]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
21 years agoUnbreak kernel build.
mux [Wed, 13 Nov 2002 17:34:12 +0000 (17:34 +0000)]
Unbreak kernel build.

Submitted by: Hiten Pandya <hiten@angelica.unixdaemons.com>

21 years agoAdd ia64 documents to RELNOTESng snapshot.
bmah [Wed, 13 Nov 2002 17:28:44 +0000 (17:28 +0000)]
Add ia64 documents to RELNOTESng snapshot.

21 years agoDon't set atime / mtime if we don't have an output file, which can happen
des [Wed, 13 Nov 2002 16:04:20 +0000 (16:04 +0000)]
Don't set atime / mtime if we don't have an output file, which can happen
in certain failure scenarii (mostly DNS trouble).

21 years agoIntroduce a condition variable to avoid returning EBUSY when
rwatson [Wed, 13 Nov 2002 15:47:09 +0000 (15:47 +0000)]
Introduce a condition variable to avoid returning EBUSY when
the MAC policy list is busy during a load or unload attempt.
We assert no locks held during the cv wait, meaning we should
be fairly deadlock-safe.  Because of the cv model and busy
count, it's possible for a cv waiter waiting for exclusive
access to the policy list to be starved by active and
long-lived access control/labeling events.  For now, we
accept that as a necessary tradeoff.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoAdd support for the C99 %t format modifier.
mux [Wed, 13 Nov 2002 15:15:59 +0000 (15:15 +0000)]
Add support for the C99 %t format modifier.

21 years agoAdd the sys/stddef.h header, so that we can have ptrdiff_t
mux [Wed, 13 Nov 2002 15:14:57 +0000 (15:14 +0000)]
Add the sys/stddef.h header, so that we can have ptrdiff_t
definition in the kernel.  It also includes the offsetof()
definition which, according to bde@ should be here and not
in sys/types.h, and the definition of NULL so that we're
closer to make it closer to standard C.

Reviewed by: mike

21 years agoTake __FreeBSD_version into account when BOOTSTRAPPING.
ru [Wed, 13 Nov 2002 13:49:29 +0000 (13:49 +0000)]
Take __FreeBSD_version into account when BOOTSTRAPPING.

21 years agoActually check if stdbool.h exists when BOOTSTRAPPING.
ru [Wed, 13 Nov 2002 13:47:48 +0000 (13:47 +0000)]
Actually check if stdbool.h exists when BOOTSTRAPPING.

21 years agomtree(8) wasn't a bootstrap tool since 2000/07/23 (Makefile.inc1,v 1.161).
ru [Wed, 13 Nov 2002 12:03:27 +0000 (12:03 +0000)]
mtree(8) wasn't a bootstrap tool since 2000/07/23 (Makefile.inc1,v 1.161).

21 years agoBootstrapping aid for 4.0-RELEASE.
ru [Wed, 13 Nov 2002 11:50:40 +0000 (11:50 +0000)]
Bootstrapping aid for 4.0-RELEASE.

21 years agoFix a constant in the standard namespace not to depend on another
mike [Wed, 13 Nov 2002 11:49:24 +0000 (11:49 +0000)]
Fix a constant in the standard namespace not to depend on another
constant in the BSD namespace.

21 years agoDue to a memory alignment sizeof(struct ipfw_flow_id) is bigger than
maxim [Wed, 13 Nov 2002 11:31:44 +0000 (11:31 +0000)]
Due to a memory alignment sizeof(struct ipfw_flow_id) is bigger than
ipfw_flow_id structure actual size and bcmp(3) may fail to compare
them properly. Compare members of these structures instead.

PR: kern/44078
Submitted by: Oleg Bulyzhin <oleg@rinet.ru>
Reviewed by: luigi
MFC after: 2 weeks

21 years agoDocument loader tunables hw.pci.enable_io_modes and
mdodd [Wed, 13 Nov 2002 09:43:53 +0000 (09:43 +0000)]
Document loader tunables hw.pci.enable_io_modes and
hw.pci.allow_unsupported_io_range.

