Warner Losh [Fri, 30 Jan 2009 03:41:45 +0000 (03:41 +0000)]
Spawn one fewer shells on startup. We don't use dhcp_interfaces at
all in this function, and grep shows no other instances of it
(besides, this is a function, and in a sub-shell, so all changes are
local).
David E. O'Brien [Fri, 30 Jan 2009 00:22:08 +0000 (00:22 +0000)]
For files not named on the command line, only the basename is compared
to the exclude pattern.
Change this so that "grep --exclude='*/.svn/*' -[Rr] foo *" DWIM.
Roman Divacky [Thu, 29 Jan 2009 16:51:09 +0000 (16:51 +0000)]
Define NULL to be __null in a case of gnu c++. This makes sentinel attribute
work ok in C++. Note that we enable this only for gcc 4.x for any value
of x. The __null was introduced in gcc 4.1 (in fact it was commited
12 days after release of gcc 4.0) and as we have never released any version
of FreeBSD with gcc 4.0 nor ports support gcc 4.0.x this is a safe check.
Using __GNUC_PREREQ__ would require us to include cdefs.h in params.h so
we just check __GNUC__.
Approved by: kib (mentor)
Tested by: exp build of ports (done by pav)
Tested by: make universe (done by me)
Robert Watson [Thu, 29 Jan 2009 09:32:56 +0000 (09:32 +0000)]
If a process is a zombie and we couldn't identify another useful state,
print out the state as "zombine" in preference to "unknown" when ^T is
pressed.
Jeff Roberson [Thu, 29 Jan 2009 09:22:56 +0000 (09:22 +0000)]
- Allocate apic vectors on a per-cpu basis. This allows us to allocate
more irqs as we have more cpus. This is principally useful on systems
with msi devices which may want many irqs per-cpu.
John Baldwin [Wed, 28 Jan 2009 20:35:16 +0000 (20:35 +0000)]
Use a different value for the initial control word for the FPU state for
32-bit processes. The value matches the initial setting used by
FreeBSD/i386. Otherwise, 32-bit binaries using floating point would use
a slightly different initial state when run on FreeBSD/amd64.
Sam Leffler [Wed, 28 Jan 2009 19:25:51 +0000 (19:25 +0000)]
Remove gsm hacks now that we can do this "right":
o no need for special country codes; it's sufficient to use the sku
o no need to specify bands w/ 2.4G frequencies, use the real values
o remove duplicate band specs
Sam Leffler [Wed, 28 Jan 2009 19:24:29 +0000 (19:24 +0000)]
include the channel list in list caps when -v is set; ieee channel
#'s are not available and we have to hack around the mapchan routine
but it lets us see the calibration table w/o forcing the debug regdomain
Sam Leffler [Wed, 28 Jan 2009 19:23:12 +0000 (19:23 +0000)]
pritize the channel we display with list chans so that, among other
things, 1/2 and 1/4 width channels are hidden behind the full width
channel; this is needed because they are ordered such that they
appear after in the channel table
Sam Leffler [Wed, 28 Jan 2009 19:15:52 +0000 (19:15 +0000)]
Fix 1/2 and 1/4 width channel handling:
o only include 1/2 and 1/4 width channels when they are specified in the
regulatory database description; previously we treated them as if they
were part of the band and blindly added them for 11a/g
o check the channel list returned in the devcaps to identify whether a
device supports 1/2 or 1/4 width channels on a band; this might be
better brought out as a capability bit to avoid filling the channel
list w/ 1/2 and 1/4 width channels but then cards that only support
these channels in a range of frequencies could not be described (though
right now we don't check frequency range only band)
Warner Losh [Wed, 28 Jan 2009 19:09:49 +0000 (19:09 +0000)]
Use the correct field name for the size of the sierra_id. While this
is the same size as id, and is unlikely to change, it seems better to
use the correct field here. There's no difference in the generated
code.
John Baldwin [Wed, 28 Jan 2009 19:05:18 +0000 (19:05 +0000)]
Convert the global mutex protecting the directory lookup name cache from a
mutex to a reader/writer lock. Lookup operations first grab a read lock and
perform the lookup. If the operation results in a need to modify the cache,
then it tries to do an upgrade. If that fails, it drops the read lock,
obtains a write lock, and redoes the lookup.
