skra [Mon, 25 Jan 2016 12:55:24 +0000 (12:55 +0000)]
Do not use blk_write_cont() and remove it. There si no need to call
blk_flush() between two writes by physical address when these are
PAGE_SIZE aligned.
skra [Mon, 25 Jan 2016 12:49:08 +0000 (12:49 +0000)]
Make minidump more like its i386 original back as with new pmap dump
interface all used physical addresses are PAGE_SIZE aligned.
Add missing copyright.
skra [Mon, 25 Jan 2016 12:43:07 +0000 (12:43 +0000)]
Create new pmap dump interface for minidump and use it for existing
pmap implementations on ARM. This way minidump code can be used without
any platform specific modification.
Also, this is the last piece missing for ARM_NEW_PMAP.
melifaro [Mon, 25 Jan 2016 06:33:15 +0000 (06:33 +0000)]
MFP r287070,r287073: split radix implementation and route table structure.
There are number of radix consumers in kernel land (pf,ipfw,nfs,route)
with different requirements. In fact, first 3 don't have _any_ requirements
and first 2 does not use radix locking. On the other hand, routing
structure do have these requirements (rnh_gen, multipath, custom
to-be-added control plane functions, different locking).
Additionally, radix should not known anything about its consumers internals.
So, radix code now uses tiny 'struct radix_head' structure along with
internal 'struct radix_mask_head' instead of 'struct radix_node_head'.
Existing consumers still uses the same 'struct radix_node_head' with
slight modifications: they need to pass pointer to (embedded)
'struct radix_head' to all radix callbacks.
Routing code now uses new 'struct rib_head' with different locking macro:
RADIX_NODE_HEAD prefix was renamed to RIB_ (which stands for routing
information base).
New net/route_var.h header was added to hold routing subsystem internal
data. 'struct rib_head' was placed there. 'struct rtentry' will also
be moved there soon.
sephe [Mon, 25 Jan 2016 05:33:18 +0000 (05:33 +0000)]
hyperv/vmbus: Avoid extra copy of page information.
The page information array could contain up to 32 elements (i.e. 512B).
And on network side w/ TSO, 11+ (176B+) elements, i.e. ~44K TSO packet,
in the page information array is quite common.
This saves us some cpu cycles.
Reviewed by: adrian, delphij
Approved by: adrian (mentor)
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D4992
sephe [Mon, 25 Jan 2016 05:18:57 +0000 (05:18 +0000)]
hyperv/hn: Remove unnecessary zeroing out the netvsc_packet
All used fields are setup one by one, so there is no need to zero
out this large struct.
While I'm here, move the stack variable near its usage.
Reviewed by: adrian, delphij, Jun Su <junsu microsoft com>
Approved by: adrian (mentor)
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D4978
sephe [Mon, 25 Jan 2016 05:12:00 +0000 (05:12 +0000)]
hyperv/hn: Use m_copydata for chimney sending.
While I'm here, move stack variables near their usage.
Reviewed by: adrian, delphij, Jun Su <junsu microsoft com>
Approved by: adrian (mentor)
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D4977
sephe [Mon, 25 Jan 2016 05:01:32 +0000 (05:01 +0000)]
hyperv/hn: Partly rework transmission path
- Avoid unnecessary malloc/free on transmission path.
- busdma(9)-fy transmission path.
- Properly handle IFF_DRV_OACTIVE. This should fix the network
stalls reported by many.
- Properly setup TSO parameters.
- Properly handle bpf(4) tapping. This 5 times the performance
during TCP sending test, when there is one bpf(4) attached.
- Allow size of chimney sending be tuned on a running system.
Default value still needs more test to determine.
Reviewed by: adrian, delphij
Approved by: adrian (mentor)
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D4972
andrew [Mon, 25 Jan 2016 00:24:57 +0000 (00:24 +0000)]
Update the Allwinner kernels:
* Use the ARM PLATFORM framework
* Use ARM_INTRNG on teh A20 as it has a GICv2
* Add a method to find which Allwinner SoC we are running on
avos [Sun, 24 Jan 2016 23:35:20 +0000 (23:35 +0000)]
net80211: reduce stack usage for ieee80211_ioctl*() methods.
