mjacob [Wed, 31 Oct 2001 20:43:48 +0000 (20:43 +0000)]
Make the actual volatile int smp_started live *somewhere*. This is
a temporary fix so that we can compile kernels. I waited 30 minutes
for a response from the person who would likely know, but any longer
is too long to wait with breakage at ToT.
dfr [Wed, 31 Oct 2001 20:02:28 +0000 (20:02 +0000)]
Experiment with rewriting the syscall() wrapper using explicit bundling
and trying to reduce stalls from reading certain high latency registers.
This should be faster than the old syscall code. Its certainly a lot
smaller.
marcel [Wed, 31 Oct 2001 09:03:05 +0000 (09:03 +0000)]
Make smp_started volatile in sys/smp.h and remove the volatile
declaration in subr_smp.c. This solves a compile problem with
gcc 3.0.1 (ia64 cross-build).
dillon [Wed, 31 Oct 2001 03:06:33 +0000 (03:06 +0000)]
Don't let pmap_object_init_pt() exhaust all available free pages
(allocating pv entries w/ zalloci) when called in a loop due to
an madvise(). It is possible to completely exhaust the free page list and
cause a system panic when an expected allocation fails.
nsayer [Wed, 31 Oct 2001 02:04:33 +0000 (02:04 +0000)]
Sheldon asked me to make a note of the Message-ID of the e-mail from
Dave approving the copyright statement.
The message was like this:
Date: Mon, 29 Oct 2001 13:18:36 -0500
From: David Safford <safford@watson.ibm.com>
Message-Id: <20011029181842.0DCAAD7C59@safford.watson.ibm.com>
Subject: Re: SRA copyright? In-Reply-To: Your message of "Mon, 29 Oct 2001 09:13:14 PST."
<3846.66.126.254.34.1004375594.squirrel@medusa.kfu.com>
looks fine.
dave safford
> >> Some time ago we spoke about SRA and importing it into FreeBSD. I
> >> forgot to ask if you had a prefered license boilerplate for the top of
> >> the files. It has come up recently, and the SRA code in FreeBSD
> >> doesn't have one.
> >>
> >>
> > I really have no preference - use whatever is most convenient in the
> > FreeBSD environment.
> >
> > dave safford
>
> Here is a representative diff. Let me know if you feel any changes are
> necessary.
>
> ===================================================================
> RCS file: /c/ncvs/src/crypto/telnet/libtelnet/pk.h,v
> retrieving revision 1.3
> retrieving revision 1.4
tobez [Tue, 30 Oct 2001 20:15:32 +0000 (20:15 +0000)]
Implement -e option. It modifies the output produced by sysctl(8) in
such a way that the name and the value of the variable(s) are separated
with `=' instead of the usual `: '. This is useful for producing output
that can be fed back to the sysctl utility (pasted to sysctl.conf, for
example).
markm [Tue, 30 Oct 2001 19:45:00 +0000 (19:45 +0000)]
Install libssh and libssh_pic. These are needed when building
statically, and when building things (like login(8)) standalone.
libssh_pic is needed for libpam and modules.
wpaul [Tue, 30 Oct 2001 18:15:48 +0000 (18:15 +0000)]
Fix a (long standing?) bug in ip_output(): if ip_insertoptions() is
called and ip_output() encounters an error and bails (i.e. host
unreachable), we will leak an mbuf. This is because the code calls
m_freem(m0) after jumping to the bad: label at the end of the function,
when it should be calling m_freem(m). (m0 is the original mbuf list
_without_ the options mbuf prepended.)
des [Tue, 30 Oct 2001 15:57:20 +0000 (15:57 +0000)]
Make sure the netmask always has an address family. This fixes Linux
ifconfig, which expects the address returned by the SIOCGIFNETMASK ioctl
to have a valid sa_family. Similar changes may be necessary for IPv6.
While we're here, get rid of an unnecessary temp variable.
green [Tue, 30 Oct 2001 15:54:09 +0000 (15:54 +0000)]
Add -K support to gprof(1), which enables dynamic symbol resolution from
the currently-running kernel (and supercedes an executable file argument
given). With this change, properly-compiled KLD modules are now
able to be profiled.