John Baldwin [Wed, 28 Jan 2009 18:54:56 +0000 (18:54 +0000)]
Mark cd9660 MPSAFE and add support for using shared vnode locks during
pathname lookups.
- Remove 'i_offset' and 'i_ino' from the ISO node structure and replace
them with local variables in the lookup routine instead.
- Cache a copy of 'i_diroff' for use during a lookup in a local variable.
- Save a copy of the found directory entry in a malloc'd buffer after a
successfull lookup before getting the vnode. This allows us to release
the buffer holding the directory block before calling vget() which
otherwise resulted in a LOR between "bufwait" and the vnode lock.
- Use an inlined version of vn_vget_ino() to handle races with ..
lookups. I had to inline the code here since cd9660 uses an internal
vget routine to save a disk I/O that would otherwise re-read the
directory block.
- Honor the requested locking flags during lookups to allow for shared
locking.
- Honor the requested locking flags passed to VFS_ROOT() and VFS_VGET()
similar to UFS.
- Don't make every ISO 9660 vnode hold a reference on the vnode of the
underlying device vnode of the mountpoint. The mountpoint already
holds a suitable reference.
Sam Leffler [Wed, 28 Jan 2009 18:37:46 +0000 (18:37 +0000)]
change null_getradiocaps to return the actual count of channels instead
of the true number; otherwise the caller may use the count to do things
with the data it should not (e.g. sort the channel table)
Sam Leffler [Wed, 28 Jan 2009 18:00:22 +0000 (18:00 +0000)]
Overhaul regulatory support:
o remove HAL_CHANNEL; convert the hal to use net80211 channels; this
mostly involves mechanical changes to variable names and channel
attribute macros
o gut HAL_CHANNEL_PRIVATE as most of the contents are now redundant
with the net80211 channel available
o change api for ath_hal_init_channels: no more reglass id's, no more outdoor
indication (was a noop), anM contents
o add ath_hal_getchannels to have the hal construct a channel list without
altering runtime state; this is used to retrieve the calibration list for
the device in ath_getradiocaps
o add ath_hal_set_channels to take a channel list and regulatory data from
above and construct internal state to match (maps frequencies for 900MHz
cards, setup for CTL lookups, etc)
o compact the private channel table: we keep one private channel
per frequency instead of one per HAL_CHANNEL; this gives a big
space savings and potentially improves ani and calibration by
sharing state (to be seen; didn't see anything in testing); a new config
option AH_MAXCHAN controls the table size (default to 96 which
was chosen to be ~3x the largest expected size)
o shrink ani state and change to mirror private channel table (one entry per
frequency indexed by ic_devdata)
o move ani state flags to private channel state
o remove country codes; use net80211 definitions instead
o remove GSM regulatory support; it's no longer needed now that we
pass in channel lists from above
o consolidate ADHOC_NO_11A attribute with DISALLOW_ADHOC_11A
o simplify initial channel list construction based on the EEPROM contents;
we preserve country code support for now but may want to just fallback
to a WWR sku and dispatch the discovered country code up to user space
so the channel list can be constructed using the master regdomain tables
o defer to net80211 for max antenna gain
o eliminate sorting of internal channel table; now that we use ic_devdata
as an index, table lookups are O(1)
o remove internal copy of the country code; the public one is sufficient
o remove AH_SUPPORT_11D conditional compilation; we always support 11d
o remove ath_hal_ispublicsafetysku; not needed any more
o remove ath_hal_isgsmsku; no more GSM stuff
o move Conformance Test Limit (CTL) state from private channel to a lookup
using per-band pointers cached in the private state block
o remove regulatory class id support; was unused and belongs in net80211
o fix channel list construction to set IEEE80211_CHAN_NOADHOC,
IEEE80211_CHAN_NOHOSTAP, and IEEE80211_CHAN_4MSXMIT
o remove private channel flags CHANNEL_DFS and CHANNEL_4MS_LIMIT; these are
now set in the constructed net80211 channel
o store CHANNEL_NFCREQUIRED (Noise Floor Required) channel attribute in one
of the driver-private flag bits of the net80211 channel
o move 900MHz frequency mapping into the hal; the mapped frequency is stored
in the private channel and used throughout the hal (no more mapping in the
driver and/or net80211)
o remove ath_hal_mhz2ieee; it's no longer needed as net80211 does the
calculation and available in the net80211 channel
o change noise floor calibration logic to work with compacted private channel
table setup; this may require revisiting as we no longer can distinguish
channel attributes (e.g. 11b vs 11g vs turbo) but since the data is used
only to calculate status data we can live with it for now
o change ah_getChipPowerLimits internal method to operate on a single channel
instead of all channels in the private channel table
o add ath_hal_gethwchannel to map a net80211 channel to a h/w frequency
(always the same except for 900MHz channels)
o add HAL_EEBADREG and HAL_EEBADCC status codes to better identify regulatory
problems
o remove CTRY_DEBUG and CTRY_DEFAULT enum's; these come from net80211 now
o change ath_hal_getwirelessmodes to really return wireless modes supported
by the hardware (was previously applying regulatory constraints)
o return channel interference status with IEEE80211_CHANSTATE_CWINT (should
change to a callback so hal api's can take const pointers)
o remove some #define's no longer needed with the inclusion of
<net80211/_ieee80211.h>
Ed Schouten [Wed, 28 Jan 2009 17:57:16 +0000 (17:57 +0000)]
Last step of splitting up minor and unit numbers: remove minor().
