strvis().  This will allow stripping of
leading/trailing whitespace and also handle unprintable characters
in model/serial numbers.  This function goes into a new nvme_util.c
file which is used by both the driver and nvmecontrol.

Sponsored by:	Intel
Reviewed by:	carl
MFC after:	3 days
2013-07-19 21:40:57 +00:00
Jim Harris
2fb37e8f1a Fix nvme(4) and nvd(4) to support non 512-byte sector sizes.
Recent testing with QEMU that has variable sector size support for
NVMe uncovered some of these issues.  Chatham prototype boards supported
only 512 byte sectors.

Sponsored by:	Intel
Reviewed by:	carl
MFC after:	3 days
2013-07-19 21:33:24 +00:00
Jim Harris
dddb618e3f Use _PATH_DEV (from paths.h) for the "/dev/" string, rather than
hard-coding it.

Sponsored by:	Intel
Suggested by:	kib
Reviewed by:	kib, carl
MFC after:	3 days
2013-07-18 23:25:57 +00:00
Jim Harris
b378da2781 Simplify open_dev() by returning errno values rather than just 0 or 1.
Also remove stat() call and just rely on errno from open() call to discern
whether dev node exists or not.

Sponsored by:	Intel
Reviewed by:	kib, carl
MFC after:	3 days
2013-07-18 23:12:44 +00:00
Joel Dahl
050da67038 Minor mdoc fixes. 2013-07-18 05:46:33 +00:00
Hiroki Sato
8207f478fc Fix a gcc warning.
Pointy hat to:	hrs
2013-07-18 02:34:25 +00:00
Jim Harris
e8f25c6266 Define constants for the lengths of the serial number, model number
and firmware revision in the controller's identify structure.

Also modify consumers of these fields to ensure they only use the
specified number of bytes for their respective fields.

Sponsored by:	Intel
Reviewed by:	carl
MFC after:	3 days
2013-07-17 23:23:38 +00:00
Jim Harris
a648fac32d Always initialize fd to 0 in open_dev().
Sponsored by:	Intel
Reviewed by:	carl
MFC after:	3 days
2013-07-17 23:14:41 +00:00
Craig Rodrigues
0d6422dd87 In this GRN, Marcel Moolenaar overhauled the logic for mounting
the root file system on bootup:

   |------------------------------------------------------------------------
   |r214006 | marcel | 2010-10-17 22:01:53 -0700 (Sun, 17 Oct 2010) | 20 lines
   |
   | Re-implement the root mount logic using a recursive approach, whereby each
   |root file system (starting with devfs and a synthesized configuration) can
   |contain directives for mounting another file system as root.
   |------------------------------------------------------------------------

This commit adds a mount.conf(8) man page which documents
the root mount logic.  mount.conf(8) also provides some examples
for the /.mount.conf file, which can be used to change the root mount behavior.

Reviewed by: marcel bjk
2013-07-17 19:32:07 +00:00
Hiroki Sato
171863ac75 Use NET_RT_DUMP.0.FIB leaf node instead of setting td_proc->p_fibnum. 2013-07-17 14:15:00 +00:00
Hiroki Sato
826b6d6978 - Add support of MK_INET_SUPPORT=no.
- Fix a bug in sodump() which prevented struct sockaddr_in6 from displaying.
- Fix a bug in in fiboptlist_csv() which could cause free() of uninitialized
  pointer.
- Style cleanups:
 . Add missing "static" keywords.
 . Use an array of struct sockaddr_storage instead of sockunion for rtmsg.
 . Use err() and errx() instead of pair of fprintf(stderr, "...") + exit(1).
 . Use nitems() macro.
 . Various style(9) fixes.
2013-07-17 14:05:20 +00:00
Hiroki Sato
0365230a1a Simplify keywords.h generation. 2013-07-17 13:47:13 +00:00
Jim Harris
181c4ec270 Do not throw an error if the user requests to activate the image from
an empty firmware slot, as long as the user has specified a firmware
image to download into the empty firmware slot.

Sponsored by:	Intel
Reported by:	Joe Golio <joseph.golio@emc.com>
MFC after:	3 days
2013-07-16 15:45:37 +00:00
Robert Millan
fc98db27df Add -n flag for compatibility with Linux version of mount(8).
Reviewed by: freebsd-fs, eadler, mckusick, jh, wblock
2013-07-15 21:57:21 +00:00
Jim Harris
008ac71e0a %d should be used for printing int32_t instead of %zd.
clang does not complain about this - only gcc.

MFC after:	3 days
2013-07-12 18:13:41 +00:00
Oleg Bulyzhin
a34becb373 Fix 'SEE ALSO' list. 2013-07-12 09:20:55 +00:00
Jaakko Heinonen
36ad3f479c Clarify how "hide" and "unhide" commands work on directories. 2013-07-12 06:03:25 +00:00
Jim Harris
08c29873bf Ensure controller or namespace node name is specified before trying to
access it.

While here, also fix the identify usage message to show the -v and -x
parameters.

Sponsored by:	Intel
MFC after:	3 days
2013-07-09 21:33:12 +00:00
Jim Harris
9c0871b260 Condense the output for displaying LBA formats.
Sponsored by:	Intel
MFC after:	3 days
2013-07-09 21:31:58 +00:00
Jim Harris
2528d6a326 Send per-namespace logpage commands to the controller devnode, so they
are processed as admin commands, not I/O commands.

