mikeb [Mon, 12 Jan 2015 10:40:51 +0000 (10:40 +0000)]
When setting up advanced TX descriptor use m_getptr to locate the IP
or IPv6 header instead of assuming contiguousness of the target buffer
across Ethernet and IP/IPv6 headers.
Tested by Kapetanakis Giannis <bilias at edu ! physics ! uoc ! gr>,
thanks! Problem analysis and initial diff by dlg@.
nicm [Mon, 12 Jan 2015 10:11:07 +0000 (10:11 +0000)]
Add a note that HOST_NAME_MAX does not include a terminating \0, unlike
MAXHOSTNAMELEN.
ok deraadt guenther
deraadt [Mon, 12 Jan 2015 07:12:58 +0000 (07:12 +0000)]
pretty up some malloc() calls
deraadt [Mon, 12 Jan 2015 07:11:41 +0000 (07:11 +0000)]
hibernate_suspend() should not pmap_kremove by itself; hibernate_free()
must do that. otherwise, pmap_kremove is called twice. i386 in particular
does not tolerate that, found by sebastia
ok mlarkin kettenis
brad [Mon, 12 Jan 2015 04:49:41 +0000 (04:49 +0000)]
Some fixes and tidying up of the receive filter handling bits.
tedu [Mon, 12 Jan 2015 03:20:04 +0000 (03:20 +0000)]
rename blocks to words. bcrypt "blocks" are unrelated to blowfish blocks,
nor are they the same size.
doug [Mon, 12 Jan 2015 02:15:23 +0000 (02:15 +0000)]
Fix a memory leak in bss_dgram.
Free data->saved_message.data. Based on OpenSSL commit:
41cd41c4416f545a18ead37e09e437c75fa07c95 except this version sets a->ptr
to NULL to avoid accidental reuse and handles malloc failing.
ok beck@, input + ok miod@
dlg [Mon, 12 Jan 2015 00:07:55 +0000 (00:07 +0000)]
stash the softc in the memory allocated for setting up a task for
creating sensors.
espie [Sun, 11 Jan 2015 22:34:08 +0000 (22:34 +0000)]
replace fw_update (shell script) with a specialized version of pkg_add.
This has numerous advantages, in terms of speed and control.
Initial version should be (mostly) identical to the old one.
thanks to rpe@ and sthen@ for tests.
miod [Sun, 11 Jan 2015 21:24:32 +0000 (21:24 +0000)]
Remove orphaned defines.
miod [Sun, 11 Jan 2015 21:02:55 +0000 (21:02 +0000)]
No more etc set, also sparc nowadays needs more than 16MB to boot multiuser.
kettenis [Sun, 11 Jan 2015 19:59:56 +0000 (19:59 +0000)]
Tolerate checksum errors in ACPI tables (but complain about them). Apparently
there are machines out there with broken BIOSen with such checksum errors in
thier RSDT tables. Windows runs fine on these machines, so it probably does
not check. Linux ignores checksum errors as well (and prints a warning).
ok deraadt@, krw@
guenther [Sun, 11 Jan 2015 19:34:52 +0000 (19:34 +0000)]
LOCKDEBUG is dead; perform the funeral rites
pointed out by Helg (xx404 (at) msn.com)
ok deraadt@ miod@
tedu [Sun, 11 Jan 2015 19:25:13 +0000 (19:25 +0000)]
switch prototype warnings to implicit-declaration warnings.
This should catch all the same bad cases, but be a little less aggravating
in circumstances where a prototype isn't necessary
ok deraadt
gilles [Sun, 11 Jan 2015 18:25:54 +0000 (18:25 +0000)]
revert two latest commits until a crash is fixed
miod [Sun, 11 Jan 2015 18:10:33 +0000 (18:10 +0000)]
When using sun4 PROM I/O routines, make sure we never attempt to transfer more
than the maximum DMA transfer size the PROM reports.
Crank version to 2.9.
mpi [Sun, 11 Jan 2015 15:41:16 +0000 (15:41 +0000)]
There's no simple way so check for DMA buffer overrun with isochronous
transfers, so skip the check for such xfers for the moment.