Inside the kernel, the minor() function was responsible for obtaining
the device minor number of a character device. Because we made device
numbers dynamically allocated and independent of the unit number passed
to make_dev() a long time ago, it was actually a misnomer. If you really
want to obtain the device number, you should use dev2udev().
We already converted all the drivers to use dev2unit() to obtain the
device unit number, which is still used by a lot of drivers. I've
noticed not a single driver passes NULL to dev2unit(). Even if they
would, its behaviour would make little sense. This is why I've removed
the NULL check.
Ths commit removes minor(), minor2unit() and unit2minor() from the
kernel. Because there was a naming collision with uminor(), we can
rename umajor() and uminor() back to major() and minor(). This means
that the makedev(3) manual page also applies to kernel space code now.
I suspect umajor() and uminor() isn't used that often in external code,
but to make it easier for other parties to port their code, I've
increased __FreeBSD_version to 800062.
Bjoern A. Zeeb [Wed, 28 Jan 2009 15:31:16 +0000 (15:31 +0000)]
While OpenBSD's crypto/ framework has sha1 and md5 implementations that
can cope with a result buffer of NULL in the "Final" function, we cannot.
Thus pass in a temporary buffer long enough for either md5 or sha1 results
so that we do not panic.
Luigi Rizzo [Wed, 28 Jan 2009 13:11:22 +0000 (13:11 +0000)]
For some reason (probably dating ages ago) an #ifdef SYSCTL_NODE / #endif
section included a lot of stuff that did not belong there.
So split the block in multiple components each around the relevant stuff.
This said, I wonder if building a kernel where SYSCTL_NODE is not
defined is supported at all.
Robert Watson [Wed, 28 Jan 2009 09:33:00 +0000 (09:33 +0000)]
Add a new flag to dumpfs(8), -f, which causes dumpfs to list all free
fragments in the file system by fragment (block) number. This new
mode does the necessary arithmetic to generate absolute fragment
numbers rather than than the cg-relative numbers printed in the default
mode.
If -f is passed once, contiguous fragment ranges are collapsed into
an X-Y format as free block lists are currently printed in regular
dumpfs output, but if specified twice, all block numbers are printed
individually, allowing both compact and more script-friendly
representation.
This proves quite handy when attempting to recover deleted data, as it
allows exclusion of non-deleted data from blocks searched.
MFC after: 1 week
Discussed with: jeff, Richard Clayton <richard dot clayton at cl.cam.ac.uk>
Sponsored by: Google, Inc.
