openbsd
10 years agoReimplement apropos -s NUM -S ARCH EXPR by internally converting it to
schwarze [Sun, 5 Jan 2014 03:06:36 +0000 (03:06 +0000)]
Reimplement apropos -s NUM -S ARCH EXPR by internally converting it to
apropos \( EXPR \) -a 'sec~^NUM$' -a 'arch~^(ARCH|any)$'
in preparation for removal of sec and arch from the mpage table.

Almost no functional change except for the following bonus:
This also makes sure that for cross-section and cross-arch MLINKs,
all of the following work:
apropos -s 1 encrypt
apropos -s 8 encrypt
apropos -s 1 makekey
apropos -s 8 makekey
Before this commit, they don't, neither for the
old makewhatis(8)/apropos(1) combo nor for the new one.

While here, print error messages about invalid regexps to stderr.

10 years agoLike upgrade, use feed_random right after mounting filesystems. This
deraadt [Sun, 5 Jan 2014 01:56:52 +0000 (01:56 +0000)]
Like upgrade, use feed_random right after mounting filesystems.  This
is less than ideal.  We hope the kernel has good entropy already, because
this is a reinstall??  Or maybe the ftp server startup has run to completion?
Or if this is an autoinstall, that config has enough unique in it?  This
needs more work, but the existing late-entropy feed sequence is worse.
discussed at length with rpe, who tested it

10 years agoWe can feed_random right after mounting the filesystems, hoping we pick
deraadt [Sun, 5 Jan 2014 01:53:35 +0000 (01:53 +0000)]
We can feed_random right after mounting the filesystems, hoping we pick
up the previous kernel's /var/db/host.random
ok rpe

10 years agoprovide feed_random() and store_random(). The first is used by install or
deraadt [Sun, 5 Jan 2014 01:52:17 +0000 (01:52 +0000)]
provide feed_random() and store_random().  The first is used by install or
upgrade to feed as much additional usable entropy (as early as possible)
to the current bsd.rd install kernel.  The latter is used late in the
procedure to store entropy for the next boot (of a real kernel)
based on an issue described by dtucker, must conversation with halex and rpe
tested by rpe

10 years agosync
deraadt [Sun, 5 Jan 2014 01:27:58 +0000 (01:27 +0000)]
sync

10 years agoWe need /dev/random on the install media
deraadt [Sun, 5 Jan 2014 01:16:51 +0000 (01:16 +0000)]
We need /dev/random on the install media
discussed with rpe and halex

10 years agoPut section and architecture info into the keys table,
schwarze [Sun, 5 Jan 2014 00:29:49 +0000 (00:29 +0000)]
Put section and architecture info into the keys table,
in preparation for removing them from the mpages table,
aiming for cleaner and more uniform interfaces.
Database growth is below 4%, part of which will be reclaimed.

As a bonus, this allows searches like:
./obj/apropos An=kettenis -a arch=ppc
./obj/apropos An=kettenis -a sec~[^4]

10 years agoNew implementation of complex search criteria using \(, \), -a because
schwarze [Sat, 4 Jan 2014 23:42:32 +0000 (23:42 +0000)]
New implementation of complex search criteria using \(, \), -a because
the old implementation got lost in the Berkeley to SQLite switch.
Note that this is not just feature creep, but required for upcoming
database format cleanup and simplification.

The following takes less than half a second on my Z61m ThinkPad:
time ./obj/apropos An=kettenis -a \( Cd=vnet -o Xr=vbus \)

10 years agoFix of_cmap[] size. Makes the textmode cursor reappear on vgafb (i.e.
miod [Sat, 4 Jan 2014 20:28:24 +0000 (20:28 +0000)]
Fix of_cmap[] size. Makes the textmode cursor reappear on vgafb (i.e.
non-radeon) systems.

