- move list of supported adapters into a HARDWARE section
- list the Digital EtherWORKS adapter as supported [1]
- properly capitalize (Fast|Gigabit) Ethernet
- bump document date
Introduce ata_udelay() that uses tsleep instead of DELAY if possible.
In places where we have long delays that doesn't depend on too accurate
timing, use ata_udelay() instead of DELAY() so we dont uselessly spin
the CPU if not nessesary;
Scott Long [Fri, 3 Sep 2004 08:44:23 +0000 (08:44 +0000)]
Panic if given a CAM_DATA_PHYS pointer from CAM instead of trying to handle it.
It makes no sense in a PAE environment and is impossible to handle correctly.
This case is also never used right now. This should make the iir(4) driver
ready for PAE.
Brooks Davis [Fri, 3 Sep 2004 07:11:42 +0000 (07:11 +0000)]
If the argument to the -r flag starts with a ':' or a '|', don't try to
make sure it is a device. GDB special cases these prefixes and treats
:#### as a tcp port on localhost and executes what ever follows '|'.
Peter Pentchev [Fri, 3 Sep 2004 06:24:25 +0000 (06:24 +0000)]
Add a BUGS section and copy the wording from mmap(2)'s MAP_NOSYNC,
documenting the obsoleteness of the msync(2) syscall and its single
remaining purpose.
Alan Cox [Fri, 3 Sep 2004 05:11:32 +0000 (05:11 +0000)]
Push Giant deep into vm_forkproc(), acquiring it only if the process has
mapped System V shared memory segments (see shmfork_myhook()) or requires
the allocation of an ldt (see vm_fault_wire()).
Hiroki Sato [Thu, 2 Sep 2004 20:30:05 +0000 (20:30 +0000)]
New release notes (changes in the second half of July and
the first half of August 2004):
NO_ADAPTIVE_MUTEXES,
ADAPTIVE_GIANT,
acpi_panasonic(4) added,
contigmalloc(9) reimplemented (vm.old_contigmalloc added),
hme(4) checksum offload support,
{mem,io}.ko added,
pcic(4) removed completely from GENERIC kernel configuration file,
psm(4) Synaptics TouchPad support,
Thread Local Storage (TLS) support,
ipfw(4) antispoof option,
GEOM_MIRROR class and gmirror(8) added,
col(1), colcrt(1), colrm(1), column(1), and fmt(1) multibyte
characters support,
df(1) bugfix,
ftpd(8) uses its euid for opening a data transfer socket,
gpt(8) -i option, remove command added,
kgdb(1) added,
LIBPTHREAD_SYSTEM_SCOPE for libpthread,
look(1) -d bugfix,
make(1) "+" flag added,
make(1) MAKEFLAGS change,
tgmath.h C99 header added,
GNU extensions of mbsnrtowcs(3) and wcsnrtombs(3) implemented,
patch(1) replaced with a BSD-licensed version from OpenBSD,
ppp(8) "set pppoe [standard|3Com]" for ng_pppoe(4),
ul(1) multibyte characters support,
FILE 3.41 --> 4.10,
GCC 3.3.3 --> 3.4.2,
NTP 4.1.1.a --> 4.2.0,
rc.d localpkg support in the Ports Collection, and
crypto distribution now part of base.
Update release notes:
snd_*(4) reorganization, and
use the term "geom class" for consistency.
Robert Watson [Thu, 2 Sep 2004 20:14:03 +0000 (20:14 +0000)]
The KAME IPSEC implementation at one point used its own pseudo-random
number generator, which was re-seeded via a timeout. Now centralized
randomness/entropy is used, we can garbage collect the timeout and
re-seeding code (which was largely a no-op).
Discussed with: itojun, suz, JINMEI Tatuya < jinmei at isl dot rdc dot toshiba dot co dot jp >
Scott Long [Thu, 2 Sep 2004 18:59:15 +0000 (18:59 +0000)]
Turn PREEMPTION into a kernel option. Make sure that it's defined if
FULL_PREEMPTION is defined. Add a runtime warning to ULE if PREEMPTION is
enabled (code inspired by the PREEMPTION warning in kern_switch.c). This
is a possible MT5 candidate.