Use malloc(9) for
- struct ieee80211req_wpaie2 (518 bytes, used in
ieee80211_ioctl_getwpaie())
- struct ieee80211_scan_req (128 bytes, used in setmlme_assoc_adhoc()
and ieee80211_ioctl_scanreq())
Also, drop __noinline workarounds; stack overflow is not reproducible
with recent compilers.
Tested with Clang 3.7.1, GCC 4.2.1 (from 9.3-RELEASE) and 4.9.4
(with -fstack-usage flag)
Approved by: adrian (mentor)
Differential Revision: https://reviews.freebsd.org/D5041
pfg [Sun, 24 Jan 2016 23:24:47 +0000 (23:24 +0000)]
ext2fs: passthrough any extra timestamps to the dinode struct.
In general we don't trust any of the extended timestamps unless the
EXT2F_ROCOMPAT_EXTRA_ISIZE feature is set. However, in the case where
we freshly allocated a new inode the information is valid and it is
better to pass it along instead of leaving the value undefined.
This should have no practical effect but should reduce the amount of
garbage if EXT2F_ROCOMPAT_EXTRA_ISIZE is set, like in cases where the
filesystem is converted from ext3 to ext4.
sobomax [Sun, 24 Jan 2016 22:20:13 +0000 (22:20 +0000)]
Fix bug in the readpassphrase(3) function, which can be exposed
by application closing its stdin (i.e. STDIN_FILENO) prior to
calling readpassphrase WITHOUT setting RPP_STDIN. What happens
then is that the readpassphrase would open /dev/tty, and since
file descriptors are reused, the call would return first unused
fd, which is 0 which is also STDIN_FILENO. Then due to the usage
of "input != STDIN_FILENO" in the code to do its logic, that
would result in noecho flags not set on that file descriptor,
which was original issue I've been trying to fix.
In addition to that, the readpassphrase() would leak file
descriptor on its way out, so fix that one as well.
This problem can be tested with:
$ ssh-add - < /tmp/myprivate.key
The password will not be hidden as it should and ktrace will
show:
53326 ssh-add CALL open(0x80142443c,0x100002<O_RDWR|O_CLOEXEC>,<unused>0x165f030)
53326 ssh-add NAMI "/dev/tty"
53326 ssh-add RET open 0
53326 ssh-add CALL sigprocmask(SIG_SETMASK,0x802eb1324,0x7fffffffd5e0)
53326 ssh-add RET sigprocmask 0
53326 ssh-add CALL sigaction(SIGALRM,0x7fffffffd630,0x7fffffffd610)
Instead of:
57690 ssh-add CALL open(0x80142443c,0x100002<O_RDWR|O_CLOEXEC>,<unused>0x165f030)
57690 ssh-add NAMI "/dev/tty"
57690 ssh-add RET open 4
57690 ssh-add CALL ioctl(0x4,TIOCGETA,0x7fffffffd860)
57690 ssh-add RET ioctl 0
57690 ssh-add CALL ioctl(0x4,TIOCSETAF,0x7fffffffd680)
57690 ssh-add RET ioctl 0
57690 ssh-add CALL sigprocmask(SIG_SETMASK,0x802eb1324,0x7fffffffd620)
57690 ssh-add RET sigprocmask 0
57690 ssh-add CALL sigaction(SIGALRM,0x7fffffffd670,0x7fffffffd650)
For the case when the key is read from the file.
Technically this can also be workaround'ed at the application side
by not closing the STDIN_FILENO in the first place, but readpassphrase(3)
doesn't need to make any assumptions about that. Plus the file descriptor
leak confirms that this is an oversight, rather than a deliberate behaviour.
mmel [Sun, 24 Jan 2016 11:00:38 +0000 (11:00 +0000)]
Add clock framework, a first part of new 'extended resources' family of
support frameworks(i.e. reset/regulators/phy/tsensors/fuses...).