Submitted by:  Hiten Pandya <hiten@angelica.unixdaemons.com>
Approved by:  re (murray)

21 years agoConvert kernel compile option PCI_ALLOW_UNSUPPORTED_IO_RANGE to
mdodd [Wed, 13 Nov 2002 09:42:25 +0000 (09:42 +0000)]
Convert kernel compile option PCI_ALLOW_UNSUPPORTED_IO_RANGE to
a loader tunable hw.pci.allow_unsupported_io_range.

Submitted by:  Hiten Pandya <hiten@angelica.unixdaemons.com>
Approved by:  re (murray)

21 years agoStaticize local variable.
mdodd [Wed, 13 Nov 2002 09:40:09 +0000 (09:40 +0000)]
Staticize local variable.

Submitted by:  Hiten Pandya <hiten@angelica.unixdaemons.com>
Obtained from:  re (murray)

21 years agoLoader tunable 'machdep.disable_mtrrs'.
mdodd [Wed, 13 Nov 2002 09:37:43 +0000 (09:37 +0000)]
Loader tunable 'machdep.disable_mtrrs'.
Sysctl of same name to reflect status.

Submitted by:  jhb
Approved by:  re (murray)
MFC after:  1 day

21 years agoAlso test for type efi everywhere we currently test for type fat.
marcel [Wed, 13 Nov 2002 05:39:59 +0000 (05:39 +0000)]
Also test for type efi everywhere we currently test for type fat.
With this change there's no a priori difference between EFI and
FAT partitions. With this change and the corresponding change to
libdisk, we can create EFI partitions, just like regular FAT
partitions.

21 years agoMove pmap_collect() out of the machine-dependent code, rename it
alc [Wed, 13 Nov 2002 05:39:58 +0000 (05:39 +0000)]
Move pmap_collect() out of the machine-dependent code, rename it
to reflect its new location, and add page queue and flag locking.

Notes: (1) alpha, i386, and ia64 had identical implementations
of pmap_collect() in terms of machine-independent interfaces;
(2) sparc64 doesn't require it; (3) powerpc had it as a TODO.

21 years agoHandle EFI partitions the same as regular FAT partitions. The only
marcel [Wed, 13 Nov 2002 05:31:32 +0000 (05:31 +0000)]
Handle EFI partitions the same as regular FAT partitions. The only
difference between the two from a low-level point of view is that
the partition type is different. This change adds EFI related cases
to existing switch statements with existing FAT related cases.

21 years agoMajor #46 was also MIA.
peter [Wed, 13 Nov 2002 01:42:59 +0000 (01:42 +0000)]
Major #46 was also MIA.

21 years agoPut echo on a diet, removing unnecessary use of stdio and getopt.
njl [Wed, 13 Nov 2002 01:39:02 +0000 (01:39 +0000)]
Put echo on a diet, removing unnecessary use of stdio and getopt.

Before...
-r-xr-xr-x  1 root  wheel  58636 Oct 28 05:16 /bin/echo
After...
-rwxr-xr-x  1 root  wheel  12824 Nov 12 17:39 /usr/obj/usr/src/bin/echo/echo

Submitted by: Tim Kientzle <kientzle@acm.org>

21 years agoNote that 101 is unused, rather than leave it out of the file.
peter [Wed, 13 Nov 2002 00:25:01 +0000 (00:25 +0000)]
Note that 101 is unused, rather than leave it out of the file.
Add some historical information.  This can be removed as majors
are recycled.

