]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
20 years agoo Reduce path names in RRQ and WRQ packets by:
Brian Somers [Mon, 21 Jun 2004 08:01:16 +0000 (08:01 +0000)]
o Reduce path names in RRQ and WRQ packets by:

    Reducing "/+./" strings to "/"
    Reducing "/[^/]+/../" to "/"

o Don't send an OACK when the result of the [RW]RQ is an error.

These changes allow tftpd to interact with pxelinux.bin from the syslinux
package.

Whilst the path reducing code doesn't properly handle situations where the
path component before the "/../" is a symlink to (say) ".", I would suggest
that it does the right thing in terms of the clients perception of what
their path string actually represents.  This seems better than using
realpath() and breaking environments where symlinks point outside of the
directory hierarchy that tftpd is configured to allow.

20 years agoInclude module.h
Scott Long [Mon, 21 Jun 2004 07:27:34 +0000 (07:27 +0000)]
Include module.h

Submitted by: Koop Mast

20 years agoExtend the description of SIOC[GS]IFCAP.
Yaroslav Tykhiy [Mon, 21 Jun 2004 05:13:22 +0000 (05:13 +0000)]
Extend the description of SIOC[GS]IFCAP.

20 years agoMerge next step in socket buffer locking:
Robert Watson [Mon, 21 Jun 2004 00:20:43 +0000 (00:20 +0000)]
Merge next step in socket buffer locking:

- sowakeup() now asserts the socket buffer lock on entry.  Move
  the call to KNOTE higher in sowakeup() so that it is made with
  the socket buffer lock held for consistency with other calls.
  Release the socket buffer lock prior to calling into pgsigio(),
  so_upcall(), or aio_swake().  Locking for this event management
  will need revisiting in the future, but this model avoids lock
  order reversals when upcalls into other subsystems result in
  socket/socket buffer operations.  Assert that the socket buffer
  lock is not held at the end of the function.

- Wrapper macros for sowakeup(), sorwakeup() and sowwakeup(), now
  have _locked versions which assert the socket buffer lock on
  entry.  If a wakeup is required by sb_notify(), invoke
  sowakeup(); otherwise, unconditionally release the socket buffer
  lock.  This results in the socket buffer lock being released
  whether a wakeup is required or not.

- Break out socantsendmore() into socantsendmore_locked() that
  asserts the socket buffer lock.  socantsendmore()
  unconditionally locks the socket buffer before calling
  socantsendmore_locked().  Note that both functions return with
  the socket buffer unlocked as socantsendmore_locked() calls
  sowwakeup_locked() which has the same properties.  Assert that
  the socket buffer is unlocked on return.

- Break out socantrcvmore() into socantrcvmore_locked() that
  asserts the socket buffer lock.  socantrcvmore() unconditionally
  locks the socket buffer before calling socantrcvmore_locked().
  Note that both functions return with the socket buffer unlocked
  as socantrcvmore_locked() calls sorwakeup_locked() which has
  similar properties.  Assert that the socket buffer is unlocked
  on return.

- Break out sbrelease() into a sbrelease_locked() that asserts the
  socket buffer lock.  sbrelease() unconditionally locks the
  socket buffer before calling sbrelease_locked().
  sbrelease_locked() now invokes sbflush_locked() instead of
  sbflush().

- Assert the socket buffer lock in socket buffer sanity check
  functions sblastrecordchk(), sblastmbufchk().

- Assert the socket buffer lock in SBLINKRECORD().

- Break out various sbappend() functions into sbappend_locked()
  (and variations on that name) that assert the socket buffer
  lock.  The !_locked() variations unconditionally lock the socket
  buffer before calling their _locked counterparts.  Internally,
  make sure to call _locked() support routines, etc, if already
  holding the socket buffer lock.

- Break out sbinsertoob() into sbinsertoob_locked() that asserts
  the socket buffer lock.  sbinsertoob() unconditionally locks the
  socket buffer before calling sbinsertoob_locked().

- Break out sbflush() into sbflush_locked() that asserts the
  socket buffer lock.  sbflush() unconditionally locks the socket
  buffer before calling sbflush_locked().  Update panic strings
  for new function names.

- Break out sbdrop() into sbdrop_locked() that asserts the socket
  buffer lock.  sbdrop() unconditionally locks the socket buffer
  before calling sbdrop_locked().

- Break out sbdroprecord() into sbdroprecord_locked() that asserts
  the socket buffer lock.  sbdroprecord() unconditionally locks
  the socket buffer before calling sbdroprecord_locked().

- sofree() now calls socantsendmore_locked() and re-acquires the
  socket buffer lock on return.  It also now calls
  sbrelease_locked().