Issue reported by Ingo Feinerer and fix confirmed by Jan Stary.
miod [Sun, 11 Jan 2015 15:35:38 +0000 (15:35 +0000)]
Use the softc-specific function pointers for bcopy() and bzero() in iestart(),
instead of, well, bcopy and bzero. This lets ie@vme transmit packets instead of
panicing due to memset() trying to use 32-bit stores on a device address.
miod [Sun, 11 Jan 2015 15:29:03 +0000 (15:29 +0000)]
Make sure to initialize nbpg to a sensible value in bootxx, when running on
a sun4 system - the prom i/o routines depend on it.
miod [Sun, 11 Jan 2015 14:56:23 +0000 (14:56 +0000)]
Unbreak sun4
deraadt [Sun, 11 Jan 2015 13:00:05 +0000 (13:00 +0000)]
global "list" is not the nicest variable name
seen my mlarkin, ok miod
kirby [Sun, 11 Jan 2015 12:17:44 +0000 (12:17 +0000)]
add/fix/extend entries of exUSSR airports I've been to.
ok sthen@
gilles [Sun, 11 Jan 2015 11:49:36 +0000 (11:49 +0000)]
unbreak the delivery_filename backend, now that it no longer prepends the
From separator line but gets it injected in its iobuf by the mda process,
we need to make sure not to escape the first 'From ' we receive.
spotted by deraadt@
dlg [Sun, 11 Jan 2015 11:18:36 +0000 (11:18 +0000)]
armv6 introduced opcodes for reversing words in registers. we can
use these on armv7 as a backend for byteswapping things that endian.h
provide.
i dunno if its faster, but it makes smaller code. saves 30k on
GENERIC-OMAP.
ok jsing@ bmercer@ jsg@
deraadt [Sun, 11 Jan 2015 04:14:40 +0000 (04:14 +0000)]
correctly use HOST_NAME_MAX.
Some notes:
POSIX HOST_NAME_MAX doesn't include the NUL.
POSIX LOGIN_NAME_MAX and TTY_NAME_MAX do include the NUL.
BSD MAXHOSTNAMELEN includes the NUL. Actually, most of the historical
BSD MAX* defines did include the NUL, except for the historical
mistake of utmp fields without NULs in the string, which directly led
to strncpy.. just showing how error prone this kind of accounting is.
CSRG did right. Somehow POSIX missed the memo on the concepts of
carefulness and consistancy, and we are still paying the price when
people trip over this. Of course, glibc is even more amazing (that is
a hint to blackhats)
ok guenther
deraadt [Sun, 11 Jan 2015 03:08:38 +0000 (03:08 +0000)]
use mallocarray(), then set the size afterwards. Prefer if people learn
and use this idiom everywhere, it is always safest.
deraadt [Sun, 11 Jan 2015 03:07:44 +0000 (03:07 +0000)]
use mallocarray()
deraadt [Sun, 11 Jan 2015 03:06:19 +0000 (03:06 +0000)]
mallocarray() for the rx_buffer memory
guenther [Sun, 11 Jan 2015 03:05:03 +0000 (03:05 +0000)]
Casting a gid_t to gid_t for a gid_t argument is overkill
deraadt [Sun, 11 Jan 2015 03:04:53 +0000 (03:04 +0000)]
use mallocarray() when reallocating the mux table
deraadt [Sun, 11 Jan 2015 02:59:59 +0000 (02:59 +0000)]
KUE_MCFILTCNT() is non-constant, so use mallocarray()
deraadt [Sun, 11 Jan 2015 02:57:09 +0000 (02:57 +0000)]
unnecessary cast
deraadt [Sun, 11 Jan 2015 02:35:44 +0000 (02:35 +0000)]
snc
kettenis [Sat, 10 Jan 2015 22:14:30 +0000 (22:14 +0000)]
Get the MAC address from the PCI ROM on systems without Open Firmware. Code
largely stolen from cas(4), but with a few tweaks.
kettenis [Sat, 10 Jan 2015 19:52:37 +0000 (19:52 +0000)]
Bye, bye simple_lock.