As part of this change, pull out the code for parsing a namespace node
string into a separate function, since it is used for both identify and
logpage commands.

Sponsored by:	Intel
MFC after:	3 days
2013-07-09 21:31:21 +00:00
Jim Harris
fdfa4d2d35 Try to read firmware image before prompting the user to confirm
firmware download.  This correctly prints an error and exits for
an incorrect firmware image name before prompting the user to
confirm the download.

Sponsored by:	Intel
MFC after:	3 days
2013-07-09 21:20:08 +00:00
Jim Harris
821ef73ca6 Incorporate feedback from bde@ based on r252672 changes:
* Use 0/1 instead of sysexits.  Man pages are confusing on this topic,
  but 0/1 is sufficient for nvmecontrol.
* Use err function family where possible instead of fprintf/exit.
* Fix some typing errors.
* Clean up some error message inconsistencies.

Sponsored by:	Intel
Submitted by:	bde (parts of firmware.c changes)
MFC after:	3 days
2013-07-09 21:14:15 +00:00
Rick Macklem
b54de2127e Document the "gssname" and "allgssname" mount options added by the
host-based initiator credential patches.
This is a content change.
2013-07-09 01:31:36 +00:00
Alan Somers
3449b15aa2 Correct the printf format specifier for total_events.
Add __printflike argument checking for devdlog().

Reported by: pjd
Approved by: gibbs (co-mentor)
2013-07-08 21:10:30 +00:00
Pawel Jakub Dawidek
374a8a322b Fix dhclient for interfaces that are down. The discover_interfaces() function
that looks for interface skips interfaces that are not UP. We need to call
dhclient-script PREINIT before we call discover_interfaces(), so the script has
a chance to bring the interface UP.

Reported by:	alfred
2013-07-04 12:27:10 +00:00
Jim Harris
960116ec24 Fix printf argument mismatch reported by gcc on i386.
Reported by: kargl
2013-07-04 00:26:24 +00:00
Pawel Jakub Dawidek
8da93e6861 MFp4 @229488:
Sandbox unprivileged process using capability mode.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:23:25 +00:00
Pawel Jakub Dawidek
a6f38228d4 MFp4 @229487:
Revoke all capability rights from STDIN and allow only for write to STDOUT and
STDERR. All those descriptors are redirected to /dev/null.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:22:29 +00:00
Pawel Jakub Dawidek
4c7a48b7a9 MFp4 @229486:
Once PID is written to the pidfile, revoke all capability rights.
We just want to keep the pidfile open.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:21:11 +00:00
Pawel Jakub Dawidek
fe5c716394 MFp4 @229485:
Only allow to overwrite lease file.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:19:43 +00:00
Pawel Jakub Dawidek
f73ac8b9de MFp4 @229484:
Limit routing socket so only poll(2) and read(2) are allowed (CAP_POLL_EVENT
and CAP_READ). This prevents unprivileged process from adding, removing or
modifying system routes.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:18:40 +00:00
Pawel Jakub Dawidek
de2c882f5c MFp4 @229483:
Limit communication pipe with privileged process to CAP_READ and CAP_WRITE.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:17:29 +00:00
Pawel Jakub Dawidek
3b2ed0659c MFp4 @229482:
- Limit bpf descriptor in unprivileged process to CAP_POLL_EVENT, CAP_READ and
  allow for SIOCGIFFLAGS, SIOCGIFMEDIA ioctls.
- While here limit bpf descriptor in privileged process to only CAP_WRITE.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:16:02 +00:00
Pawel Jakub Dawidek
235eb53002 MFp4 @229481:
Currently it was allowed to send any UDP packets from unprivileged process and
possibly any packets because /dev/bpf was open for writing.

Move sending packets to privileged process. Unprivileged process has no longer
access to not connected UDP socket and has only access to /dev/bpf in read-only
mode.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:12:54 +00:00
Pawel Jakub Dawidek
e374cef518 MFp4 @229480:
Shutdown write direction of the routing socket. We only need to read from it.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:09:02 +00:00
Pawel Jakub Dawidek
7fdc27bc51 MFp4 @229479:
- Add new request (IMSG_SEND_PACKET) that will be handled by privileged process.
- Add $FreeBSD$.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:07:55 +00:00
Pawel Jakub Dawidek
0bbe83068e MFp4 @229477:
The gethostname(3) function won't work in capability mode, because reading
kern.hostname sysctl is not permitted there.  Cache hostname early and use
cached value later.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:05:36 +00:00
Pawel Jakub Dawidek
c786bc9089 Remove redundant white-spaces. 2013-07-03 22:03:19 +00:00
Pawel Jakub Dawidek
b0f1b32afb MFp4 @229476,229478:
Make use of two fields: rfdesc and wfdesc to keep bpf descriptor open for
reading only in rfdesc and bpf descriptor open for writing only in wfdesc.
In the end they will be used by two different processes.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 22:01:52 +00:00
Pawel Jakub Dawidek
3a0fc7d8a9 MFp4 @229474:
iov_base field is 'void *' in FreeBSD, no need to cast.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 21:58:26 +00:00
Pawel Jakub Dawidek
e8da500388 MFp4 @229473:
No caller checks send_packet() return value, so make it void.

Reviewed by:	brooks
Sponsored by:	The FreeBSD Foundation
2013-07-03 21:57:24 +00:00