Obtained from: NAI Labs CBOSS project
Funded by: DARPA
imp [Tue, 30 Oct 2001 15:31:49 +0000 (15:31 +0000)]
Restore the main BAR for the bridge on resume. Some machines don't save
this accross suspend/resume events and this was causing the dreaded false
positive hit on my "static bug" test.
Note: the PCI bus code should do this for us.
Note2: We don't do the same for I/O based pci devices since it is
more code and doesn't appear to be necessary.
green [Tue, 30 Oct 2001 15:21:45 +0000 (15:21 +0000)]
Add the sysctl "kern.function_list", which currently exports all
function symbols in the kernel in a list of C strings, with an extra
nul-termination at the end.
This sysctl requires addition of a new linker operation. Now,
linker_file_t's need to respond to "each_function_name" to export
their function symbols.
Note that the sysctl doesn't currently allow distinguishing multiple
symbols with the same name from different modules, but could quite
easily without a change to the linker operation. This will be a nicety
to have when it can be used.
Obtained from: NAI Labs CBOSS project
Funded by: DARPA
green [Tue, 30 Oct 2001 15:04:57 +0000 (15:04 +0000)]
Add kmupetext(), a function that expands the range of memory covered
by the profiler on a running system. This is not done sparsely, as
memory is cheaper than processor speed and each gprof mcount() and
mexitcount() operation is already very expensive.
Obtained from: NAI Labs CBOSS project
Funded by: DARPA
jlemon [Tue, 30 Oct 2001 14:58:27 +0000 (14:58 +0000)]
When dropping a packet because there is no room in the queue (which itself
is somewhat bogus), update the statistics to indicate something was dropped.
iwasaki [Tue, 30 Oct 2001 14:24:26 +0000 (14:24 +0000)]
Some improvements of control method battery driver.
- Add a new MIB for battery info expire time in order to make it changeable.
Battery info expire time can be specified by
hw.acpi.battery.info_expire in sec.
- Add own MALLOC type and fix some potential memory leakages.
- Change some frequent printings to verbose printing.
- Stop timeout during acpi_cmbat_get_bst() too. This should reduce
the races with BIF evaluation.
- Remove acpi_cmbat_get_bif() invocation from acpi_cmbat_attach().
This was redundant because this should be called from
acpi_cmbat_timeout() now.
julian [Tue, 30 Oct 2001 07:28:17 +0000 (07:28 +0000)]
A node that allows ethernet type packets to be filtered to different
hooks depending on ethertype. Great for prototyping protocols.
connects to the lower and upper hooks of an ethernet type of node.
Obtained from: Monzoon Networks.
Thanks to Andre Oppermann, May 2001.
jake [Tue, 30 Oct 2001 06:31:45 +0000 (06:31 +0000)]
Add definitions for network support, doesn't work yet.
Pass the right arguments to the kernel.
Replace magic numbers with symbolic constants.
Pass the real openfirmware entry point to OF_init.
jake [Tue, 30 Oct 2001 06:27:34 +0000 (06:27 +0000)]
Use ENTRY() for defining functions in asm.
Remove asm functions to call the openfirmware and kernel entry points;
we can just call them directly.
Don't use the stack pointer for an intermediate result in setx.
Put the stack in the bss.
imp [Tue, 30 Oct 2001 06:08:56 +0000 (06:08 +0000)]
Move device lnc to isa section, since it no longer uses the compat shims.
Add comment about lnc.
Remove probe order comment from isa_compat.c. That appears to no longer
be the case.
dwmalone [Mon, 29 Oct 2001 20:04:03 +0000 (20:04 +0000)]
When scanning for control messages, don't process the data mbufs.
This could cause hangs if a unix domain socket was closed with data
still to be read from it.
iwasaki [Mon, 29 Oct 2001 18:09:43 +0000 (18:09 +0000)]
Some small improvements of ACPI thermal driver.
- Give a guaranteed minimum cooling run time to avoid too frequent
cooling system On/Off switching. The minimum cooling run time can be
specified by hw.acpi.thermal.min_runtime in sec.
- Refine message printing (_AC-1 -> NONE).