Robert Watson [Thu, 2 Sep 2004 16:04:09 +0000 (16:04 +0000)]
Back out pseudo_vnops.c:1.45, which was a workaround for pfind()
returning incompletely initialized processes. This problem was
eliminated by kern_proc.c:1.215, which causes pfind() not to
return processes in the PRS_NEW state.
Brooks Davis [Thu, 2 Sep 2004 05:07:29 +0000 (05:07 +0000)]
Back out ifi_epoch. The ABI breakage is too disruptive this close to
5-STABLE. ifi_epoch will shortly be reintroduced with less precistion
using the space currently allocated to ifi_unused.
Don't enter the debugger when executing an AML breakpoint instruction
unless ACPI_DEBUG is defined. Users don't typically care about errant
breakpoint instructions. The HP Pavilion 7915 has this in its PCI0
_INI method for rev 0x6040000 of the RSDT.
Added support for Intel PRO/1000 GT Desktop Adapter(Device ID 8086 107C)
Removed support for Intel 82541ER
Added fix for 82547 which corrects an issue with Jumbo frames larger than 10k.
Added fix for vlan tagged frames not being properly bridged.
Corrected TBI workaround.
Corrected incorrect LED operation issues
Max Laier [Wed, 1 Sep 2004 19:56:47 +0000 (19:56 +0000)]
Fix an assertion when if_down()ing a ALTQ managed interface. The lock should
have been in place all the time the mtx_assert in the ALTQ code just
discovered the shortcoming.
Alan Cox [Wed, 1 Sep 2004 19:04:09 +0000 (19:04 +0000)]
Correction to the previous revision: I forgot to apply the ones complement
to a constant. This didn't show in testing because the broken expression
produced the same result in my tests as the correct expression.
Brooks Davis [Wed, 1 Sep 2004 18:22:14 +0000 (18:22 +0000)]
Use a spare byte in struct if_data to store the structure size without
increasing it. Add code to ifconfig to use this size to find the
sockaddr_dl after the struct if_data in the routing message. This
allows struct if_data to grow (up to 255 bytes) without breaking
ifconfig.
*Blush* forgot to test non SMP builds.. oddly enough some UP code (particularly
in the acpi code) seems to want this in a UP build. (I guess so you can have
a sigle kernel module that works for both)
Disable links after getting the possible resources. Even though _DIS
should only affect current resources, it seems best to wait until all
configuration is done before disabling it. If this fixes any problems, it
is a MT5 candidate.
Ruslan Ermilov [Wed, 1 Sep 2004 14:36:48 +0000 (14:36 +0000)]
Tidy up the list of docproj dependencies. Here comes a long
explanation...
This makefile tries to be smart in determining the minimal list
of ports comprising the docproj meta-port, necessary to build
the FreeBSD documentation and release notes.
To get it absolutely right, it would have to generate the list
at run time, parsing the output of "make all-depends-list" for
ports/textproc/docproj, supplying the same options like in
src/release/Makefile, and trying for different platforms and
FreeBSD versions!
Besides from the complexity involved, this is just not possible
to do from here, because the ports tree may be not up-to-date or
not be available at all -- it's not required to have a checked
out ports tree to start "make release". So makefile copies
some logic from the ports system to try to come up with the
right list...
Unfortunately, it's nearly impossible nor practical to come up
with a proper and complete list from here this way. Such a
list would have to support all possible __FreeBSD_version's,
Perl versions, and so on, and their combinations -- everything
that the ports system would do.
The fix is to stop pretending be smart, and only provide
lists enough to build a docproj port on recent of 4.x, 5.x,
and 6.x systems, where this is practical.
What can break with this change is "make release -DNOPORTS" of
HEAD or RELENG_5, *without* -DNODOC, on older FreeBSD systems
like 5.1-RELEASE. The change in the docproj dependency list
between 5.1 and 5.3 is as follows:
Since doing "make release" on these old systems is not officially
supported anyway (only one tiny tweak is necessary to make it
work), the practical damage of this change is zero.
Original problem spotted by Scott Long doing "make release
-DNOPORTS RELEASETAG=RELENG_5". That particular problem
was already fixed by the previous change to this file, while
this change is the result of some additional analyzis of the
problem.
Fix was tested by doing a "make release -DNOPORTS" of HEAD on
the 4.10-STABLE machine (with one small tweak I mentioned, to
make it possible).