10 years agoaccept [0-9]lua as a valid section name (netbsd only);
jmc [Sat, 4 Jan 2014 18:27:03 +0000 (18:27 +0000)]
accept [0-9]lua as a valid section name (netbsd only);
from wiz@netbsd

10 years agouse standard types and formats for size_t like variables. ok dtucker
tedu [Sat, 4 Jan 2014 17:50:55 +0000 (17:50 +0000)]
use standard types and formats for size_t like variables. ok dtucker

10 years agoto do: this is a fatal error, we want a decent error message
espie [Sat, 4 Jan 2014 17:10:56 +0000 (17:10 +0000)]
to do: this is a fatal error, we want a decent error message

10 years agoProperly validate index value received from CDDB server.
tobias [Sat, 4 Jan 2014 15:39:17 +0000 (15:39 +0000)]
Properly validate index value received from CDDB server.

ok deraadt@, millert@

10 years agorecognize @option firmware
espie [Sat, 4 Jan 2014 14:14:55 +0000 (14:14 +0000)]
recognize @option firmware

10 years agomore forbidden elements
espie [Sat, 4 Jan 2014 14:13:39 +0000 (14:13 +0000)]
more forbidden elements

10 years agooops, it's global to the set, no need to do it several times
espie [Sat, 4 Jan 2014 14:13:05 +0000 (14:13 +0000)]
oops, it's global to the set, no need to do it several times

10 years agoAdd rnddata[] placeholders to make these boot blocks build again.
miod [Sat, 4 Jan 2014 10:49:21 +0000 (10:49 +0000)]
Add rnddata[] placeholders to make these boot blocks build again.

10 years agomark @vendor as old keyword so that transitory packages can be handled
espie [Sat, 4 Jan 2014 01:42:46 +0000 (01:42 +0000)]
mark @vendor as old keyword so that transitory packages can be handled
thru pkg_delete -q

10 years agosync
deraadt [Sat, 4 Jan 2014 00:20:08 +0000 (00:20 +0000)]
sync

10 years agovendor->signer
espie [Sat, 4 Jan 2014 00:14:08 +0000 (00:14 +0000)]
vendor->signer
shorten pubkey -> pub (naddy@)
\\ -> \e (jmc@)

10 years agoUse kern.securelevel to determine whether or not we are in single
millert [Fri, 3 Jan 2014 23:24:19 +0000 (23:24 +0000)]
Use kern.securelevel to determine whether or not we are in single
user mode now that init no longer raises securelevel during reboot.
OK deraadt@

10 years agosignify into the build
tedu [Fri, 3 Jan 2014 22:47:18 +0000 (22:47 +0000)]
signify into the build

10 years agoDo not raise the securelevel when transitioning from catatonia to
millert [Fri, 3 Jan 2014 22:29:00 +0000 (22:29 +0000)]
Do not raise the securelevel when transitioning from catatonia to
multiuser since we are not actually going multiuser.  Fixes a problem
where the securelevel was raised for rc.shutdown even when reboot
was run from single user mode.  OK deraadt@

10 years agoPurging a queue requires it to be non-empty, not empty.
pelikan [Fri, 3 Jan 2014 19:58:54 +0000 (19:58 +0000)]
Purging a queue requires it to be non-empty, not empty.

ok millert

10 years agodon't initialize declared variable with a function call
tedu [Fri, 3 Jan 2014 17:14:47 +0000 (17:14 +0000)]
don't initialize declared variable with a function call

10 years agoreplace hand rolled strchr with strchr
tedu [Fri, 3 Jan 2014 17:13:42 +0000 (17:13 +0000)]
replace hand rolled strchr with strchr

10 years agosynch with signify usage
espie [Fri, 3 Jan 2014 17:10:57 +0000 (17:10 +0000)]
synch with signify usage

10 years agolet signify have an actual parameters: the file to sign/verify
espie [Fri, 3 Jan 2014 17:10:27 +0000 (17:10 +0000)]
let signify have an actual parameters: the file to sign/verify
clarify SYNOPSIS, as options are highly dependent on the mode.
okay tedu@,
usage suggestion by jmc@

10 years agoPATH is now fixed
espie [Fri, 3 Jan 2014 16:56:22 +0000 (16:56 +0000)]
PATH is now fixed

10 years agosort SEE ALSO;
jmc [Fri, 3 Jan 2014 16:24:06 +0000 (16:24 +0000)]
sort SEE ALSO;

10 years agoconsistent Nd lines for the various dhcp bits; help/ok krw
jmc [Fri, 3 Jan 2014 16:21:58 +0000 (16:21 +0000)]
consistent Nd lines for the various dhcp bits; help/ok krw