- Add verbose mode enable/disable capability by hw.acpi.verbose in bool.
bde [Mon, 29 Oct 2001 16:31:15 +0000 (16:31 +0000)]
Don't set CR0_NE in cpu_setregs() for the SMP case, since setting it
is npx.c's job and setting it here breaks the edit-time option of not
setting it in npx.c. (It is not set in the right places for the SMP
case, but always setting it here is harmless because there isn't even
an edit-time option to not set it.)
nsayer [Mon, 29 Oct 2001 16:12:16 +0000 (16:12 +0000)]
Add Berkeley copyright to SRA.
This is by the kind permission of Dave Safford, formerly of TAMU who wrote the
original code. Here is an excerpt of the e-mail exchange concerning this
issue:
Dave Safford wrote:
>Nick Sayer wrote:
>> Some time ago we spoke about SRA and importing it into FreeBSD. I forgot to
>> ask if you had a prefered license boilerplate for the top of the files. It
>> has come up recently, and the SRA code in FreeBSD doesn't have one.
>I really have no preference - use whatever is most convenient in the
>FreeBSD environment.
>dave safford
This is the standard BSD license with clause 3 removed and clause 4
suitably renumbered.
dfr [Mon, 29 Oct 2001 12:04:23 +0000 (12:04 +0000)]
* Factor out common code for manipulating the RSE backing store.
* Implement a fairly simplistic parser for unwinding stack frames.
* Use unwind records for DDB's 'trace' command. Also add support for
tracing past exceptions to the context which generated the exception.
The stack unwind code requires a toolchain based on binutils-2.11.2 or
later and gcc-3.0.1 or later.
peter [Mon, 29 Oct 2001 10:18:58 +0000 (10:18 +0000)]
Mostly cosmetic. Use indentation that is reasonably close to other
ia64 *.S files in our tree (eg: locore.s). Put the rest of the args
in the .section lines.
peter [Mon, 29 Oct 2001 10:10:10 +0000 (10:10 +0000)]
Update rtld for the "new" ia64 ABI. In the old toolchain, the
DT_INIT and DT_FINI tags pointed to fptr records. In 2.11.2, it points
to the actuall address of the function. On IA64 you cannot just take
an address of a function, store it in a function pointer variable and
call it.. the function pointers point to a fptr data block that has the
target gp and address in it. This is absolutely necessary for using
the in-tree binutils toolchain, but (unfortunately) will not work with
old shared libraries. Save your old ld-elf.so.1 if you want to use
old ones still. Do not mix-and-match.
marcel [Mon, 29 Oct 2001 07:34:48 +0000 (07:34 +0000)]
o Send a test IPI from the BSP to itself at the same time APs
are woken up.
o Make IPIs synchronuous by default. If we want asynchronuous
IPIs, we may want to make the memory fence controllable.
mike [Mon, 29 Oct 2001 07:29:03 +0000 (07:29 +0000)]
Use a more sophisticated check that works better with older versions
of GCC when setting __func__ to NULL. This also better supports GCC
when using C++. Move the definition to a better location.
marcel [Mon, 29 Oct 2001 05:17:36 +0000 (05:17 +0000)]
Make the clock vector 255 instead of 240. On Lion boxes, 240 is
the AP wake-up vector. We probably want a more dynamic approach
to assigning vectors in the future...
bmah [Mon, 29 Oct 2001 05:08:52 +0000 (05:08 +0000)]
Add infrastructure for an ia64 hardware list. It's a long way from
being usable at this point, so it's not even hooked up to the rest
of the RELNOTESng build; however it will be helpful for testing the content
as it approaches reality.
marcel [Mon, 29 Oct 2001 04:59:35 +0000 (04:59 +0000)]
Small correction in the LOCAL_SAPIC structure. The Flags field
starts at offset 8; not 6. Hence the structure is 12 bytes and
not 10 bytes. Adjust the definition so that the ProcessorEnabled
flag is moved from bit 15 to bit 31 in the Flags field.
bmah [Mon, 29 Oct 2001 04:45:08 +0000 (04:45 +0000)]
Start to unravel some of the mess in the Ethernet NIC section. The
original hardware list was mostly organized by driver except for the
case of NICs by Intel, DEC, and 3Com, whose NICs tended to be
organized by manufacturer. Most of these entries are now organized by
driver, except for a few 3Com NICs; also the multitude of ed(4)
drivers needs to be coalesced.
While I'm here, fix a markup glitch in the ISDN section that caused
any non-i386 hardware list to be incorrectly numbered.