kettenis [Sat, 10 Jan 2015 18:56:55 +0000 (18:56 +0000)]
Disable (most) interrupts when the interface goes down.
kettenis [Sat, 10 Jan 2015 17:19:02 +0000 (17:19 +0000)]
Enable nep(4).
kettenis [Sat, 10 Jan 2015 17:17:48 +0000 (17:17 +0000)]
Hook up nep(4).
kettenis [Sat, 10 Jan 2015 17:02:17 +0000 (17:02 +0000)]
A few more loose ends.
kettenis [Sat, 10 Jan 2015 16:41:04 +0000 (16:41 +0000)]
Remove debug printfs.
kettenis [Sat, 10 Jan 2015 16:39:12 +0000 (16:39 +0000)]
nep(4)
kettenis [Sat, 10 Jan 2015 16:26:17 +0000 (16:26 +0000)]
Hook up nep(4).
kettenis [Sat, 10 Jan 2015 16:25:33 +0000 (16:25 +0000)]
Add support for promiscuous mode and multicase.
sobrado [Sat, 10 Jan 2015 15:03:50 +0000 (15:03 +0000)]
tell the truth about DES.
joint work with djm@ and jsing@, who suggested stronger words warning people
away from single-DES.
ok djm@
kettenis [Sat, 10 Jan 2015 14:55:29 +0000 (14:55 +0000)]
Properly stop DMA and tear down the rings when taking the interface down.
tedu [Sat, 10 Jan 2015 13:48:02 +0000 (13:48 +0000)]
1. They're flags, not counters. Set to one instead of incrementing.
2. The G flag is useless and never checked. Remove it.
ok millert
tedu [Sat, 10 Jan 2015 13:47:05 +0000 (13:47 +0000)]
don't check for a return value that host() doesn't return, so future
generations don't try to change any of the values and break the code.
ok deraadt
mpi [Sat, 10 Jan 2015 11:43:37 +0000 (11:43 +0000)]
Correct some comments and merge in6_if_up() into in6_ifattach() to
reflect that IPv6 link-local addresses are no longer automagically
configured the first time an interface is brought up.
ok henning@, stsp@
stsp [Sat, 10 Jan 2015 10:45:46 +0000 (10:45 +0000)]
Remove pointless empty 64bit support code stubs from bwi(4).
It seems no 64bit bus space support is forthcoming. Nothing
has happened in DragonflyBSD for several years and FreeBSD
has a different driver for 64bit chips called bwn(4).
ok mpi@ some time ago
stsp [Sat, 10 Jan 2015 10:38:12 +0000 (10:38 +0000)]
Merge two bwi(4) fixes from FreeBSD:
------------------------------------------------------------------------
r192306 | imp | 2009-05-18 17:31:26 +0200 (Mon, 18 May 2009) | 10 lines
Fix a typo from the original driver. We need to write ctrl2 into RF
register 0x52, not ctrl1. This appears to be a mistake in the bcm
reverse engineering page, and has been corrected there. Tracing
through the code, this is more in keeping with the "documented"
register. Sephe thinks it looks interesting and may be worth
fixing. :)
Submitted by: ddkprog at yahoo com
Reviewed by: Sepherosa Ziehau
------------------------------------------------------------------------
r192042 | nwhitehorn | 2009-05-13 16:25:55 +0200 (Wed, 13 May 2009) | 4 lines
Add a short delay after programming PHY registers to give some time for
the engine to catch up. This prevents a machine check exception from
illegal memory requests with a BCM4318.
------------------------------------------------------------------------
No regressions observed by myself, miod@, and Vesa Norrman (vesbula at gmail).
guenther [Sat, 10 Jan 2015 07:56:16 +0000 (07:56 +0000)]
Correct reversed test of setmountent() return
Fix from Dennis Lindroos (denafcm (at) gmail.com)
bcook [Sat, 10 Jan 2015 01:56:52 +0000 (01:56 +0000)]
revert host() back to correct behavior.
unbreak config file address parsing
bcook [Fri, 9 Jan 2015 23:48:21 +0000 (23:48 +0000)]
return -1 on host() address parsing failure, not 1.