10 years agonote that our date supports traditional bsd date format, which is
jmc [Fri, 3 Jan 2014 15:52:04 +0000 (15:52 +0000)]
note that our date supports traditional bsd date format, which is
different to what posix/xpg specifies;

diff originally posted to tech, changed after some feedback from
kettenis, then a bit more from schwarze and sobrado;

ok schwarze sobrado

10 years agoterminate b64 encoded lines. decoding happily skips \n.
espie [Fri, 3 Jan 2014 15:42:22 +0000 (15:42 +0000)]
terminate b64 encoded lines. decoding happily skips \n.
okay tedu@

10 years agosome more information, based on a diff by sven falempin
tedu [Fri, 3 Jan 2014 15:35:23 +0000 (15:35 +0000)]
some more information, based on a diff by sven falempin

10 years agoThe whois() function is called in a loop so make sure we close the
millert [Fri, 3 Jan 2014 15:25:18 +0000 (15:25 +0000)]
The whois() function is called in a loop so make sure we close the
socket to the whois server before returning.  Adapted from a diff
from Loganaden Velvindron.

10 years agoremove trailing whitespace and unneccessary macros from previous;
jmc [Fri, 3 Jan 2014 15:23:16 +0000 (15:23 +0000)]
remove trailing whitespace and unneccessary macros from previous;

10 years agoexplicitly tells pkg_add we're fw_update, will have some nice side-effects
espie [Fri, 3 Jan 2014 14:58:54 +0000 (14:58 +0000)]
explicitly tells pkg_add we're fw_update, will have some nice side-effects
later.

okay deraadt@

10 years agoKNF, update comments, and delete LUNA(m68k) specific DIPSW comments.
aoyama [Fri, 3 Jan 2014 13:48:25 +0000 (13:48 +0000)]
KNF, update comments, and delete LUNA(m68k) specific DIPSW comments.
No functional change.

10 years agomake sure elements are not set manually, namely forbid
espie [Fri, 3 Jan 2014 13:26:24 +0000 (13:26 +0000)]
make sure elements are not set manually, namely forbid
@depend (use -P)
@wantlib (use -W)
@name (duh)
@arch
@localbase
@option manual-installation
@comment pkgpath=...
(some of this is actually redundant, but will give more explicit error
messages)

10 years agoexplain why digital-sig must be unique
espie [Fri, 3 Jan 2014 13:12:35 +0000 (13:12 +0000)]
explain why digital-sig must be unique

10 years agofinish moving annotations around
espie [Fri, 3 Jan 2014 13:10:04 +0000 (13:10 +0000)]
finish moving annotations around

10 years agostart moving internal annotations' description into package(5) proper,
espie [Fri, 3 Jan 2014 12:54:59 +0000 (12:54 +0000)]
start moving internal annotations' description into package(5) proper,
pkg_create(1) should only document what users can use.
Accordingly, describe dependencies in terms of -P and -W directly.

10 years agoSwitch frequently allocated structs from malloc(M_DEVBUF) to separate pools.
pelikan [Fri, 3 Jan 2014 12:48:58 +0000 (12:48 +0000)]
Switch frequently allocated structs from malloc(M_DEVBUF) to separate pools.

ok henning, "looks fine" mikeb, input from guenther.

10 years agoMake queues disappear correctly on interfaces being destroyed.
pelikan [Fri, 3 Jan 2014 12:43:09 +0000 (12:43 +0000)]
Make queues disappear correctly on interfaces being destroyed.

ok henning

10 years agoregen
jsing [Fri, 3 Jan 2014 10:32:44 +0000 (10:32 +0000)]
regen

10 years agoFix previous commit, which clearly was not even compile tested, since it
jsing [Fri, 3 Jan 2014 10:31:09 +0000 (10:31 +0000)]
Fix previous commit, which clearly was not even compile tested, since it
broke the tree.

10 years agoregen
brad [Fri, 3 Jan 2014 05:52:03 +0000 (05:52 +0000)]
regen

10 years agoAdd some newer Intel wireless PCI ids from 2000, 2030, 135, 100, 3160
brad [Fri, 3 Jan 2014 05:50:23 +0000 (05:50 +0000)]
Add some newer Intel wireless PCI ids from 2000, 2030, 135, 100, 3160
and 7265 controllers.

10 years agoAdd the second 7260 PCI id.
brad [Fri, 3 Jan 2014 05:05:38 +0000 (05:05 +0000)]
Add the second 7260 PCI id.