21 years agoAdd LOGIN_SETMAC to the list of flags that can't be set without class
rwatson [Tue, 12 Nov 2002 22:30:01 +0000 (22:30 +0000)]
Add LOGIN_SETMAC to the list of flags that can't be set without class
information, since we rely on the pwd entry to know what MAC labels
to set as part of the login process.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoUse a clean flags variable when creating chunks from scripts instead of
jhb [Tue, 12 Nov 2002 21:18:54 +0000 (21:18 +0000)]
Use a clean flags variable when creating chunks from scripts instead of
leaking flags from earlier chunks into later ones.

PR: bin/40655
Submitted by: Thomas Zenker <thz@Lennartz-electronic.de>

21 years agoFix a "bug" in sysctl(8). Limit the length when we print a string from
peter [Tue, 12 Nov 2002 21:18:21 +0000 (21:18 +0000)]
Fix a "bug" in sysctl(8).  Limit the length when we print a string from
the kernel rather than ignoring the length and keeping on going till we
finally hit a \0 character in the buffer.

21 years agoRemove a line that set the status to success. We already do that at the
jhb [Tue, 12 Nov 2002 21:12:42 +0000 (21:12 +0000)]
Remove a line that set the status to success.  We already do that at the
beginning, so the best this could achieve would be to mask an earlier
failure.  Break instead of continue for another failure case.

21 years agoTry to cleanup the non-interactive disk labeling code a bit. Rework
jhb [Tue, 12 Nov 2002 21:09:58 +0000 (21:09 +0000)]
Try to cleanup the non-interactive disk labeling code a bit.  Rework
the loop that runs through the environment variables to be a bit more
intuitive.  Also, change some 'continue's in failure cases to 'break's
instead.  If we are going to fail, we should just do it.

PR: bin/40654
Submitted by: Thomas Zenker <thz@Lennartz-electronic.de> (partially)

21 years agoDo a bit of cleanup. new_part() basically ignored the passed in size
jhb [Tue, 12 Nov 2002 20:48:49 +0000 (20:48 +0000)]
Do a bit of cleanup.  new_part() basically ignored the passed in size
argument as of revision 1.52 (July 12, 1996, about a month after I
graduated from high school) when 'newfs -u' support was axed, so remove it.
This also allows us to remove a hack in the create partition case where we
created the partition twice since we didn't have the size the first time.

21 years agoTurn off duplicate lock checking for inp locks because udp_input()
hsu [Tue, 12 Nov 2002 20:44:38 +0000 (20:44 +0000)]
Turn off duplicate lock checking for inp locks because udp_input()
intentionally locks two inp records simultaneously.

21 years agoDoh, fix a bug in previous commit. The default is to newfs for new
jhb [Tue, 12 Nov 2002 20:40:15 +0000 (20:40 +0000)]
Doh, fix a bug in previous commit.  The default is to newfs for new
partitions, not to !newfs.

21 years agoWhen setting the mountpoint name, remember any previous setting of the
jhb [Tue, 12 Nov 2002 20:26:52 +0000 (20:26 +0000)]
When setting the mountpoint name, remember any previous setting of the
newfs flag for this partition.

PR: bin/31837
Reported by: Oliver Breuninger <ob@www.partner.de>

21 years agoMake 'agp', 'tun', and 'if_txp' be modules on the mfsroot floppy instead of
jhb [Tue, 12 Nov 2002 19:30:00 +0000 (19:30 +0000)]
Make 'agp', 'tun', and 'if_txp' be modules on the mfsroot floppy instead of
just removing them from BOOTMFS using dokern.sh.

21 years agoFix bogus return values from libc_r's writev() routine in situations where
archie [Tue, 12 Nov 2002 19:01:49 +0000 (19:01 +0000)]
Fix bogus return values from libc_r's writev() routine in situations where
a partial-write is followed by an error.

PR: 43335
MFC after: 3 days

21 years agorelease.9 for ia64 cannot be currently cross-built -- gpt(8)
ru [Tue, 12 Nov 2002 17:51:02 +0000 (17:51 +0000)]
release.9 for ia64 cannot be currently cross-built -- gpt(8)
is built on ia64 only.