Match what parse.y expects it to return.
ok millert@
bcook [Fri, 9 Jan 2015 23:44:07 +0000 (23:44 +0000)]
return -1 on host() address parsing failure, not 1.
Match what parse.y expects it to return.
ok millert@
tedu [Fri, 9 Jan 2015 21:49:59 +0000 (21:49 +0000)]
absurd optimization. reduce initial space by one to avoid leaving one byte
hanging off on a page by itself.
kettenis [Fri, 9 Jan 2015 20:45:40 +0000 (20:45 +0000)]
The "SunMicro Virtual Eth Device" found on Oracle's more recent SPARC systems
claims to support both RNDIS and CDC Ethernet. However, RNDIS doesn't seem
to work, at least not with our driver. So blacklist it here such that cdce(4)
takes over.
ok miod@, armani@ (both a while back)
kettenis [Fri, 9 Jan 2015 20:42:40 +0000 (20:42 +0000)]
regen
kettenis [Fri, 9 Jan 2015 20:41:48 +0000 (20:41 +0000)]
SunMicro Virtual Eth Device
sthen [Fri, 9 Jan 2015 20:34:21 +0000 (20:34 +0000)]
Change rssi passed to ifconfig(8) to a signed value to fix printing signal
strengths on 802.11 interfaces. ok stsp@
Power on wireless interfaces is usually scaled in dBm. rssi (received signal
strength) is expected to be a -ve dBm values (i.e. [much] <1mW). Some (though
not all!) drivers store this as negative values, so it needs to be passed to
ifconfig that way for printing, not cast to an unsigned value. Valid range is
something like -40 to -90dBm, so the range available with a signed char is
reasonable whether it's stored as a +ve or -ve number.
kettenis [Fri, 9 Jan 2015 20:17:05 +0000 (20:17 +0000)]
Properly unwind from a failure in usbd_dma_contig_alloc(). Calling
bus_dmamap_unload(9) on a map that failed to load is a bad idea and causes
panics on some architectures (such as sparc64).
ok mpi@
kettenis [Fri, 9 Jan 2015 14:23:25 +0000 (14:23 +0000)]
Fix loading memory allocated with bus_dmamem_alloc(9). The old could would
always load all allocated pages instead of the size specified in the
bus_dmamap_load_raw(9) call. Also fixes the corner case where a specified
boundary is less than the page size, which would always create multiple
segments, even if the specified size was smaller than the boundary.
Fixes xhci(4) on sparc64.
mpi [Fri, 9 Jan 2015 12:15:48 +0000 (12:15 +0000)]
Always check for DMA buffer overrun, not just for some specific xfers.
mpi [Fri, 9 Jan 2015 12:09:51 +0000 (12:09 +0000)]
Correctly report memory allocation failure in uhidev_set_report_async().
mpi [Fri, 9 Jan 2015 12:07:50 +0000 (12:07 +0000)]
Add an asynchronous version of uhidev_get_report(), from David Higgs
with some tweaks.
Nothing use it for the moment, but upcoming upd(4) improvements need
it.
dlg [Fri, 9 Jan 2015 11:25:05 +0000 (11:25 +0000)]
document the chips we support. lots of raid controllers use them
and im too lazy to find out what they are or list them all.
stsp [Fri, 9 Jan 2015 11:19:12 +0000 (11:19 +0000)]
Document that ftp-proxy breaks explicit RFC4217 FTP TLS. While here,
chroot and privdrop is a feature, not a bug, so move it out of CAVEATS.
ok sthen
yasuoka [Fri, 9 Jan 2015 11:17:29 +0000 (11:17 +0000)]
Add support DIOC{G,S}CACHE to mfi(4). Cache of disk on mfi(4) will
be enabled by default. Also if the controller doesn't have any RAM,
the cache on the physical disk will be enabled.