- sorflush() now calls socantrcvmore_locked() and re-acquires the
  socket buffer lock on return.  Clean up/mess up other behavior
  in sorflush() relating to the temporary stack copy of the socket
  buffer used with dom_dispose by more properly initializing the
  temporary copy, and selectively bzeroing/copying more carefully
  to prevent WITNESS from getting confused by improperly
  initialized mutexes.  Annotate why that's necessary, or at
  least, needed.

- soisconnected() now calls sbdrop_locked() before unlocking the
  socket buffer to avoid locking overhead.

Some parts of this change were:

Submitted by: sam
Sponsored by: FreeBSD Foundation
Obtained from: BSD/OS

20 years agoAdd the `-O emul' format option, which prints the name of the system-call
Garance A Drosehn [Sun, 20 Jun 2004 23:40:54 +0000 (23:40 +0000)]
Add the `-O emul' format option, which prints the name of the system-call
emulation environment the process is in.  "emul" as a keyword is picked
up from OpenBSD.

PR: bin/65803
Submitted by: Cyrille Lefevre

20 years agoFix the following error:
Marcel Moolenaar [Sun, 20 Jun 2004 22:32:19 +0000 (22:32 +0000)]
Fix the following error:
ld: locore.o: non-pic code with imm relocation against dynamic
symbol `__gp'

With binutils 2.15, ld(1) defines the implicit/automatic symbol __gp
as a dynamic symbol and thus will now complain when used in a non-PIC
fashion (the immediate relocation used to set the GP register). Resolve
this by defining __gp in the linker script. Make sure __gp is aligned
on a 16-byte boundary.

Note: the 0x200000 magic offset is due to having a 22-bit GP-relative
relocation. The GOT will be accessed with negative offsets from GP.

20 years agoFrom SUSv3:
Garance A Drosehn [Sun, 20 Jun 2004 22:31:37 +0000 (22:31 +0000)]
From SUSv3:
    Any [standard output] field need not be meaningful in all
    implementations. In such a case a hyphen ('-') should be
    output in place of the field value

So have the `-O label' option print out the string "  -" if the
process has no label.

Approved by: Silence from rwatson and green (when asked in March...)

20 years agoAdd new output-format keywords of LWP and NLWP, which show the thread-id
Garance A Drosehn [Sun, 20 Jun 2004 22:22:49 +0000 (22:22 +0000)]
Add new output-format keywords of LWP and NLWP, which show the thread-id
and number-of-threads tied to a process.  Result can be seen by typing,
e.g.:   ps -HO lwp,nlwp
These new options are not documented yet.  More options will be coming,
and I will update the man page after I get farther along.

PR: bin/65803  (though adjusted to fit our present source)
Submitted by: Cyrille Lefevre

20 years agoFill in the values for the ki_tid and ki_numthreads which have been
Garance A Drosehn [Sun, 20 Jun 2004 22:17:22 +0000 (22:17 +0000)]
Fill in the values for the ki_tid and ki_numthreads which have been
added to kproc_info.

PR: bin/65803  (a tiny part...)
Submitted by: Cyrille Lefevre

20 years agoChange the architecture-based setting of KINFO_PROC_SIZE and KI_NSPARE so
Garance A Drosehn [Sun, 20 Jun 2004 22:05:41 +0000 (22:05 +0000)]
Change the architecture-based setting of KINFO_PROC_SIZE and KI_NSPARE so
that it is a series of alphabetically-ordered #fidef's, from Bruce Evans.
Define two new thread-related values in kproc_info, from Cyrille Lefevre.
Also remove a few values from kproc_info that were not needed, and change
around a few comments, from me.  Changes are combined into a single commit
simply because it is a hassle to make sure that alignments and sizes are
not changed on any platform when modifying kproc_info.

20 years agoInclude an annotation of NET_{LOCK,UNLOCK}_GIANT() noting that these
Robert Watson [Sun, 20 Jun 2004 21:49:12 +0000 (21:49 +0000)]
Include an annotation of NET_{LOCK,UNLOCK}_GIANT() noting that these
calls do not have the same recursion semantics as DROP_GIANT and
PICKUP_GIANT.

20 years agoIt's now the responsibility of the consumer of soabort() to remove a
Robert Watson [Sun, 20 Jun 2004 21:47:12 +0000 (21:47 +0000)]
It's now the responsibility of the consumer of soabort() to remove a
socket from its accept queue when aborting it during a new inbound
connection.  Update spx_input() to acquire the accept lock, assert
the condition of the socket on its parent queue, and approriately
disconnect it from the queue before calling soabort() on it.