21 years ago[DAIVD O'BRIEN's OPINION]
obrien [Tue, 12 Nov 2002 17:31:12 +0000 (17:31 +0000)]
[DAIVD O'BRIEN's OPINION]
Head off what I think is an abuse of the TRB, and disable lukemftpd.

21 years ago- Add fwmem_xfer_req() and use this in other functions.
simokawa [Tue, 12 Nov 2002 17:19:21 +0000 (17:19 +0000)]
- Add fwmem_xfer_req() and use this in other functions.

21 years ago- Remove unused code.
simokawa [Tue, 12 Nov 2002 17:16:20 +0000 (17:16 +0000)]
- Remove unused code.
- Fix style.

21 years agofix wording.
simokawa [Tue, 12 Nov 2002 17:11:32 +0000 (17:11 +0000)]
fix wording.

21 years agoSplit userland services to fwdev.c.
simokawa [Tue, 12 Nov 2002 15:22:19 +0000 (15:22 +0000)]
Split userland services to fwdev.c.

21 years agoFix XSI (X/Open) namespace support.
mike [Tue, 12 Nov 2002 14:41:49 +0000 (14:41 +0000)]
Fix XSI (X/Open) namespace support.

21 years agoo Fix usage().
maxim [Tue, 12 Nov 2002 14:15:59 +0000 (14:15 +0000)]
o Fix usage().
o Explicitly initialize domain pointer.
o Fix passwd file parsing.

PR: bin/39671 (3)
MFC after: 2 weeks

21 years agoUse struct fw_device in fwmem.c
simokawa [Tue, 12 Nov 2002 13:49:17 +0000 (13:49 +0000)]
Use struct fw_device in fwmem.c
- specify fwmem target by EUI64.
- fw_noderesolve returns struct fw_device rather than dst.
- include struct firewire_comm in struct fw_device.

21 years agoAdd a prototype for suspend/resume.
simokawa [Tue, 12 Nov 2002 13:46:09 +0000 (13:46 +0000)]
Add a prototype for suspend/resume.

21 years agoChange interface of fwmem_read_*() functions to be usable from other part
simokawa [Tue, 12 Nov 2002 11:15:01 +0000 (11:15 +0000)]
Change interface of fwmem_read_*() functions to be usable from other part
of the module.

21 years agoChange interface of fwmem_read_*() functions to be usable from other part
simokawa [Tue, 12 Nov 2002 11:08:50 +0000 (11:08 +0000)]
Change interface of fwmem_read_*() functions to be usable from other part
of the module.

21 years agoUse better return types and a couple of casts to eliminate warnings on
scottl [Tue, 12 Nov 2002 10:22:49 +0000 (10:22 +0000)]
Use better return types and a couple of casts to eliminate warnings on
alpha.  This will take the file out of sync with the private version
that we maintain, but alpha tinderbox has been broken for too long.

Tested on: i386, sparc64, alpha

21 years ago- clean up
simokawa [Tue, 12 Nov 2002 10:02:50 +0000 (10:02 +0000)]
- clean up

21 years agoFix two typos from the previous commit. This code is definitely infectious.
scottl [Tue, 12 Nov 2002 09:45:59 +0000 (09:45 +0000)]
Fix two typos from the previous commit.  This code is definitely infectious.

21 years agoStep one of cleaning and fixing cardbus:
scottl [Tue, 12 Nov 2002 08:23:27 +0000 (08:23 +0000)]
Step one of cleaning and fixing cardbus:
 - Fix some especially bad style in the CIS BAR tuple parsing code.
 - activate Option ROMS correctly.
 - de-obfuscate the Option ROM image selection code.
 - Fix mis-interpretation of the PCI spec that prevented Option ROMs whose
   CIS section wasn't in the first image from working.
 - Fix mis-interpretation of the PCI spec that prevented CIS's mapped into
   MEMIO space from working at all.
 - Reject invalid CIS pointers.