discuss with jsg tsubai
ok dlg
jsg [Fri, 9 Jan 2015 09:35:20 +0000 (09:35 +0000)]
regen
jsg [Fri, 9 Jan 2015 09:34:07 +0000 (09:34 +0000)]
add rate matching hubs seen on a C610 system
gilles [Fri, 9 Jan 2015 08:28:02 +0000 (08:28 +0000)]
rename a variable to avoid a warning
spotted by deraadt@
henning [Fri, 9 Jan 2015 08:09:39 +0000 (08:09 +0000)]
move the ignore-microsec-timestamp bit from mrt_parse to mrt_parse_dump_mp
fixes a coredump reachable via show mrt file
From: Benjamin Baier <programmer at netzbasis dot de>, ok claudio
deraadt [Fri, 9 Jan 2015 07:35:37 +0000 (07:35 +0000)]
remove excessive/wrong use of sys/param.h
peanuts -- but all work has to start somewhere.
jsg [Fri, 9 Jan 2015 07:29:45 +0000 (07:29 +0000)]
match on C610
jsg [Fri, 9 Jan 2015 07:28:14 +0000 (07:28 +0000)]
regen
jsg [Fri, 9 Jan 2015 07:27:25 +0000 (07:27 +0000)]
Add some Intel Xeon E5 v3 and C610/X99 ids.
tedu [Fri, 9 Jan 2015 05:04:22 +0000 (05:04 +0000)]
save the bufq pointer from the buf before we turn it loose so it won't
change on us. also, remove unused second arg to bufq_wait.
from pedro at bitrig via david hill.
ok beck kettenis
tedu [Fri, 9 Jan 2015 05:01:56 +0000 (05:01 +0000)]
rename desiredvnodes to initialvnodes. less of a lie. ok beck deraadt
tedu [Fri, 9 Jan 2015 04:59:54 +0000 (04:59 +0000)]
correctly handle no timeouts and make timeout handling in general better.
problem reported by Mages Simon
ok guenther
mlarkin [Fri, 9 Jan 2015 03:43:52 +0000 (03:43 +0000)]
Cleanup some macros and #defines in i386 pmap. Previously committed and
backed out because of libkvm breakage, recommitting now with libkvm fix.
ok deraadt@
dlg [Fri, 9 Jan 2015 03:34:40 +0000 (03:34 +0000)]
implement mfii_scsi_cmd_io for handling actual io. previously i got
away with being lazy and just passing everything to the vanilla
scsi cdb path.
sending io via the cdb path with chained sgls seems to trigger a
firmware fault on the new invader boards. sending the same io via
the ldio path works fine though.
tested on invader and thunderbolt boards:
mfii0 at pci1 dev 0 function 0 "Symbios Logic MegaRAID SAS3108" rev 0x02: msi
mfii0: "PERC H730 Mini", firmware 25.2.1.0037, 1024MB cache
mfii0 at pci1 dev 0 function 0 "Symbios Logic MegaRAID SAS2208" rev 0x05: msi
mfii0: "PERC H710 Mini", firmware 21.3.0-0009, 512MB cache
ok jmatthew@
lteo [Fri, 9 Jan 2015 03:32:58 +0000 (03:32 +0000)]
Add test suites for SHA-224 and SHA-384.
Relevant lines from testsuite.sha224 were verified against the test vectors in
RFC 3874, while testsuite.384 was verified against
https://www.cosic.esat.kuleuven.be/nessie/testvectors/hash/sha/Sha-2-384.unverified.test-vectors
ok millert@
millert [Thu, 8 Jan 2015 22:10:08 +0000 (22:10 +0000)]
Avoid a crash (free of bogus pointer) when there is a syntax error
reading the config socket by clearing the entire blacklist struct
instead of just the tag string. Also avoid holes in the blacklist
array on error since the code can't cope with them. OK beck@
bluhm [Thu, 8 Jan 2015 20:22:47 +0000 (20:22 +0000)]
Use the field f_file also for the forwarding file descriptor. Remove
the special f_fd field. Also remove an useless F_FORWUDP case.
test and OK jasper@
tedu [Thu, 8 Jan 2015 18:07:35 +0000 (18:07 +0000)]
increase namecache size to follow maxvnodes. seems better than desiredvnodes.
(accuracy of variable names, aside)
ok kettenis
florian [Thu, 8 Jan 2015 17:21:01 +0000 (17:21 +0000)]
Clear autoconf6 flag after autoconf6 cleanup happened, not before.