10 years agoFix spelling in comments
guenther [Fri, 3 Jan 2014 03:00:39 +0000 (03:00 +0000)]
Fix spelling in comments

10 years agocrank version after random instruction fix from jsing
deraadt [Thu, 2 Jan 2014 23:27:36 +0000 (23:27 +0000)]
crank version after random instruction fix from jsing

10 years agoAvoid "utf8" in the names of a function and a struct member
schwarze [Thu, 2 Jan 2014 22:44:07 +0000 (22:44 +0000)]
Avoid "utf8" in the names of a function and a struct member
that don't necessarily have anything to do with UTF-8.
Just renaming, no functional change.

10 years agoDo not put UTF-8-encoded strings into the database by default, use ASCII.
schwarze [Thu, 2 Jan 2014 22:19:38 +0000 (22:19 +0000)]
Do not put UTF-8-encoded strings into the database by default, use ASCII.
Just like for mandoc(1), provide a -Tutf8 option for people who want that.

10 years agoupdate currency exchange rates;
jmc [Thu, 2 Jan 2014 21:37:28 +0000 (21:37 +0000)]
update currency exchange rates;

10 years agopesky double space;
jmc [Thu, 2 Jan 2014 21:16:37 +0000 (21:16 +0000)]
pesky double space;

10 years agoPolish the mlink_add() interface:
schwarze [Thu, 2 Jan 2014 20:24:35 +0000 (20:24 +0000)]
Polish the mlink_add() interface:
Allocate memory inside, not in the callers.
No functional change.

10 years agosimplify handling of special files some more, that stuff has fixed names
espie [Thu, 2 Jan 2014 20:13:43 +0000 (20:13 +0000)]
simplify handling of special files some more, that stuff has fixed names
anyways

10 years agoCheck all MLINKS for consistency with the content of the manual page,
schwarze [Thu, 2 Jan 2014 18:51:51 +0000 (18:51 +0000)]
Check all MLINKS for consistency with the content of the manual page,
not just the first one.  This doesn't change how the check is done,
but just which MLINKS are checked.

10 years agomake the special file creation more explicit, so that the order actually
espie [Thu, 2 Jan 2014 18:19:46 +0000 (18:19 +0000)]
make the special file creation more explicit, so that the order actually
doesn't matter

10 years agoLDFLAGS -> LDADD, fixes build on static arches.
miod [Thu, 2 Jan 2014 17:05:02 +0000 (17:05 +0000)]
LDFLAGS -> LDADD, fixes build on static arches.

10 years agoSomeone accidentally the verb (Henri Kemppainen)
espie [Thu, 2 Jan 2014 17:02:48 +0000 (17:02 +0000)]
Someone accidentally the verb (Henri Kemppainen)

10 years agodisplay proper error messages for read/write.
espie [Thu, 2 Jan 2014 16:34:02 +0000 (16:34 +0000)]
display proper error messages for read/write.
after a lot of bickering with tedu...

"fine, fix it" tedu@

10 years agoSince the functions in read.c are part of the mandoc(3) library,
schwarze [Thu, 2 Jan 2014 16:29:46 +0000 (16:29 +0000)]
Since the functions in read.c are part of the mandoc(3) library,
do not print to stderr.  Instead, properly use the mmsg callback.
Issue noticed by Abhinav Upadhyay <er dot abhinav dot upadhyay
at gmail dot com> and Thomas Klausner <wiz at NetBSD>.

10 years agobe more explicit in why we don't write DESC
espie [Thu, 2 Jan 2014 16:27:10 +0000 (16:27 +0000)]
be more explicit  in why we don't write DESC
return early instead of adding an extra level of test.

10 years agozap PKG_PREFIX (undocumented and unused, now)
espie [Thu, 2 Jan 2014 16:05:42 +0000 (16:05 +0000)]
zap PKG_PREFIX (undocumented and unused, now)
okay landry@

10 years agoCall fuse callbacks on FBT_INIT and FBT_DESTROY.
syl [Thu, 2 Jan 2014 15:39:12 +0000 (15:39 +0000)]
Call fuse callbacks on FBT_INIT and FBT_DESTROY.

Issue reported by Helg Bredow, thanks!

With inputs from tedu@
ok tedu@

10 years agoAdd primary support for LUNA-88K 4bpp/8bpp frame buffer.
aoyama [Thu, 2 Jan 2014 15:30:34 +0000 (15:30 +0000)]
Add primary support for LUNA-88K 4bpp/8bpp frame buffer.