The clock framework significantly simplifies handling of complex clock
structures found in modern SoCs. It provides the unified consumers
interface, holds and manages actual clock topology, frequency and gating.
It's tested on three different ARM boards (Nvidia Tegra TK1, Inforce 6410 and
Odroid XU2) and on one MIPS board (Creator Ci20) by kan@.
The framework is still far from perfect and probably doesn't have stable
interface yet, but we want to start testing it on more real boards and
different architectures.
pfg [Sun, 24 Jan 2016 02:41:49 +0000 (02:41 +0000)]
Rename some directory index constants.
Directory index was introduced in ext3. We don't always use the
prefix to denote the ext2 variant they belong to but when we
do we should try to be accurate.
ngie [Sat, 23 Jan 2016 22:49:13 +0000 (22:49 +0000)]
Don't run the t_cmsg_len testcase on 64-bit architectures
It always fails when trying to send through the sendit(9) private KPI in the
kernel due to a size mismatch between the msghdr and data being sent [*], which
suspiciously seems like it's related to sizeof pointers instead of scalars, or
something of that ilk
ngie [Sat, 23 Jan 2016 22:44:00 +0000 (22:44 +0000)]
- Don't return immediately in check_xucred, check_scm_creds_cmsgcred, and
check_scm_creds_sockcred after initial != NULL checks have been done for
debugging purposes
- Use more terse names for bintime (bt), cmesgcred (cmcred),
sockcred (sc), and timeval (tv) [*]
- Add some debug messages to better understand some of the flow of the test
program
MFC after: 1 week
Requested by: bde [*]
Use of the word "terse" (^.^) corrected by: jhb, rpokala [*]
Sponsored by: EMC / Isilon Storage Division
trasz [Sat, 23 Jan 2016 12:13:09 +0000 (12:13 +0000)]
Fix ru_oublocks accounting for ZFS. There are two code paths that can be
called from zfs_write() - one of them, through dmu_write(), was handled
correctly; the other wasn't.
Reviewed by: avg@
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D4923
araujo [Sat, 23 Jan 2016 04:18:44 +0000 (04:18 +0000)]
Add an IOCTL rr_limit to let users fine tuning the number of packets to be
sent using roundrobin protocol and set a better granularity and distribution
among the interfaces. Tuning the number of packages sent by interface can
increase throughput and reduce unordered packets as well as reduce SACK.
Example of usage:
# ifconfig bge0 up
# ifconfig bge1 up
# ifconfig lagg0 create
# ifconfig lagg0 laggproto roundrobin laggport bge0 laggport bge1 \
192.168.1.1 netmask 255.255.255.0
# ifconfig lagg0 rr_limit 500
fanf [Sat, 23 Jan 2016 01:20:58 +0000 (01:20 +0000)]
Fix a regression in the .de and .dk whois special cases
Ensure the special cases trigger whether we come via a referral
or via the -c option. Match host names case-insensitively.
Use the default character set supported by .de (UTF-8) since that
is more compatible with the modern world than ISO 8859-1. Persuade
them to give us a useful answer whether an internationalized
domain name is given in UTF-8 or in punycode.
fanf [Sat, 23 Jan 2016 00:28:18 +0000 (00:28 +0000)]
A lot of the cleverness in whois is no longer needed!
The IANA whois server has the right referral information for domain
names, IP addresses, and AS numbers, so whois does not need to be
able to choose servers itself (except for a few cases where referrals
do not work). We can delete a chunk of code, which is always fun.
This change improves the referral handling to be less sensitive to
all the various formats, and to allow multi-hop referral chains,
such as IANA -> registry -> registrar.
ARIN queries have the "+" flag added if no flags are present, so we
get full details if the query matches multiple objects. The Verisign
anti-spam logic is also now suppressed if the user provided a non-
trivial query string.