20 years agoIf debug.mpsafenet is set, initialize TCP callouts as CALLOUT_MPSAFE.
Robert Watson [Sun, 20 Jun 2004 21:44:50 +0000 (21:44 +0000)]
If debug.mpsafenet is set, initialize TCP callouts as CALLOUT_MPSAFE.

20 years agoIn uipc_rcvd(), lock the socket buffers at either end of the UNIX
Robert Watson [Sun, 20 Jun 2004 21:43:13 +0000 (21:43 +0000)]
In uipc_rcvd(), lock the socket buffers at either end of the UNIX
domain sokcet when updating fields at both ends.

Submitted by: sam
Sponsored by: FreeBSD Foundation

20 years agoMerge changes from the tzcode2004a import.
Stefan Farfeleder [Sun, 20 Jun 2004 21:41:11 +0000 (21:41 +0000)]
Merge changes from the tzcode2004a import.

20 years agoAnnotate so_state as locked with SOCK_LOCK(so).
Robert Watson [Sun, 20 Jun 2004 21:39:46 +0000 (21:39 +0000)]
Annotate so_state as locked with SOCK_LOCK(so).

Add a commenting indicating that the SB_ constants apply to sb_flags.

20 years agoHold SOCK_LOCK(so) when frobbing so_state when disconnecting a
Robert Watson [Sun, 20 Jun 2004 21:29:56 +0000 (21:29 +0000)]
Hold SOCK_LOCK(so) when frobbing so_state when disconnecting a
connected UNIX domain datagram socket.

20 years agoHave the main() routine calculate %CPU and (if needed) memory information
Garance A Drosehn [Sun, 20 Jun 2004 21:25:10 +0000 (21:25 +0000)]
Have the main() routine calculate %CPU and (if needed) memory information
when copying per-process info before starting to sort the list.  This way,
sort-by-CPU or sort-by-memory will only calculate values once-per-process,
instead of twice-per-comparison.  Also take advantage of this to simplify
the pscomp() routine.

20 years agoRemove one more unneeded reference to arpcom.ac_netgraph.
Bill Paul [Sun, 20 Jun 2004 21:08:58 +0000 (21:08 +0000)]
Remove one more unneeded reference to arpcom.ac_netgraph.

Tweak things so that ng_fec has a chance of working with things
other than ethernet. Use ifp->if_output of the underlying interfaces
and use IF_HANDOFF() rather than depending on ether_output() and
ether_output_frame() explicitly. Also, don't insist that underlying
devices be IFM_ETHER when checking their link states in the link
monitor code.

With these changes, I was able to create a two channel bundle
consisting of one ethernet interface and one 802.11 wireless
device (via ndis). Note that this only works because both devices
use the same if_output vector: ng_fec will not let you bundle
devices with different output vectors together (it really doesn't
make sense to do that).

20 years ago - Simplify pmap_remove_pages(), eliminating unnecessary indirection.
Alan Cox [Sun, 20 Jun 2004 20:57:06 +0000 (20:57 +0000)]
 - Simplify pmap_remove_pages(), eliminating unnecessary indirection.
 - Simplify the locking of pmap_is_modified() by converting control flow to
   data flow.

20 years agoRemove from HEAD, that what is not on the vendor branch anymore.
Marcel Moolenaar [Sun, 20 Jun 2004 20:40:44 +0000 (20:40 +0000)]
Remove from HEAD, that what is not on the vendor branch anymore.

20 years agoRevive files which were pulled from the vendor branch only to be
Marcel Moolenaar [Sun, 20 Jun 2004 20:36:15 +0000 (20:36 +0000)]
Revive files which were pulled from the vendor branch only to be
removed on HEAD, but which are part of the GDB 6.1.1 import. This
makes HEAD identical to the vendor branch, for better or for
worse.

20 years agoAssert the inpcb lock before letting MAC check whether we can deliver
Robert Watson [Sun, 20 Jun 2004 20:17:29 +0000 (20:17 +0000)]
Assert the inpcb lock before letting MAC check whether we can deliver
to the inpcb in tcp_input().

20 years agoIP multicast code no longer needs to acquire Giant before appending
Robert Watson [Sun, 20 Jun 2004 20:10:05 +0000 (20:10 +0000)]
IP multicast code no longer needs to acquire Giant before appending
an mbuf onto a socket buffer.  This is left over from debug.mpsafenet
affecting the forwarding/bridging plane only.

20 years agoUpdate to GDB 6.1.1. These files are identical to the vendor branch.
Marcel Moolenaar [Sun, 20 Jun 2004 19:47:29 +0000 (19:47 +0000)]
Update to GDB 6.1.1. These files are identical to the vendor branch.