Reviewed by: imp

21 years agoWe don't use libpam, libopie, or libmd.
obrien [Tue, 12 Nov 2002 07:41:59 +0000 (07:41 +0000)]
We don't use libpam, libopie, or libmd.

21 years agoWe have fparseln(3). Also libskey on RELENG_4.
obrien [Tue, 12 Nov 2002 07:37:15 +0000 (07:37 +0000)]
We have fparseln(3).  Also libskey on RELENG_4.

21 years agoTurn off devd until I have a devd.conf file that I can install one
imp [Tue, 12 Nov 2002 07:10:45 +0000 (07:10 +0000)]
Turn off devd until I have a devd.conf file that I can install one
that does no harm.

21 years agoDon't divide by zero to calculate the ETA, even if we haven't gotten
fenner [Tue, 12 Nov 2002 07:09:43 +0000 (07:09 +0000)]
Don't divide by zero to calculate the ETA, even if we haven't gotten
any data in 30 seconds.

21 years agoUpdate for version 1.2 Beta 2.
obrien [Tue, 12 Nov 2002 06:48:35 +0000 (06:48 +0000)]
Update for version 1.2 Beta 2.

21 years agoThis commit was generated by cvs2svn to compensate for changes in r106792,
obrien [Tue, 12 Nov 2002 06:42:39 +0000 (06:42 +0000)]
This commit was generated by cvs2svn to compensate for changes in r106792,
which included commits to RCS files with non-trunk default branches.

21 years agoImport of LukeM's ftpd version 1.2 Beta 2.
obrien [Tue, 12 Nov 2002 06:42:39 +0000 (06:42 +0000)]
Import of LukeM's ftpd version 1.2 Beta 2.

21 years ago- Remove unused code.
simokawa [Tue, 12 Nov 2002 05:52:01 +0000 (05:52 +0000)]
- Remove unused code.
- Style fixes.

21 years agoRestrict process count in fwohci_arcv() and fwohci_ircv().
simokawa [Tue, 12 Nov 2002 04:54:30 +0000 (04:54 +0000)]
Restrict process count in fwohci_arcv() and fwohci_ircv().

21 years agoGarbage collect mac_create_devfs_vnode() -- it hasn't been used since
rwatson [Tue, 12 Nov 2002 04:20:36 +0000 (04:20 +0000)]
Garbage collect mac_create_devfs_vnode() -- it hasn't been used since
we brought in the new cache and locking model for vnode labels.  We
now rely on mac_associate_devfs_vnode().

Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoSchedule an idle context to block until timeouts expire without blocking
mini [Tue, 12 Nov 2002 00:55:01 +0000 (00:55 +0000)]
Schedule an idle context to block until timeouts expire without blocking
further upcalls.

21 years ago- Document the chunk_name() function that replaces chunk_n[].
davidc [Mon, 11 Nov 2002 19:37:19 +0000 (19:37 +0000)]
- Document the chunk_name() function that replaces chunk_n[].
- Rearrange things a tiny bit.

21 years agoRemove parameter names from function prototypes (at least one collided
davidc [Mon, 11 Nov 2002 18:55:57 +0000 (18:55 +0000)]
Remove parameter names from function prototypes (at least one collided
with stdlib.h).

discussed with: phk

21 years agoRemove my 'pre-commit' on sa(4).
mjacob [Mon, 11 Nov 2002 16:56:40 +0000 (16:56 +0000)]
Remove my 'pre-commit' on sa(4).

21 years agoCorrect an assertion in the code to traverse the list of locks to find an
jhb [Mon, 11 Nov 2002 16:36:20 +0000 (16:36 +0000)]
Correct an assertion in the code to traverse the list of locks to find an
earlier acquired lock with the same witness as the lock currently being
acquired.  If we had released several earlier acquired locks after
acquiring enough locks to require another lock_list_entry bucket in the
lock list, then subsequent lock_list_entry buckets could contain only one
lock instance in which case i would be zero.