Otherwise clean up code will never run.
OK mpi@, benno@, henning@
mpi [Thu, 8 Jan 2015 15:05:44 +0000 (15:05 +0000)]
Include the Routing Table ID in "struct rtentry" and reorder the fields
such that there's no size change on, at least, amd64.
This will allow us to simplify some APIs and deprecate the remaining
flavors of "struct route".
ok claudio@, phessler@
mpi [Thu, 8 Jan 2015 14:29:18 +0000 (14:29 +0000)]
Factorize various duplicated chunks of (old and horrible) code, checking
for the validity of a given outgoing route entry into a single function.
This change is inspired from FreeBSD r111767. The function introduced
here, rt_checkgate(), should hopefully die in a near future. Why should
it die? Well, it is way too late to do such validity checks: if your
kernel can ends up in ether_output() with an invalid route, please do
not let it try to find a new one that might do the job.
Go read this function if you're wondering why you're getting messages
like:
"arpresolve: X.X.X.X: route without link local address"
Since this horrible logic has survived 20 years of copy & past and small
modifications for workarounds without a single clear commit message, let's
assume it is full of dragons and try to play safe. This factorization is
not intended to change any behavior.
With much inputs from bluhm@, tested by weerd@ and florian@ on setups
with p2p IPv6 interfaces.
ok bluhm@, benno@, florian@
djm [Thu, 8 Jan 2015 13:44:36 +0000 (13:44 +0000)]
fix ssh_config FingerprintHash evaluation order; from Petr Lautrbach
ajacoutot [Thu, 8 Jan 2015 13:29:34 +0000 (13:29 +0000)]
"default" is gone.
djm [Thu, 8 Jan 2015 13:10:58 +0000 (13:10 +0000)]
adjust for sshkey_load_file() API change
doug [Thu, 8 Jan 2015 11:08:50 +0000 (11:08 +0000)]
Convert spkac.c to the new option handling code.
input + ok jsing@
doug [Thu, 8 Jan 2015 11:06:12 +0000 (11:06 +0000)]
Convert pkcs8.c to the new option handling code.
Minor KNF in a few places too.
input + ok jsing@
doug [Thu, 8 Jan 2015 11:00:12 +0000 (11:00 +0000)]
Convert asn1pars.c to the new option handling.
Also, removed a few useless if null checks.
input from bcook@
input + ok jsing@
mpi [Thu, 8 Jan 2015 10:55:45 +0000 (10:55 +0000)]
Do not unconditionally set a link-local address.
carp(4) has a hack to update its Ethernet address which was also
generating the corresponding IPv6 link-local address. Since the
removal of the NOINET6 flag, this link-local address was generated
even if no IPv6 address has been configured on the interface.
This unbreak carp setup without v6 addresses, found the hard way by
sebastia@.
ok sebastia@, benno@, stsp@, @phessler
djm [Thu, 8 Jan 2015 10:15:45 +0000 (10:15 +0000)]
reorder hostbased key attempts to better match the default
hostkey algorithms order in myproposal.h; ok markus@
djm [Thu, 8 Jan 2015 10:14:08 +0000 (10:14 +0000)]
deprecate key_load_private_pem() and sshkey_load_private_pem()
interfaces. Refactor the generic key loading API to not require
pathnames to be specified (they weren't really used).
Fixes a few other things en passant:
Makes
ed25519 keys work for hostbased authentication (ssh-keysign
previously used the PEM-only routines).
Fixes key comment regression bz#2306: key pathnames were being lost as
comment fields.
ok markus@
ajacoutot [Thu, 8 Jan 2015 09:03:26 +0000 (09:03 +0000)]
Clarify and simplify.
yasuoka [Thu, 8 Jan 2015 07:34:05 +0000 (07:34 +0000)]
Count number of sessions correctly.
bcook [Thu, 8 Jan 2015 05:34:21 +0000 (05:34 +0000)]
log as LOG_DAEMON in log_init again.
I mirrored from smtpd a little too literally.
deraadt [Thu, 8 Jan 2015 03:41:14 +0000 (03:41 +0000)]
missing , found by Dongsheng Song