* Use FreeBSD's process library instead of the Solaris one.
* There are a few placeholders in here for which there isn't libproc
support code yet. This is relevent to userland tracing. This set of
commits is designed to get kernel tracing up and running, with the
userland stuff to follow later.
* Add a couple of action definitions for FreeBSD extensions.
* Handle the different ioctl design.
* Add support for the get and set error location.
* Add support for freopen().
There are many places in libdtrace where errno can be set. When an
error is returned all the way back to the dtrace app, it's hard to
figure out where that error came from.
Add a couple of functions to get and set the error location which can
be optionally compiled into the library.
* Make this file compile cleanly.
* Use FreeBSD's kld syscalls to get kernel module information instead
of using an object file system like Solaris uses.
* Get the maximum number of CPUs via a sysctl.
* Handle the different ioctl design.
* Support the freopen() changes.
* Use functions in FreeBSD's process library rather than the CDDL
library that Solaris has which sits on top of their process file
system and is therefore unsuitable for use on FreeBSD. The libproc
API for FreeBSD is deliberately different to that on Solaris because
Sun wouldn't release the libproc.h header under a BSD license.
* Define YY_INPUT for flex since input() can't be re-defined.
* On FreeBSD define both LITTLE_ENDIAN and BIG_ENDIAN and then set the
BYTE_ORDER to the one we are using. On Solaris they define one or
the other but not both. For us to keep using FreeBSD header files,
we need to use endian definitions the same way we do in pure BSD
code.
* Fix a function prototype.
* Allow for the different level of indirection between SYSV's ioctl and BSD's.
* Map a couple of sysconf definitions to one that FreeBSD has.
Solaris has code which makes zlib optional, so they have to jump through
some hoops to get the compression functions. On FreeBSD have libz and
can just link against it.
We need to be consistent with prototype definitions. It isn't OK to
use 'const' and just override it whenever we feel like it. If we use
it at all, then we need to do it properly.
Add a couple of functions that were useful in getting this code ported.
* Set the path to perl on FreeBSD.
* Use the FreeBSD shell.
* On FreeBSD the tests run from the OBJDIR, so output files go there
rather than in the source tree like they do on Solaris.
* FreeBSD doesn't need a special path to the compiler.
Have the man page catch up with the namespace pollution cleanup that
occurred between 2001-2003. Thanks to bde for the history lesson[1]
concerning sys/types.h and the many system calls that at one time
(pre-2001) were required by POSIX to include it.
Vendor import of the full userland contrib part of DTrace support from
OpenSolaris. This commit resets files to match the versions in the
OpenSolaris tree as of 2008/04/10.
The changes in this import from the previous import are the ones that
will subsequently re-applied to take files off the vendor branch. This
is unfortunately necessary because the Solaris developers won't allow
FreeBSD support #ifdefs in their source code because that creates
'dead code' (stuff that they never compile).
Replace the prototype vaps_<ifn> and vap_create_<ifn> variables with
more wlans_<ifn> and create_args_<ifn>
Add documentation for these variants and generally update the wireless
device example.
There is are very short lived shim from vaps_<ifn> which produces
a warning and vap_create_<ifn> which does not. Misuse the MFC
notification service to remind me to remove them.
Second, the issue on building modules since the static kernel
build changes is now resolved. I was not able to get the fancier
directory hierarchy working, but this works, both em and igb
build as modules now.
Third, there is now support in em for two new NICs, Hartwell
(or 82574) is a low cost PCIE dual port adapter that has MSIX,
for this release it uses 3 vectors only, RX, TX, and LINK. In
the next release I will add a second TX and RX queue. Also, there
is support here for ICH10, the followon to ICH9. Both of these are
early releases, general availability will follow soon.
Fourth: On Hartwell and ICH10 we now have IEEE 1588 PTP support,
I have implemented this in a provisional way so that early adopters
may try and comment on the functionality. The IOCTL structure may
change. This feature is off by default, you need to edit the Makefile
and add the EM_TIMESYNC define to get the code.
sam [Fri, 25 Apr 2008 19:45:42 +0000 (19:45 +0000)]
hookup the parent device's if_input and if_output to stub routines
to catch unintended use (one might argue about if_output but it's
behaviour is ill-defined without vap context)
Add "ddb capture print" and "ddb capture status" commands do ddb(8),
alowing the DDB output capture buffer to be easily extracted from
user space. Both of these commands include -M/-N arguments, allowing
them to be used with kernel crash dumps (or /dev/mem).
This makes it easier to use DDB scripting and output capture with
minidumps or full dumps rather than with text dumps, allowing DDB
output (scripted or otherwise) to be easily extracted from a crash
dump.
Initialize tz_active to a new constant TZ_ACTIVE_UNKNOWN and make no
assumptions about the state of the cooling devices. Instead, switch them
off on init and, only after that, we are in TZ_ACTIVE_NONE.
Submited by: Andriy Gapon <avg at icyb.net.ua>
Reviewed by: njl
jeff [Fri, 25 Apr 2008 05:18:50 +0000 (05:18 +0000)]
- Add an integer argument to idle to indicate how likely we are to wake
from idle over the next tick.