Reported by: Joel M. Baldwin <qumqats@outel.org>

21 years agoCapitalize the first letter of device descriptions
mux [Mon, 11 Nov 2002 15:26:08 +0000 (15:26 +0000)]
Capitalize the first letter of device descriptions
for consistency with the rest of the drivers.

21 years agoGarbage collect definition of M_MACOPVEC -- we no longer perform a
rwatson [Mon, 11 Nov 2002 14:15:58 +0000 (14:15 +0000)]
Garbage collect definition of M_MACOPVEC -- we no longer perform a
dynamic mapping of an operation vector into an operation structure,
rather, we rely on C99 sparse structure initialization.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoRemove extra #include<sys/vmmeter.h>.
cognet [Mon, 11 Nov 2002 13:57:50 +0000 (13:57 +0000)]
Remove extra #include<sys/vmmeter.h>.

21 years agoAdd Ian Freislich's excellent, BSD-licensed POP3 server
sheldonh [Mon, 11 Nov 2002 12:32:30 +0000 (12:32 +0000)]
Add Ian Freislich's excellent, BSD-licensed POP3 server
(ports/mail/popd) to the list of packages that should appear on CDROM1.

Approved by: re

21 years agokenv(1) is WARNS=6 clean.
mux [Mon, 11 Nov 2002 11:56:28 +0000 (11:56 +0000)]
kenv(1) is WARNS=6 clean.

21 years agoatomic_set_8 isn't MI. Instead, follow Jake's suggestions about
mjacob [Mon, 11 Nov 2002 11:50:03 +0000 (11:50 +0000)]
atomic_set_8 isn't MI. Instead, follow Jake's suggestions about
ZONE_LOCK.

21 years agoSort by floppy, type, driver name.
ru [Mon, 11 Nov 2002 11:44:57 +0000 (11:44 +0000)]
Sort by floppy, type, driver name.

21 years agoRemoved redundant entries (these are handled by drivers.conf).
ru [Mon, 11 Nov 2002 11:02:14 +0000 (11:02 +0000)]
Removed redundant entries (these are handled by drivers.conf).

21 years agoUpdate to match recent snp fixes.
tmm [Mon, 11 Nov 2002 10:59:28 +0000 (10:59 +0000)]
Update to match recent snp fixes.

Reviewed by: dd

21 years agoThe FIONREAD sysctl operates on an int *, not a size_t *.
tmm [Mon, 11 Nov 2002 10:56:57 +0000 (10:56 +0000)]
The FIONREAD sysctl operates on an int *, not a size_t *.

Reviewed by: dd

21 years agoDon't attempt to strip obsolete, non-existing, or commented out
ru [Mon, 11 Nov 2002 10:53:28 +0000 (10:53 +0000)]
Don't attempt to strip obsolete, non-existing, or commented out
options and devices.

21 years agoDeclare the snp ioctl()s to work on udev_t, since that is what they
tmm [Mon, 11 Nov 2002 10:45:31 +0000 (10:45 +0000)]
Declare the snp ioctl()s to work on udev_t, since that is what they
really do and dev_t is defined differently in kernel and userland.
Return a correctly formed udev from SNPGTTY.

Reviewed by: dd

21 years agoWarning fixes.
mux [Mon, 11 Nov 2002 10:28:44 +0000 (10:28 +0000)]
Warning fixes.

21 years agoWarning fix.
mux [Mon, 11 Nov 2002 10:17:01 +0000 (10:17 +0000)]
Warning fix.

21 years agoThere's no longer a P1003_1B option.
ru [Mon, 11 Nov 2002 10:15:03 +0000 (10:15 +0000)]
There's no longer a P1003_1B option.