Tim Kientzle [Wed, 1 Sep 2004 05:06:42 +0000 (05:06 +0000)]
Per Ruslan, bsd.lib.mk already has support for dynamically-generated
.h files. This simplifies the Makefile here a bit and makes it behave
better in a couple of situations. While I'm here, clean up some comments
and try to improve the organization a bit.
Thanks to: Ruslan Ermilov (The Marvelous Makefile Guru)
Brooks Davis [Wed, 1 Sep 2004 00:08:15 +0000 (00:08 +0000)]
When an USB keyboard is plugged in to a machine with a builtin keyboard,
cause the USB keyboard to take over from the builtin one. This means my
laptop just works when I'm using it as a desktop.
Warner Losh [Tue, 31 Aug 2004 20:37:10 +0000 (20:37 +0000)]
When ISA_PNP_PROBE is called, it will return 0 when it finds a device,
ENOENT when there's no PNP ID for this device node, or ENXIO when there
is one, but it doesn't match.
In the nonPNP case (as most Alpha systems appear to be), we were
treating the error return as an error, when it should be have ignored
it. Version 1.9 properly ignored it, but the attach re-write of 1.10
introduced this logic error.
Pointy Hat to: phk (for breaking it then asking me to fix it :-)
Sponsored by: The Voices in Bill Paul's Head, LLC
Max Laier [Tue, 31 Aug 2004 14:23:51 +0000 (14:23 +0000)]
Don't rely on properly setup linker.hints to figure out that pflog is now
part of the pf module.
While here fix a comment that was c'n'ped from rc.d/pf
David Xu [Tue, 31 Aug 2004 11:52:05 +0000 (11:52 +0000)]
Remove TDP_USTATCLOCK, we no longer need it because we now always
update tick count for userland in thread_userret. This change
also removes a "no upcall owned" panic because fuword() schedules
an upcall under heavily loaded, and code assumes there is no upcall
can occur.
Reported and Tested by: Peter Holm <peter@holm.cc>
Julian Elischer [Tue, 31 Aug 2004 07:34:54 +0000 (07:34 +0000)]
Remove an unneeded argument..
The removed argument could trivially be derived from the remaining one.
That in turn should be the same as curthread, but it is possible that curthread could be expensive to derive on some syste,s so leave it as an argument.
Having both proc and thread as an argumen tjust gives an opportunity for
them to get out sync.
Julian Elischer [Tue, 31 Aug 2004 06:12:13 +0000 (06:12 +0000)]
Remove sched_free_thread() which was only used
in diagnostics. It has outlived its usefulness and has started
causing panics for people who turn on DIAGNOSTIC, in what is otherwise
good code.
Peter Wemm [Mon, 30 Aug 2004 23:03:58 +0000 (23:03 +0000)]
Kill count device support from config. I've changed the last few
remaining consumers to have the count passed as an option. This is
i4b, pc98/wdc, and coda.
Bump configvers.h from 500013 to 600000.
Remove heuristics that tried to parse "device ed5" as 5 units of the ed
device. This broke things like the snd_emu10k1 device, which required
quotes to make it parse right. The no-longer-needed quotes have been
removed from NOTES, GENERIC etc. eg, I've removed the quotes from:
device snd_maestro
device "snd_maestro3"
device snd_mss
I believe everything will still compile and work after this.
When configuring RAID3 with verification option, force synchronization
of parity component, because we can't return an EIO error for read of
every sector which wasn't written first.
Alan Cox [Mon, 30 Aug 2004 21:39:22 +0000 (21:39 +0000)]
- Introduce a lock for synchronizing access to the pvo and pteg tables.
- In pmap_enter(), only the acquisition and release of the page queues
lock needs to check the bootstrap flag.
Nate Lawson [Mon, 30 Aug 2004 21:35:34 +0000 (21:35 +0000)]
Fix _FDI drive type probing. The new fd child is not an ACPI device while
the old one is. Hence we need to evaluate the old one for _FDI since it
has a valid ACPI_HANDLE ivar. This is a minimal fix. Make a note that a
more complete one is to make fdc support the ACPI_HANDLE ivar for its
children.
Nate Lawson [Mon, 30 Aug 2004 21:13:03 +0000 (21:13 +0000)]
Fix _FDE probing by using the buffer contents instead of the buffer
object itself. ACPI-CA returns an ACPI_OBJECT of type Buffer rather than
the buffer contents itself.
projects / FreeBSD / FreeBSD.git / log