20 years agoStash our node context pointer somewhere else within struct ifnet of
Bill Paul [Sun, 20 Jun 2004 19:22:22 +0000 (19:22 +0000)]
Stash our node context pointer somewhere else within struct ifnet of
underlying interfaces rather than using ac_netgraph in struct arpcom.
The latter is meant only for use by ng_ether, and using it breaks
interoperability with the rest of netgraph.

20 years agoUpdate to match the GDB 6.1.1 import.
Marcel Moolenaar [Sun, 20 Jun 2004 19:15:58 +0000 (19:15 +0000)]
Update to match the GDB 6.1.1 import.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r130805,
Marcel Moolenaar [Sun, 20 Jun 2004 19:05:30 +0000 (19:05 +0000)]
This commit was generated by cvs2svn to compensate for changes in r130805,
which included commits to RCS files with non-trunk default branches.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r130803,
Marcel Moolenaar [Sun, 20 Jun 2004 18:26:14 +0000 (18:26 +0000)]
This commit was generated by cvs2svn to compensate for changes in r130803,
which included commits to RCS files with non-trunk default branches.

20 years agoImport of GDB version 6.1.1 (stripped; see FREEBSD-Xlist rev. 1.6).
Marcel Moolenaar [Sun, 20 Jun 2004 18:26:14 +0000 (18:26 +0000)]
Import of GDB version 6.1.1 (stripped; see FREEBSD-Xlist rev. 1.6).

20 years agoBump document date.
Dag-Erling Smørgrav [Sun, 20 Jun 2004 18:15:20 +0000 (18:15 +0000)]
Bump document date.

Requested by: ru

20 years agoWhen retrieving the SO_LINGER socket option for user space, hold the
Robert Watson [Sun, 20 Jun 2004 17:50:42 +0000 (17:50 +0000)]
When retrieving the SO_LINGER socket option for user space, hold the
socket lock over pulling so_options and so_linger out of the socket
structure in order to retrieve a consistent snapshot.  This may be
overkill if user space doesn't require a consistent snapshot.

20 years agoConvert an if->panic in soclose() into a call to KASSERT().
Robert Watson [Sun, 20 Jun 2004 17:47:51 +0000 (17:47 +0000)]
Convert an if->panic in soclose() into a call to KASSERT().

20 years agoGive zlib the ability to be a module that can be depended on,
Mark Murray [Sun, 20 Jun 2004 17:42:35 +0000 (17:42 +0000)]
Give zlib the ability to be a module that can be depended on,
in the MODULE_DEPEND() sense.

20 years agoAnnotate some ordering-related issues in solisten() which are not yet
Robert Watson [Sun, 20 Jun 2004 17:38:19 +0000 (17:38 +0000)]
Annotate some ordering-related issues in solisten() which are not yet
resolved by socket locking: in particular, that we test the connection
state at the socket layer without locking, request that the protocol
begin listening, and then set the listen state on the socket
non-atomically, resulting in a non-atomic cross-layer test-and-set.

20 years agoAnnotate two intentionally unlocked reads with comments.
Robert Watson [Sun, 20 Jun 2004 17:35:50 +0000 (17:35 +0000)]
Annotate two intentionally unlocked reads with comments.

Annotate a potentially inconsistent result returned to user space when
performing fstaT() on a socket due to not using socket buffer locking.

20 years agoInitialize ni_cnd.cn_cred before calling lookup() (this is normally done
Thomas Moestl [Sun, 20 Jun 2004 17:31:01 +0000 (17:31 +0000)]
Initialize ni_cnd.cn_cred before calling lookup() (this is normally done
by namei(), which cannot easily be used here however). This fixes boot
time crashes on sparc64 and probably other platforms.

Reviewed by: phk

20 years agoFix the format-string in a call to err(). It was causing a warning if
Garance A Drosehn [Sun, 20 Jun 2004 17:26:47 +0000 (17:26 +0000)]
Fix the format-string in a call to err().  It was causing a warning if
compiled on 4.x-stable.

20 years agoDocument the possibility of setting default values for
Josef El-Rayes [Sun, 20 Jun 2004 15:58:24 +0000 (15:58 +0000)]
Document the possibility of setting default values for
mixer channels via boot loader.

20 years agoEnable pcm to read kenv variables to set default values for
Josef El-Rayes [Sun, 20 Jun 2004 15:38:11 +0000 (15:38 +0000)]
Enable pcm to read kenv variables to set default values for
mixer channels.

e.g.: pcm0.line=0

to muten input line per default.

Approved by:    cg
Reviewed by:    le, stefanf
Requested by:   Eugene Grosbein <eugen@grosbein.pp.ru> (implicitly)
PR:             kern/63771

20 years agoMFNetBSD 1.15, author: mycroft
Lukas Ertl [Sun, 20 Jun 2004 15:32:19 +0000 (15:32 +0000)]
MFNetBSD 1.15, author: mycroft
   Fix an unimportant typo.