- Add a new MD routine, cpu_wake_idle() to wakeup idle threads who are
suspended in cpu specific states. This function can fail and cause the
scheduler to fall back to another mechanism (ipi).
- Implement support for mwait in cpu_idle() on i386/amd64 machines that
support it. mwait is a higher performance way to synchronize cpus
as compared to hlt & ipis.
- Allow selecting the idle routine by name via sysctl machdep.idle. This
replaces machdep.cpu_idle_hlt. Only idle routines supported by the
current machine are permitted.
marius [Thu, 24 Apr 2008 23:12:03 +0000 (23:12 +0000)]
o Disable HMEDEBUG by default.
o Add CTASSERTs ensuring that HME_NRXDESC and HME_NTXDESC are set to
legal values.
o Use appropriate maxsize, nsegments and maxsegsize parameters when
creating DMA tags and correct some comments related to them.
o The FreeBSD bus_dmamap_sync(9) supports ored together flags for quite
some time now so collapse calls accordingly.
o Add missing BUS_DMASYNC_PREREAD when syncing the control DMA maps in
hme_rint() and hme_start_locked().
o Keep state of the link state and use it to enable or disable the MAC
in hme_mii_statchg() accordingly as well as to return early from
hme_start_locked() in case the link is down.
o Introduce a sc_flags and use it to replace individual members like
sc_pci.
o Add bus_barrier(9) calls to hme_mac_bitflip(), hme_mii_readreg(),
hme_mii_writereg() and hme_stop() to ensure the respective bit
has been written before we starting polling on it and for the right
bits to change.
o Rather just returning in case hme_mac_bitflip() fails and leaving us
in an undefined state report the problem and move on; chances are
the requested configuration will become active shortly after.
o Don't call hme_start_locked() in hme_init_locked() unconditionally
but only after calls to hme_init_locked() when it's appropriate, i.e.
in hme_watchdog().
o Add a KASSERT which asserts nsegs is valid also to hme_load_txmbuf().
o In hme_load_txmbuf():
- use a maximum of the newly introduced HME_NTXSEGS segments instead
of the incorrect HME_NTXQ, which reflects the maximum TX queue
length, for loading the mbufs and put the DMA segments back onto
the stack instead of the softc as 16 should be ok there.
- use the common errno(2) return values instead of homegrown ones,
- given that hme_load_txmbuf() is allowed to fail resulting in a
packet drop for quite some time now implement the functionality of
hme_txcksum() by means of m_pullup(9), which de-obfuscates the code
and allows to always retrieve the correct length of the IP header, [1]
- also add a KASSERT which asserts nsegs is valid,
- take advantage of m_collapse(9) instead of m_defrag(9) for
performance reasons.
o Don't bother to check whether the interface is running or whether its
queue is empty before calling hme_start_locked() in hme_tint(), the
former will check these anyway.
o In hme_intr() call hme_rint() before hme_tint() as gem_tint() may
take quite a while to return when it calls hme_start_locked().
o Get rid of sc_debug and just check if_flags for IFF_DEBUG directly.
o Add a shadow sc_ifflags so we don't reset the chip when unnecessary.
o Handle IFF_ALLMULTI correctly. [2]
o Use PCIR_BAR instead of a homegrown macro.
o Replace sc_enaddr[6] with sc_enaddr[ETHER_ADDR_LEN].
o Use the maximum of 256 TX descriptors for better performance as using
all of them has no additional static cost rather than using just half
of them.
marius [Thu, 24 Apr 2008 22:48:34 +0000 (22:48 +0000)]
- Use bus_{read,write}_*(9) instead of bus_space_{read,write}_*(9)
in order to get rid of bus space handle and tag in struct sym_hcb.
- Remove unused members related to bus addresses in struct sym_hcb.
- sym(4) takes care of allocating an instance of struct sym_hcb
itself so don't let newbus allocate it as an unused softc also.
- Add basic MPSAFE locking. This includes changing the sym(4) CCBs
to be allocated up-front instead of on demand as needed. Besides
making these allocations more likely to succeed, this also solves
the problem of calling bus_dmamap_create(9) with the SIM mutex
held.
marius [Thu, 24 Apr 2008 22:36:11 +0000 (22:36 +0000)]
- Use NULL instead of 0 for pointers.
- Remove superfluous returns in functions returning void.
- In sym_alloc_lcb_tags() return directly instead of jumping
to a label which just returns.
- Fix some spelling in comments.
- Remove trailing whitespace.
Use logic or, not binary or, when deciding whether or not a system call
exit requires entering the audit code. The result is much the same,
but they mean different things.
MFC afer: 3 days
Submitted by: Diego Giagio <dgiagio at gmail dot com>
Allow the vnode zone to return the unused memory. The vnode reference
count is/shall be properly maintained for the long time, and VFS
shall be safe against the vnode memory reclamation.
Updated the test to handle multiple sinks. The sinks cut their
transmissions by the number of them running so that they do not
overwhelm the source.