This brings color support on LUNA's wscons.  And, with the help of recent
development version of 'mlterm-fb' (frame buffer version of ports/x11/mlterm),
graphic images can be displayed in 16/256 colors on the screen.  Thanks to
arakiken, the original developer of mlterm!

This diff is based on NetBSD/luna68k work:
http://mail-index.netbsd.org/source-changes/2013/12/28/msg050266.html

Need more work to coexist with the monochrome X.Org server.

"go ahead!" miod@

10 years agoDocument new options related to signatures
espie [Thu, 2 Jan 2014 15:01:37 +0000 (15:01 +0000)]
Document new options related to signatures

10 years agoadapt to signify changes
espie [Thu, 2 Jan 2014 13:45:14 +0000 (13:45 +0000)]
adapt to signify changes
allow a separate key for firmwares

10 years agomention hardware vlan support for 6105M chips. ok sthen@
dtucker [Thu, 2 Jan 2014 11:06:22 +0000 (11:06 +0000)]
mention hardware vlan support for 6105M chips.  ok sthen@

10 years agodisable MSI for Samsung XP941.
gilles [Thu, 2 Jan 2014 08:00:35 +0000 (08:00 +0000)]
disable MSI for Samsung XP941.
This is so I can use the machine to figure out the others things on it,
the quirk will go away when interrupt routing gets fixed.

ok dlg@, jmatthew@

10 years agosync
gilles [Thu, 2 Jan 2014 07:43:14 +0000 (07:43 +0000)]
sync

10 years agoadd vendor and product
gilles [Thu, 2 Jan 2014 07:42:33 +0000 (07:42 +0000)]
add vendor and product

ok dlg@

10 years agoSimplify things slightly by using four byte indexing.
jsing [Thu, 2 Jan 2014 04:39:56 +0000 (04:39 +0000)]
Simplify things slightly by using four byte indexing.

10 years agoEnsure that we preserve callee-save registers.
jsing [Thu, 2 Jan 2014 04:23:57 +0000 (04:23 +0000)]
Ensure that we preserve callee-save registers.

10 years agoenable beep controls on ALC221 and ALC269 as well
jsg [Thu, 2 Jan 2014 04:23:35 +0000 (04:23 +0000)]
enable beep controls on ALC221 and ALC269 as well
seems to be required when tested here

10 years agoenable beep and CD controls on ALC282
jsg [Thu, 2 Jan 2014 03:05:20 +0000 (03:05 +0000)]
enable beep and CD controls on ALC282
from and tested by Mark Rowland

10 years agoloadrandom() status should be ignored, so that we do not loop.
deraadt [Thu, 2 Jan 2014 00:03:04 +0000 (00:03 +0000)]
loadrandom() status should be ignored, so that we do not loop.
issue reported by halex

10 years agoRemove __HAVE_PMAP_PHYSSEG support, nothing uses it anymore.
miod [Wed, 1 Jan 2014 22:15:18 +0000 (22:15 +0000)]
Remove __HAVE_PMAP_PHYSSEG support, nothing uses it anymore.

10 years agoSwitch alpha to __HAVE_VM_PAGE_MD. From NetBSD.
miod [Wed, 1 Jan 2014 22:13:52 +0000 (22:13 +0000)]
Switch alpha to __HAVE_VM_PAGE_MD. From NetBSD.

10 years agorevert to mostly lowercase options, with uppercase for actions.
tedu [Wed, 1 Jan 2014 17:50:33 +0000 (17:50 +0000)]
revert to mostly lowercase options, with uppercase for actions.
ok deraadt espie naddy. wth man tweaks from jmc

10 years agoDestroy the eligible list when detaching from an interface.
pelikan [Wed, 1 Jan 2014 17:46:43 +0000 (17:46 +0000)]
Destroy the eligible list when detaching from an interface.

ok henning

10 years agoFixed typos and style.
tobias [Wed, 1 Jan 2014 16:17:55 +0000 (16:17 +0000)]
Fixed typos and style.

ok jmc@

10 years agoFix 'make bsd.umg' by skipping PT_OPENBSD_RANDOMIZE segments, as well as
mcbride [Wed, 1 Jan 2014 09:24:54 +0000 (09:24 +0000)]
Fix 'make bsd.umg' by skipping PT_OPENBSD_RANDOMIZE segments, as well as
other innocuous segment types (PT_NOTE, PT_NULL). Handle PT_LOAD, fail on
any others.