20 years agoRe-add the code to automatically load the smbfs.ko module if necessary.
Maxime Henrion [Sun, 20 Jun 2004 14:30:33 +0000 (14:30 +0000)]
Re-add the code to automatically load the smbfs.ko module if necessary.
We can't realy on the mount(2) system call to do it for us here because
smb_lib_init() needs the module and we call it before mount().  The old
code has been slightly modified to not use the getvfsent(3) API which
is now retired.

Noticed by:     many

20 years agoMicro-tweaking.
Mark Murray [Sun, 20 Jun 2004 13:50:50 +0000 (13:50 +0000)]
Micro-tweaking.

20 years agoSpelling fixes.
Mike Pritchard [Sun, 20 Jun 2004 13:33:35 +0000 (13:33 +0000)]
Spelling fixes.

20 years agoSpelling fixes.
Mike Pritchard [Sun, 20 Jun 2004 13:17:37 +0000 (13:17 +0000)]
Spelling fixes.

20 years agoSpelling fix and add $ FreeBSD $ tag
Mike Pritchard [Sun, 20 Jun 2004 13:07:25 +0000 (13:07 +0000)]
Spelling fix and add $ FreeBSD $ tag

20 years agoGrr. Fixup markup on my last commit. Sorry, I'm a little out of practice :)
Mike Pritchard [Sun, 20 Jun 2004 12:51:23 +0000 (12:51 +0000)]
Grr.  Fixup markup on my last commit.  Sorry, I'm a little out of practice :)

20 years agoSync up ERRORS section with errno.h.
Mike Pritchard [Sun, 20 Jun 2004 12:44:42 +0000 (12:44 +0000)]
Sync up ERRORS section with errno.h.

20 years agoFix a spelling error.
Mike Pritchard [Sun, 20 Jun 2004 12:35:16 +0000 (12:35 +0000)]
Fix a spelling error.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r130780,
Stefan Farfeleder [Sun, 20 Jun 2004 12:04:00 +0000 (12:04 +0000)]
This commit was generated by cvs2svn to compensate for changes in r130780,
which included commits to RCS files with non-trunk default branches.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r130777,
Stefan Farfeleder [Sun, 20 Jun 2004 11:52:51 +0000 (11:52 +0000)]
This commit was generated by cvs2svn to compensate for changes in r130777,
which included commits to RCS files with non-trunk default branches.

20 years agoImport the rest of tzcode2004a, the bits needed for the timezone compiler.
Stefan Farfeleder [Sun, 20 Jun 2004 11:52:51 +0000 (11:52 +0000)]
Import the rest of tzcode2004a, the bits needed for the timezone compiler.

Obtained from: ftp://elsie.nci.nih.gov/pub/tzcode2004a.tar.gz

20 years agoRetire support for gprof's -c option. All our currently supported
Stefan Farfeleder [Sun, 20 Jun 2004 11:05:25 +0000 (11:05 +0000)]
Retire support for gprof's -c option.  All our currently supported
architectures only provide a dummy implementation.

Silence on: current@

20 years agoCompletely remove s_ilogb.S as the assembler implementation gives very little
Stefan Farfeleder [Sun, 20 Jun 2004 10:42:23 +0000 (10:42 +0000)]
Completely remove s_ilogb.S as the assembler implementation gives very little
speed improvement to none at all over the MI version.

Submitted by: bde

20 years agoUncomment some functions that we now support.
David Schultz [Sun, 20 Jun 2004 10:39:09 +0000 (10:39 +0000)]
Uncomment some functions that we now support.

20 years agoRemove outdated comments.
Tim J. Robbins [Sun, 20 Jun 2004 10:01:30 +0000 (10:01 +0000)]
Remove outdated comments.

20 years agoWith tmm's latest change to the sparc64 dynamic loader, we can remove this
David E. O'Brien [Sun, 20 Jun 2004 09:46:13 +0000 (09:46 +0000)]
With tmm's latest change to the sparc64 dynamic loader, we can remove this
hack to binutils and use the stock sparc64 sources.

20 years agoCross-reference round(3) and trunc(3) as appropriate.
David Schultz [Sun, 20 Jun 2004 09:27:17 +0000 (09:27 +0000)]
Cross-reference round(3) and trunc(3) as appropriate.

20 years agoConnect scalbln(), trunc(), and the associated documentation to the build.
David Schultz [Sun, 20 Jun 2004 09:27:03 +0000 (09:27 +0000)]
Connect scalbln(), trunc(), and the associated documentation to the build.