Uninformative rubric is now trimmed by default. The -S option
turns off trimming, and disables query fettling.
The -i option is back to its traditional pre-1999 hostname, since
whois.internic.net is more useful than whois.networksolutions.com.
Note that the old fallback/default server whois.crsnic.net is an
alias for whois.internic.net.
The manual is more informative about query syntax.
np [Fri, 22 Jan 2016 23:33:34 +0000 (23:33 +0000)]
Fix for iWARP servers that listen on INADDR_ANY.
The iWARP Connection Manager (CM) on FreeBSD creates a TCP socket to
represent an iWARP endpoint when the connection is over TCP. For
servers the current approach is to invoke create_listen callback for
each iWARP RNIC registered with the CM. This doesn't work too well for
INADDR_ANY because a listen on any TCP socket already notifies all
hardware TOEs/RNICs of the new listener. This patch fixes the server
side of things for FreeBSD. We've tried to keep all these modifications
in the iWARP/TCP specific parts of the OFED infrastructure as much as
possible.
Submitted by: Krishnamraju Eraparaju @ Chelsio (with design inputs from Steve Wise)
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D4801
kib [Fri, 22 Jan 2016 20:35:20 +0000 (20:35 +0000)]
The struct file f_advice member is overlaid with the devfs f_cdevpriv
data. If vnode bypass for devfs file failed, vn_read/vn_write are
called and might try to dereference f_advice. Limit the accesses to
f_advice to VREG vnodes only, which is the type ensured by
posix_fadvise().
The f_advice for regular files is protected by mtxpool lock. Recheck
that f_advice is not NULL after lock is taken.
Reported and tested by: bde
Sponsored by: The FreeBSD Foundation
MFC after: 3 weeks
andrew [Fri, 22 Jan 2016 16:35:01 +0000 (16:35 +0000)]
Stop including fdt_common.h in the arm64 code. We don't use anything from
it, however may have relied on header pollution to pull in the needed
headers through it
bz [Fri, 22 Jan 2016 15:03:22 +0000 (15:03 +0000)]
Change the variable to a #define in order to make gcc happy which
otherwise will complain about "variably modified 'alias' at file scope".
Unbreaks the build on gcc platforms.
jilles [Fri, 22 Jan 2016 14:52:31 +0000 (14:52 +0000)]
sem: Don't free nameinfo that is still in list when open() fails.
This bug could be reproduced easily by calling sem_open() with O_CREAT |
O_EXCL on a semaphore that is already open in the process. The struct
sem_nameinfo would be freed while still in sem_list and later calls to
sem_open() or sem_close() could access freed memory.
sephe [Fri, 22 Jan 2016 09:06:40 +0000 (09:06 +0000)]
hyperv/stor: Verify returned inquiry data before further dispatching
Windows 10 and Window 2016 will return all zero inquiry data for
non-existing slots. If we dispatched them, then a lot of useless
(0 sized) disks would be created. So we verify the returned inquiry
data (valid type, non-empty vendor/product/revision etc.), before
further dispatching.
Minor white space cleanup and wording fix.
Submitted by: Hongjiang Zhang <honzhan microsoft com>
Reviewed by: adrian, sephe, Jun Su <junsu microsoft com>
Approved by: adrian (mentor)
Modified by: sephe
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D4928
sephe [Fri, 22 Jan 2016 07:29:31 +0000 (07:29 +0000)]
hyperv/vmbus: Lookup channel through id table
Vmbus event handler will need to find the channel by its relative
id, when software interrupt for event happens. The original lookup
searches the channel list, which is not very efficient. We now
create a table indexed by the channel relative id to speed up
the channel lookup.
Submitted by: Hongjiang Zhang <honzhan microsoft com>
Reviewed by: delphij, adrain, sephe, Dexuan Cui <decui microsoft com>
Approved by: adrian (mentor)
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D4802