21 years agoTabulate, sort.
ru [Mon, 11 Nov 2002 10:13:30 +0000 (10:13 +0000)]
Tabulate, sort.

21 years agoDon't override CWARNFLAGS in these Makefiles.
mux [Mon, 11 Nov 2002 10:11:59 +0000 (10:11 +0000)]
Don't override CWARNFLAGS in these Makefiles.

21 years agoRemove debugging printfs.
phk [Mon, 11 Nov 2002 10:08:46 +0000 (10:08 +0000)]
Remove debugging printfs.

21 years agoHandle the removal of CD9660 and NFSCLIENT options in drivers.conf.
ru [Mon, 11 Nov 2002 10:07:57 +0000 (10:07 +0000)]
Handle the removal of CD9660 and NFSCLIENT options in drivers.conf.
Removed duplicate entry from drivers.conf.

21 years agoDon't override CWARNFLAGS in this Makefile.
mux [Mon, 11 Nov 2002 10:07:18 +0000 (10:07 +0000)]
Don't override CWARNFLAGS in this Makefile.

21 years agoRemoved more redundant lines: amr, mlx, pcn, ahc and mpt are handled
ru [Mon, 11 Nov 2002 09:52:59 +0000 (09:52 +0000)]
Removed more redundant lines: amr, mlx, pcn, ahc and mpt are handled
by drivers.conf already.

21 years agoia64 ABI breaker:
marcel [Mon, 11 Nov 2002 08:11:44 +0000 (08:11 +0000)]
ia64 ABI breaker:
Don't force 16-byte alignment at run-time. Do it at compile-time.
This saves us the pointer fiddling by the setjmp functions and
reduces complexity. While here, increase the jmp_buf by 16 bytes
to an even 512 bytes. Coincidentally, due to the way alignment
was handled prior to this change, the jmp_buf has not changed in
size, but only in how the space is used. Prior to this change
the 16 bytes were reserved for enforcing alignment; now they are
reserved by us for future extensions.
Therefore, this ABI breaker is relatively save: the failure is
always an alignment trap.

21 years agoDon't free the current addrinfo list, or else a pointer to a freed
yar [Mon, 11 Nov 2002 07:31:48 +0000 (07:31 +0000)]
Don't free the current addrinfo list, or else a pointer to a freed
memory area would arise.  Only an addrinfo list from an earlier
call to getaddrinfo() should be freed there because it will be
substituted by the current list referenced by "res".

Reported by: John Long <fbsd1@pruam.com>
MFC after: 5 days

21 years ago - Clear the page's PG_WRITEABLE flag in the i386's pmap_changebit()
alc [Mon, 11 Nov 2002 05:17:34 +0000 (05:17 +0000)]
 - Clear the page's PG_WRITEABLE flag in the i386's pmap_changebit()
   if we're removing write access from the page's PTEs.
 - Export pmap_remove_all() on alpha, i386, and ia64.  (It's already
   exported on sparc64.)

21 years agoMBR slices are named the same on ia64 as they are on i386.
marcel [Mon, 11 Nov 2002 04:46:39 +0000 (04:46 +0000)]
MBR slices are named the same on ia64 as they are on i386.

21 years agoComment-out USB support. A kernel doesn't boot with it. Deal with it
marcel [Mon, 11 Nov 2002 01:50:10 +0000 (01:50 +0000)]
Comment-out USB support. A kernel doesn't boot with it. Deal with it
later.

21 years agoDouble the size of the EFI boot image. 5MB is not enough. The kernel
marcel [Mon, 11 Nov 2002 00:48:35 +0000 (00:48 +0000)]
Double the size of the EFI boot image. 5MB is not enough. The kernel
is not compressed.