20 years agoDeclare scalbln(), scalblnf(), trunc(), and truncf().
David Schultz [Sun, 20 Jun 2004 09:26:41 +0000 (09:26 +0000)]
Declare scalbln(), scalblnf(), trunc(), and truncf().

20 years agoImplement trunc() and truncf().
David Schultz [Sun, 20 Jun 2004 09:25:43 +0000 (09:25 +0000)]
Implement trunc() and truncf().

20 years agoAdd trivial implementations of scalbln() and scalblnf().
David Schultz [Sun, 20 Jun 2004 09:25:27 +0000 (09:25 +0000)]
Add trivial implementations of scalbln() and scalblnf().
These routines are specified in C99 for the sake of
architectures where an int isn't big enough to represent
the full range of floating-point exponents.  However,
even the 128-bit long double format has an exponent smaller
than 15 bits, so for all practical purposes, scalbln() and
scalblnf() are aliases for scalbn() and scalbnf(), respectively.

20 years agoAdd pmap locking to pmap_is_prefaultable().
Alan Cox [Sun, 20 Jun 2004 06:11:00 +0000 (06:11 +0000)]
Add pmap locking to pmap_is_prefaultable().

20 years agoBacked out previous commit. Blind substitution of dev_t by `struct cdev *'
Bruce Evans [Sun, 20 Jun 2004 03:52:50 +0000 (03:52 +0000)]
Backed out previous commit.  Blind substitution of dev_t by `struct cdev *'
was just wrong here because the dev_t's are user dev_t's.

20 years agoFixed misformatting of code and breaking of a comment in previous commit.
Bruce Evans [Sun, 20 Jun 2004 03:36:31 +0000 (03:36 +0000)]
Fixed misformatting of code and breaking of a comment in previous commit.

20 years agoFixed misformatting in previous commit.
Bruce Evans [Sun, 20 Jun 2004 03:34:21 +0000 (03:34 +0000)]
Fixed misformatting in previous commit.

20 years agoBacked out previous commit. The dev_t -> `struct cdev *' changes have
Bruce Evans [Sun, 20 Jun 2004 03:11:19 +0000 (03:11 +0000)]
Backed out previous commit.  The dev_t -> `struct cdev *' changes have
lots of errors.  Blind substitution of "dev_t foo" by "struct cdev *foo"
in comments usually just created an English syntax error (e.g.,
"struct cdev *changes"), but here it did less than that since the dev_t
is a user dev_t.

20 years agoThis file was not part of the GDB 5.2.1 import and should have been
Marcel Moolenaar [Sun, 20 Jun 2004 02:30:48 +0000 (02:30 +0000)]
This file was not part of the GDB 5.2.1 import and should have been
deleted from the vendor branch.

20 years agoAdd a call to calcru() to update the kproc_info fields of ki_rusage.ru_utime
Garance A Drosehn [Sun, 20 Jun 2004 02:03:33 +0000 (02:03 +0000)]
Add a call to calcru() to update the kproc_info fields of ki_rusage.ru_utime
and ki_rusage.ru_stime.  This greatly improves the accuracy of those fields.

Suggested by: bde

20 years agoThis file was not part of the GDB 5.2.1 import and should have been
Marcel Moolenaar [Sun, 20 Jun 2004 01:55:44 +0000 (01:55 +0000)]
This file was not part of the GDB 5.2.1 import and should have been
deleted from the vendor branch.

20 years agoOur 'TOOLS_PREFIX' and not 'DESTDIR' matches what the stock GNU build
David E. O'Brien [Sun, 20 Jun 2004 01:44:58 +0000 (01:44 +0000)]
Our 'TOOLS_PREFIX' and not 'DESTDIR' matches what the stock GNU build
framework wants.

20 years agoInclude <sys/_lock.h>'s prerequisite <sys/queue.h> before including the
Bruce Evans [Sun, 20 Jun 2004 00:33:14 +0000 (00:33 +0000)]
Include <sys/_lock.h>'s prerequisite <sys/queue.h> before including the
former, not after.

20 years agoUpdate to binutils 2.15.
Marcel Moolenaar [Sat, 19 Jun 2004 23:18:14 +0000 (23:18 +0000)]
Update to binutils 2.15.

20 years agoPrint the `99.99% done, finished soon' message only as often as the
Ian Dowse [Sat, 19 Jun 2004 22:41:18 +0000 (22:41 +0000)]
Print the `99.99% done, finished soon' message only as often as the
normal status messages. Previously a large number of these new
messages could be spewed out towards the end of a dump.