Thanks kettenis and millert for feedback

ok deraadt

10 years agofix header to get symbols loaded correctly
miod [Wed, 1 Jan 2014 08:02:09 +0000 (08:02 +0000)]
fix header to get symbols loaded correctly

10 years agoAdd a linker script to force the .openbsd.randomdata section to be put after
miod [Wed, 1 Jan 2014 07:54:15 +0000 (07:54 +0000)]
Add a linker script to force the .openbsd.randomdata section to be put after
.data, so that it will not get mapped read-only by pmap_bootstrap().

10 years agocopyright++
fgsch [Wed, 1 Jan 2014 07:08:10 +0000 (07:08 +0000)]
copyright++

10 years agoRemove #ifdef _REENTRANT code that isn't used here. These days, the code
kettenis [Tue, 31 Dec 2013 23:31:13 +0000 (23:31 +0000)]
Remove #ifdef _REENTRANT code that isn't used here.  These days, the code
handling per-thread errno lives in librthread.

ok miod@, espie@

10 years agotedu the diagnostic code to detect the presence of the defective Xterasys XN-152
brad [Tue, 31 Dec 2013 21:09:34 +0000 (21:09 +0000)]
tedu the diagnostic code to detect the presence of the defective Xterasys XN-152
32-bit PCI NIC in a 64-bit PCI slot. The code has never been enabled, never will
be and is unused / unmaintained. Update the man page to not indicate there is a
possibility of the driver detecting this hardware issue.

Input about the man page and Ok chris@

10 years agoregen
brad [Tue, 31 Dec 2013 20:37:59 +0000 (20:37 +0000)]
regen

10 years agoanother MCP61 id
brad [Tue, 31 Dec 2013 20:37:38 +0000 (20:37 +0000)]
another MCP61 id

10 years agoYet another regression introduced by Kristaps when he switched from
schwarze [Tue, 31 Dec 2013 19:39:09 +0000 (19:39 +0000)]
Yet another regression introduced by Kristaps when he switched from
Berkeley DB to SQLite3: In the .In parser, the logic got inverted.
The resulting NULL pointer access was found by clang;
scan log provided by Ulrich Spoerlein <uqs at FreeBSD>.

The best fix is to simply remove the whole, pointless custom
handler function for .In and let the framework do its work.
Now searching for included header files actually works.

While here, remove the similarly pointless custom .St handler,
fix the return value of the .Fd handler and disentangle the
spaghetti in the .Nm handler.

10 years agozap an erroneous space in "input .sig" (should be input.sig, with .sig not
jmc [Tue, 31 Dec 2013 18:18:36 +0000 (18:18 +0000)]
zap an erroneous space in "input .sig" (should be input.sig, with .sig not
marked up;

10 years agoremove assignments that will be overwritten right afterwards,
schwarze [Tue, 31 Dec 2013 18:07:06 +0000 (18:07 +0000)]
remove assignments that will be overwritten right afterwards,
and remove pointless local variables;
found in a clang output from Ulrich Spoerlein <uqs at FreeBSD>

10 years agovarious tweaks; ok tedu
jmc [Tue, 31 Dec 2013 17:33:17 +0000 (17:33 +0000)]
various tweaks; ok tedu

10 years agosync
todd [Tue, 31 Dec 2013 15:36:04 +0000 (15:36 +0000)]
sync

10 years agoOops, add missing end of block.
schwarze [Tue, 31 Dec 2013 15:15:47 +0000 (15:15 +0000)]
Oops, add missing end of block.

10 years agoAdd missing description of .UR/.UE.
schwarze [Tue, 31 Dec 2013 15:13:25 +0000 (15:13 +0000)]
Add missing description of .UR/.UE.
While here, also add .EX/.EE and .PD to the MACRO SYNTAX table.

10 years agoDrop blocks and instert silence immediately in the socket rx/tx code
ratchov [Tue, 31 Dec 2013 12:27:49 +0000 (12:27 +0000)]
Drop blocks and instert silence immediately in the socket rx/tx code
path rather than waiting for the next device cycle. This fixes deadocks
when the slot uses the SIO_SYNC mode.

10 years agoround client block size to nearest possble
ratchov [Tue, 31 Dec 2013 12:24:55 +0000 (12:24 +0000)]
round client block size to nearest possble