21 years agoGiven that we have 3 places to document UUID related information,
marcel [Mon, 11 Nov 2002 00:29:01 +0000 (00:29 +0000)]
Given that we have 3 places to document UUID related information,
namely uuidgen(1), uuidgen(2) and uuid(3), the following division
has been choosen:
uuidgen(1) A description of the command line utility,
and other user oriented UUID information.
uuidgen(2) A mostly technical description of UUIDs.
uuid(3) A description of the functions and other
programmer oriented UUID information.

According to the division: add more technical contents.

Contributed by: Hiten Pandya <hiten@uk.FreeBSD.org>
Edited and enhanced: marcel

21 years agoAdd gerrold.limerick to the building process.
dcs [Sun, 10 Nov 2002 23:41:01 +0000 (23:41 +0000)]
Add gerrold.limerick to the building process.

21 years agoFinished at last.
dcs [Sun, 10 Nov 2002 23:37:25 +0000 (23:37 +0000)]
Finished at last.

21 years agoWrite the boot block to the first 16 sectors of all partitions, instead of
jake [Sun, 10 Nov 2002 21:07:29 +0000 (21:07 +0000)]
Write the boot block to the first 16 sectors of all partitions, instead of
always to the first 16 sectors of the disk.  The firmware reads the boot
code from a partition, defaulting to 'a' if none is specified, which only
corresponds to the first 16 sectors of the disk if 'a' is first.  Solaris
often makes the swap partition first, instead of the root partition, and
users expect to be able to do the same with freebsd as well.  This also
allows one to temporarily boot from another partition if the boot block
on the root partition gets scrambled somehow.

21 years agoA support for creating EFI partitions. This is required on ia64,
marcel [Sun, 10 Nov 2002 20:56:53 +0000 (20:56 +0000)]
A support for creating EFI partitions. This is required on ia64,
but not made conditional upon it.

21 years agoAdd efi to the list of types for which we need to return tha name.
marcel [Sun, 10 Nov 2002 20:53:10 +0000 (20:53 +0000)]
Add efi to the list of types for which we need to return tha name.
Also, return chunk type efi in case we find an EFI partition in
the GPT. We used to return FAT due to a lack of EFI type.

21 years agoAdd an efi chunk type. We need to be able to create an EFI partition
marcel [Sun, 10 Nov 2002 20:49:28 +0000 (20:49 +0000)]
Add an efi chunk type. We need to be able to create an EFI partition
on ia64, because that's where we need to put the loader and the
kernel.

21 years agoGeneral cleanup:
marcel [Sun, 10 Nov 2002 20:47:02 +0000 (20:47 +0000)]
General cleanup:
o  Remove all code guarded by !defined(__ia64__). This file is
   specifically written for ia64,
o  Handle the case when read_block() or write_block() fails. We
   don't want sysinstall(8) to signal a thumbs-up on error,
o  Set the starting (cyl,hd,sect) triple to 0xFFFFFF when either
   bios_hd or bios_sect is zero or the LBA us not representable
   with the triple. In that case automaticly initialize the
   ending triple with 0xFFFFFF as well,
o  Reindent Write_Int32() as it was different than the rest of
   the file,
o  Remove some unused variables that appeared to be used but
   were effectively useless.
o  Plug a memory leak: The second timne we read the MBR, we write
   out a modified block, but didn't free the memory after writing.
o  Replace d1->sector_size with 512 when we read/write the MBR.
   We ignore the sector size in cases we shouldn't but adhered to
   it in cases it would be wrong if the sector_size wasn't 512.

This file should eventually be rewritten to write out a GPT. For
now, a MBR will do...

21 years agoRemove inclusion of <sys/uuid.h>. We now include <sys/uuid.h> in
marcel [Sun, 10 Nov 2002 20:22:25 +0000 (20:22 +0000)]
Remove inclusion of <sys/uuid.h>. We now include <sys/uuid.h> in
<sys/gpt.h>. This avoids having to include both <sys/uuid.h> and
<uuid.h>, which is considered by your friendly committer to be
aestheticly displeasing (= ballyhoo barf barf :-)