Reviewed by: imp

20 years agoClear any pending exceptions before using frstor (in the non-FXSR case)
Bruce Evans [Sat, 19 Jun 2004 22:24:16 +0000 (22:24 +0000)]
Clear any pending exceptions before using frstor (in the non-FXSR case)
in npxsetregs() too.  npxsetregs() must overwrite the previous state, and
it is never paired with an npxgetregs() that would defuse the previous
state (since npxgetregs() would have fninit'ed the state, leaving nothing
to do).

PR: 68058 (this should complete the fix)
Tested by: Simon Barner <barner@in.tum.de>

20 years agoSeperately add regex.c and splay-tree.c to SRCS. They are both needed
Marcel Moolenaar [Sat, 19 Jun 2004 22:22:42 +0000 (22:22 +0000)]
Seperately add regex.c and splay-tree.c to SRCS. They are both needed
by gdb(1). While here, sort SRCS.

20 years agoSync to 1.182 of usbdevs
MIHIRA Sanpei Yoshiro [Sat, 19 Jun 2004 22:17:34 +0000 (22:17 +0000)]
Sync to 1.182 of usbdevs

20 years agoAdd support Microtune Bluetooth dongle
MIHIRA Sanpei Yoshiro [Sat, 19 Jun 2004 22:16:03 +0000 (22:16 +0000)]
Add support Microtune Bluetooth dongle

PR: kern/68049
Submitted by: Markus Brueffer <markus@brueffer.de>

20 years agoYield to ru's ninja-like manpage skills and update watchdog(4) to
Sean Kelly [Sat, 19 Jun 2004 21:17:07 +0000 (21:17 +0000)]
Yield to ru's ninja-like manpage skills and update watchdog(4) to
look more like other manpages. Move some content around, use an EXAMPLES
section, etc.

Submitted by: ru

20 years agoAdd the necessary branding of ELF files. This is relatively easy to
Marcel Moolenaar [Sat, 19 Jun 2004 21:14:10 +0000 (21:14 +0000)]
Add the necessary branding of ELF files. This is relatively easy to
achieve on ia64, because we need to generate the ELF64/ia64 code and
simply tag elf-fbsd-brand.c at the end of it.

This hasn't actually been tested beyond trivial compilation testing.
A buildworld has been started and it's time I wait for my changes to
loop back to my local repo anyway. I'll get back to this in a couple
of hours...

20 years agoAdd bfdwin.c to SRCS. The file contains support for mmap(2) windows
Marcel Moolenaar [Sat, 19 Jun 2004 20:50:57 +0000 (20:50 +0000)]
Add bfdwin.c to SRCS. The file contains support for mmap(2) windows
into a BFD. It's included in stock BFD on the platforms I checked
(i386 and ia64).

20 years agoFix compilation for Xscale.
Olivier Houchard [Sat, 19 Jun 2004 20:50:50 +0000 (20:50 +0000)]
Fix compilation for Xscale.

20 years agoGrrr. our rev 1.19 (FSF GCC rev 1.579) is causing some problems on 32-bit
David E. O'Brien [Sat, 19 Jun 2004 20:40:00 +0000 (20:40 +0000)]
Grrr. our rev 1.19 (FSF GCC rev 1.579) is causing some problems on 32-bit
systems.  So only use the rev 1.19 (FSF GCC rev 1.579) change on 64-bit
systems.

20 years agoMove the elf32-target.h and elf64-target.h targets from the MD
Marcel Moolenaar [Sat, 19 Jun 2004 20:37:38 +0000 (20:37 +0000)]
Move the elf32-target.h and elf64-target.h targets from the MD
makefiles to the centralized makefile. This not only reduces
duplication, it also makes the MD quirks stand out better and
thus improves maintenance.

20 years agoRemove dead code related to pv entry allocation.
Alan Cox [Sat, 19 Jun 2004 20:31:49 +0000 (20:31 +0000)]
Remove dead code related to pv entry allocation.

Reviewed by: marcel@

20 years agoThe 3C3SH573BT looks like the Xircom RealPort cards to the naked eye,
Warner Losh [Sat, 19 Jun 2004 19:46:13 +0000 (19:46 +0000)]
The 3C3SH573BT looks like the Xircom RealPort cards to the naked eye,
but it appears to have a 3CxFE575BT under the hood (and is identified
as such by the xl driver).  Add it to the list of supported cards.  No
driver changes are required or necessary to support this card.

20 years agoUpdate to binutils 2.15:
Marcel Moolenaar [Sat, 19 Jun 2004 19:44:19 +0000 (19:44 +0000)]
Update to binutils 2.15:
o  BFD_VERSION_DATE now reflects the release date of 2.15,
o  BFD_VERSION now has the correct version number.

Previous values reflected 2.14.92 from a week prior to release.

While here, fix a whitespace (tab) nit.