Sam Leffler [Tue, 27 Jan 2009 23:42:14 +0000 (23:42 +0000)]
Remove assumptions about the max # channels in ioctl's:
o change ioctl's that pass channel lists in/out to handle variable-size
arrays instead of a fixed (compile-time) value; we do this in a way
that maintains binary compatibility
o change ifconfig so all channel list data structures are now allocated
to hold MAXCHAN entries (1536); this, for example, allows the kernel
to return > IEEE80211_CHAN_MAX entries for calls like IEEE80211_IOC_DEVCAPS
Sam Leffler [Tue, 27 Jan 2009 23:09:55 +0000 (23:09 +0000)]
o add country codes from Atheros regulatory; these are not listed in
the ISO tables, mark them accordingly
o add sku's for handling 900MHz cards
o add opaque struct defs and change []'s to *'s so this file can be
included w/o requiring all of net80211 to be pulled in
o make CTRY_DEBUG and CTRY_DEFAULT public
Robert Watson [Tue, 27 Jan 2009 21:48:47 +0000 (21:48 +0000)]
Following a fair amount of real world experience with ACLs and
extended attributes since FreeBSD 5, make the following semantic
changes:
- Don't update the inode modification time (mtime) when extended
attributes (and hence also ACLs) are added, modified, or removed.
- Don't update the inode access tie (atime) when extended attributes
(and hence also ACLs) are queried.
This means that rsync (and related tools) won't improperly think
that the data in the file has changed when only the ACL has changed.
Note that ffs_reallocblks() has not been changed to not update on an
IO_EXT transaction, but currently EAs don't use the cluster write
routines so this shouldn't be a problem. If EAs grow support for
clustering, then VOP_REALLOCBLKS() will need to grow a flag argument
to carry down IO_EXT to UFS.
MFC after: 1 week
PR: ports/125739
Reported by: Alexander Zagrebin <alexz@visp.ru>
Tested by: pluknet <pluknet@gmail.com>,
Greg Byshenk <freebsd@byshenk.net>
Discussed with: kib, kientzle, timur, Alexander Bokovoy <ab@samba.org>
When synchronizing the clock at system startup time, use both
the -g and -q options. They do a slightly different thing and
both are necessary when the time difference is large.
Noticed by: danger, in the forums
Approved by: roberto
MFC after: 1 week
Nick Hibma [Tue, 27 Jan 2009 19:37:30 +0000 (19:37 +0000)]
Fix the input buffer at 1024. The previous calculated buffer size
exceeded the maximum size of 1 page for OHCI controllers. Other serial
drivers use the same size, so I assume this should be enough (1MB/s
throughput?).
Luigi Rizzo [Tue, 27 Jan 2009 11:06:59 +0000 (11:06 +0000)]
Put dummynet-related code in a separate file.
To this purpose, add prototypes for global functions in ipfw2.h
and move there also the list of tokens used in various places in the code.
Luigi Rizzo [Tue, 27 Jan 2009 10:18:55 +0000 (10:18 +0000)]
Start splitting the monster file in smaller blocks.
In this episode:
- introduce a common header with a minimal set of common definitions;
- bring the main() function and options parser in main.c
- rename the main functions with an ipfw_ prefix
No code changes except for the introduction of a global variable,
resvd_set_number, which stores the RESVD_SET value from ip_fw.h
and is used to remove the dependency of main.c from ip_fw.h
(and the subtree of dependencies) for just a single constant.
Luigi Rizzo [Tue, 27 Jan 2009 09:27:13 +0000 (09:27 +0000)]
put the usage() function inline, it was only 1 line and used once;
slightly reformat the help() text;
slightly correct the text for the 'extraneous filename' error message;
Jeff Roberson [Tue, 27 Jan 2009 07:29:37 +0000 (07:29 +0000)]
- Add support for nehalem/corei7 cpus. This supports all of the core
counters defined in the reference manual. It does not support the
'uncore' events.
Tom Rhodes [Tue, 27 Jan 2009 00:22:16 +0000 (00:22 +0000)]
Add ENOMEM to the return values.
Remove invalid return values.
Remove reference to non-existent manual pages.
Remove reference to rfork (it does not discuss RFSTOPPED).
Add sys/unistd.h to the list of includes (required for RFSTOPPED).
PR: 126227
Submitted by: Mateusz Guzik <mjguzik@gmail.com> (based on, original version)
Reviewed by: jhb, Christoph Mallon <christoph.mallon@gmx.de>
Maksim Yevmenkin [Mon, 26 Jan 2009 20:59:41 +0000 (20:59 +0000)]
Clean up ng_ubt2. Get rid of excessive use of NG_NODE_REF/UNREF().
Make detach() completely synchronous. Properly handle stalled USB
transfers (use internal mechanism instead of submitting own control
transfers). Rename/remove a couple of variables and update comments.
This work was done in close collaboration with HPS.