Added a simple shell script to kick off sinks on multiple hosts as
well as a source on the host where the shell script is run. The script
also collects the output of all the sinks and the source into files named
for the host on which the tests are run. A date is appended to each output
file to make it unique per run.
Add support for gpart:
o Correct for gpart's 1-based index, versus 0-based index used by
legacy slicers.
o Parse and understand the xs and xt parameters.
Revert rev 1.332 and keep ddb scripts off by default for now. Minidumps
are more flexable and much text-dump like output can be produced from
them so there's a good argument they are a better default.
Implement the G_PART_DUMPCONF method for all 6 schemes. Also call
the method for the (indent == NULL) case (i.e. the kern.geom.conftxt
sysctl). The purpose is to extend the conftxt output with scheme-
specific fields which can be used by libdisk. In particular, have
the schemes dump the xs and xt fields, which contain the backward
compatible values for class type and partition type. This allows
libdisk to work with the legacy slicers as well as with gpart and
helps/promotes migration.
marius [Wed, 23 Apr 2008 20:04:38 +0000 (20:04 +0000)]
o Rename ic_eoi to ic_clear to emphasize the functions it points
don't send and EOI which works like on amd64/i386 and blocks all
interrupts on the relevant interrupt controller.
o Replace the post_filter and post_inthread hooks registered when
creating the interrupt events with just ic_clear as on sparc64 we
don't need to do any disable->EOI->enable dance to unblock all but
the relevant interrupt while running the filter or handler; just
not clearing the interrupt already has the same effect.
o Merge from amd64/i386:
- Split the intr_table_lock into an sx lock used for most things,
and a spin lock to protect intrcnt_index.
- Add support for binding interrupts to CPUs, including for the
bus_bind_intr(9) interface, a assign_cpu hook and initially
shuffling interrupts arround in a round-robin fashion.
Implement red-black trees without using parent pointers, and store the
color bit in the least significant bit of the right child pointer, in
order to reduce red-black tree linkage overhead by ~2X as compared to
sys/tree.h.
Use the new red-black tree implementation in malloc, which drops
memory usage by ~0.5 or ~1%, for 32- and 64-bit systems, respectively.
Add four utility functions related to struct grp processing modeled in-part
after similar calls related to struct pwd in libutil/pw_util.c:
- gr_equal()
Perform a deep comparison of two struct grp's. It does a thorough, yet
unoptimized comparison of all the members regardless of order.
- gr_make()
Create a string (see group(5)) from a struct grp.
- gr_dup()
Duplicate a struct grp. Returns a value that is a single contiguous
block of memory.
- gr_scan()
Create a struct grp from a string (as produced by gr_make()).
Differentiate in the tests against getenv() between an empty string and a
NULL by having the test program return *NULL* for NULL's (no existing
variable) instead of an empty string (variable with empty value).
Now that all platforms use genclock, shuffle things around slightly
for better structure.
Much of this is related to <sys/clock.h>, which should really have
been called <sys/calendar.h>, but unless and until we need the name,
the repocopy can wait.
In general the kernel does not know about minutes, hours, days,
timezones, daylight savings time, leap-years and such. All that
is theoretically a matter for userland only.
Parts of kernel code does however care: badly designed filesystems
store timestamps in local time and RTC chips almost universally
track time in a YY-MM-DD HH:MM:SS format, and sometimes in local
timezone instead of UTC. For this we have <sys/clock.h>
<sys/time.h> on the other hand, deals with time_t, timeval, timespec
and so on. These know only seconds and fractions thereof.
Move inittodr() and resettodr() prototypes to <sys/time.h>.
Retain the names as it is one of the few surviving PDP/VAX references.
Move startrtclock() to <machine/clock.h> on relevant platforms, it
is a MD call between machdep.c/clock.c. Remove references to it
elsewhere.
Remove a lot of unnecessary <sys/clock.h> includes.
Move the machdep.disable_rtc_set sysctl to subr_rtc.c where it belongs.
XXX: should be kern.disable_rtc_set really, it's not MD.
Fix a leak in the recent fixes for file descriptors > SHRT_MAX. In the
case of a file descriptor we can't handle, clear the FILE structure's flags
so it can be reused.
jeff [Tue, 22 Apr 2008 12:34:16 +0000 (12:34 +0000)]
- Use a local variable for i_ino in ufs_lookup. It is only used to
communicate between two parts of this one function. This was causing
problems with shared lookups as each would trash the ino value in the
inode.
- Remove the unused i_ino field from the inode structure.
marius [Mon, 21 Apr 2008 20:40:42 +0000 (20:40 +0000)]
* gthr-posix.h (__gthread_active_p): Use the Solaris implementation
for FreeBSD as well.
This is the fix for __gthread_active_p() returning false positives
which was committed as rev. 1.1.1.8.2.1 to RELENG_7 but now looped
back to the vendor branch via the GCC repository and relicensed to
be GPLv2 by me.
Thanks go to gerald@ for getting the fix approved upstream and for
committing to the GCC repository.