20 years agoUpdate to binutils 2.15:
Marcel Moolenaar [Sat, 19 Jun 2004 19:25:33 +0000 (19:25 +0000)]
Update to binutils 2.15:
o  HAVE_SEEKO and HAVE_TELLO are defined now. These are used in bfdio.c.
o  HAVE_STRTOULL is defined now. This is used in bfd.c.

20 years agoRemove unused pt_entry_ts. Remove an unneeded semicolon.
Alan Cox [Sat, 19 Jun 2004 19:09:08 +0000 (19:09 +0000)]
Remove unused pt_entry_ts.  Remove an unneeded semicolon.

20 years agoFine-tune the last change even more and use the return value as error
Lukas Ertl [Sat, 19 Jun 2004 19:03:01 +0000 (19:03 +0000)]
Fine-tune the last change even more and use the return value as error
indicator, as it is expected.

Spotted by: Christoph Mallon <christoph.mallon@gmx.de>

20 years agoUpdate to binutils 2.15.
Marcel Moolenaar [Sat, 19 Jun 2004 18:52:31 +0000 (18:52 +0000)]
Update to binutils 2.15.

20 years agoDefine __lwpid_t as an int32_t in <sys/_types.h> and define lwpid_t
Marcel Moolenaar [Sat, 19 Jun 2004 17:58:32 +0000 (17:58 +0000)]
Define __lwpid_t as an int32_t in <sys/_types.h> and define lwpid_t
as an __lwpid_t in <sys/types.h>. Retype td_tid from an int to a
lwpid_t and change related definitions accordingly.

20 years agoInclude sys/module.h.
Olivier Houchard [Sat, 19 Jun 2004 17:38:32 +0000 (17:38 +0000)]
Include sys/module.h.

20 years agoI happened to have a sys/pool.h file in my tree, but most people do not,
Olivier Houchard [Sat, 19 Jun 2004 17:37:47 +0000 (17:37 +0000)]
I happened to have a sys/pool.h file in my tree, but most people do not,
so nuke this useless include.

20 years agoThis file was not part of the GDB 5.2.1 import and should have been
Marcel Moolenaar [Sat, 19 Jun 2004 15:43:52 +0000 (15:43 +0000)]
This file was not part of the GDB 5.2.1 import and should have been
deleted from the vendor branch.

20 years agoInclude <sys/_lock.h>'s prerequisite <sys/queue.h> before including the
Bruce Evans [Sat, 19 Jun 2004 14:58:35 +0000 (14:58 +0000)]
Include <sys/_lock.h>'s prerequisite <sys/queue.h> before including the
former, not after.

Don't hide this bug by including <sys/queue.h> in <sys/_lock.h>.

20 years agoWhen no fixed address is given in a shmat() request, pass a hint address
Tim J. Robbins [Sat, 19 Jun 2004 14:46:13 +0000 (14:46 +0000)]
When no fixed address is given in a shmat() request, pass a hint address
to vm_map_find() that is less likely to be outside of addressable memory
for 32-bit processes: just past the end of the largest possible heap.
This is the same hint that mmap() uses.

20 years agoThis is just a forced commit to note that the previous update was from:
Garance A Drosehn [Sat, 19 Jun 2004 14:09:40 +0000 (14:09 +0000)]
This is just a forced commit to note that the previous update was from:

PR: bin/65803  (a very tiny piece of the PR)

20 years agoFill in the some new fields 'struct kinfo_proc', namely ki_childstime,
Garance A Drosehn [Sat, 19 Jun 2004 14:08:10 +0000 (14:08 +0000)]
Fill in the some new fields 'struct kinfo_proc', namely ki_childstime,
ki_childutime, and ki_emul.  Also uses the timeradd() macro to correct
the calculation of ki_childtime.  That will correct the value returned
when ki_childtime.tv_usec > 1,000,000.

This also implements a new KERN_PROC_GID option for kvm_getprocs().
It also implements the KERN_PROC_RGID and KERN_PROC_SESSION options
which were added to sys/kern/kern_proc.c revision 1.203.

PR: bin/65803  (a very tiny piece of the PR)
Submitted by: Cyrille Lefevre

20 years agoFill in the some new fields 'struct kinfo_proc', namely ki_childstime,
Garance A Drosehn [Sat, 19 Jun 2004 14:03:00 +0000 (14:03 +0000)]
Fill in the some new fields 'struct kinfo_proc', namely ki_childstime,
ki_childutime, and ki_emul.  Also uses the timevaladd() routine to
correct the calculation of ki_childtime.  That will correct the value
returned when ki_childtime.tv_usec > 1,000,000.

This also implements a new KERN_PROC_GID option for kvm_getprocs().
(there will be a similar update to lib/libkvm/kvm_proc.c)

Submitted by: Cyrille Lefevre