d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
96,215 Commits 72 Branches 135 Tags 3.2 GiB
50d7c061a3
Commit Graph

43383 Commits

Author SHA1 Message Date
Sam Leffler
50d7c061a3 add some missing locking 
Supported by:	FreeBSD Foundation
 2003-11-08 22:53:41 +00:00
Sam Leffler
1d78192b35 the sbappendaddr call in socket_send must be protected by Giant 
because it can happen from an MPSAFE callout

Supported by:	FreeBSD Foundation
 2003-11-08 22:51:18 +00:00
Sam Leffler
e3f268fc89 add locking assertions that turn into noops if INET6 is configured; 
this is necessary because the ipv6 code shares the in_pcb code with
ipv4 but (presently) lacks proper locking

Supported by:	FreeBSD Foundation
 2003-11-08 22:48:27 +00:00
Sam Leffler
7902224c6b o add a flags parameter to netisr_register that is used to specify 
whether or not the isr needs to hold Giant when running; Giant-less
  operation is also controlled by the setting of debug_mpsafenet
o mark all netisr's except NETISR_IP as needing Giant
o add a GIANT_REQUIRED assertion to the top of netisr's that need Giant
o pickup Giant (when debug_mpsafenet is 1) inside ip_input before
  calling up with a packet
o change netisr handling so swi_net runs w/o Giant; instead we grab
  Giant before invoking handlers based on whether the handler needs Giant
o change netisr handling so that netisr's that are marked MPSAFE may
  have multiple instances active at a time
o add netisr statistics for packets dropped because the isr is inactive

Supported by:	FreeBSD Foundation
 2003-11-08 22:28:40 +00:00
Hidetoshi Shimokawa
ddce43e1cf Move post dmamap_load processes into the callback function. 2003-11-08 16:26:22 +00:00
Josef Karthauser
68ded8480c I've had a couple of reports that the Sony Clie_40 doesn't need the 
PALM_4 initialisation hack.  I've not confirmed it myself, but
seeing as we already don't use it for the Sony Clie_41, let's drop
it from the Clie_40 also and see what happens.

(Question: What about the Clie_S360 and Clie_NX60 devices?  Do we
 need to drop Palm4 from those as well?  Possibly, but I've not had
 any reports about those so I don't know.)

PR:		kern/56575
MFC after:	3 days
 2003-11-08 11:23:07 +00:00
Scott Long
8226fdbdce Don't be so chatty when performing manual sense. This should make ATAPICAM 
devices a lot more quiet.
 2003-11-08 10:56:57 +00:00
Søren Schmidt
aeb1d7b554 Remove the NOGIANT flag again, it was premature. 2003-11-08 09:56:34 +00:00
Peter Wemm
4cd2d525e3 Move a MD 32 bit binary support routine into the MD areas. exec_setregs 
is highly MD in an emulation environment since it operates on the host
environment.  Although the setregs functions are really for exec support
rather than signals, they deal with the same sorts of context and include
files.  So I put it there rather than create yet another file.
 2003-11-08 07:43:44 +00:00
Peter Wemm
b0211d5445 Regen 2003-11-08 07:31:49 +00:00
Peter Wemm
bc0e45efa3 "implement" vfork(). Add comments next to the other syscalls that need 
to be implemented.  This is enough to run i386 /bin/tcsh.  /bin/sh is still
not happy because of some strange job control problem.
 2003-11-08 07:31:30 +00:00
Peter Wemm
fcfe57d640 Update the graffiti. 2003-11-08 04:39:22 +00:00
Peter Wemm
398dbb11d8 Switch from having a fpu "device" to something that is more like the 
integrated part of the cpu core that it is.
 2003-11-08 04:37:54 +00:00
Sam Leffler
6536168e0a must use RTFREE instead of rtfree for proper locking 2003-11-08 03:36:32 +00:00
Peter Wemm
e04f2bba96 Remove some duplicated comments that refer to npx. XXX The setregs 
function is actually MD (not MI) though..
 2003-11-08 03:35:06 +00:00
Peter Wemm
bf2f09ee97 The great s/npx/fpu/gi 2003-11-08 03:33:38 +00:00
Warner Losh
d0ff2e64ac -Wunused 2003-11-08 03:28:43 +00:00
Peter Wemm
a4d60a7fb3 Converge with i386/GENERIC 2003-11-08 03:17:36 +00:00
David Xu
685a6c448a Return a reasonable number for top or ps to display for M:N thread, 
since there is no direct association between M:N thread and kse,
sometimes, a thread does not have a kse, in that case, return a pctcpu
from its last kse, it is not perfect, but gives a good number to be
displayed.
 2003-11-08 03:03:17 +00:00
Alan Cox
5e4a2fc9aa - Similar to post-PAE RELENG_4 split pmap_pte_quick() into two cases, 
pmap_pte() and pmap_pte_quick().  The distinction being based upon the
   locks that are held by the caller.  When the given pmap is not the
   current pmap, pmap_pte() should be used when Giant is held and
   pmap_pte_quick() should be used when the vm page queues lock is held.
 - When assigning to PMAP1 or PMAP2, include PG_A anf PG_M.
 - Reenable the inlining of pmap_is_current().

In collaboration with:	tegge
 2003-11-08 03:01:26 +00:00
Peter Wemm
0432a0f961 Rename npx.c to fpu.c (it isn't an extension, its part of the core 
architecture now).
 2003-11-08 02:40:40 +00:00
Peter Wemm
8b2454d833 Rename npx* to fpu*. I haven't done the flags/function names yet. 2003-11-08 02:39:46 +00:00
Peter Wemm
42ad9517cc Point the description of the fpu data in the context structures to 
i386/include/npx.h instead of the host's machine/npx.h (which might not
exist)
 2003-11-08 02:36:05 +00:00
Sam Leffler
27a940c9a2 unbreak compilation of FAST_IPSEC 
Supported by:	FreeBSD Foundation
 2003-11-08 00:34:34 +00:00
Robert Watson
8a4b86b9a4 Replace a '-' with a ')'. Update copyright. 
PR:	53195
 2003-11-08 00:21:20 +00:00
Peter Wemm
7538a488f5 There isn't much point printing 'npx0: INT 16 interface' because that is 
the only way it works here.
 2003-11-08 00:13:43 +00:00
John Baldwin
88861af1fb Dump the trigger and polarity of each intpin's default setting in the 
bootverbose output.
 2003-11-07 23:44:35 +00:00
Sam Leffler
aab621f060 MFp4: reminder that random id code is not reentrant 
Supported by:	FreeBSD Foundation
 2003-11-07 23:31:29 +00:00
Scott Long
eb3b7bf69f Document the lockfunc and lockfuncarg arguments to bus_dma_tag_create() in 
the busdma headers.
 2003-11-07 23:29:42 +00:00
Sam Leffler
8f1ee3683d Move uid/gid checking logic out of line and lock inpcb usage. This 
has a LOR between IPFW inpcb locks but I'm committing it now as the
lesser of two evils (the other being unlocked use of in_pcblookup).

Supported by:	FreeBSD Foundation
 2003-11-07 23:26:57 +00:00
Warner Losh
123cd74e26 Remove channeling interrupts to IRQ1. Some chipsets don't do the 
expected thing and that causes interference with keyboards.
 2003-11-07 23:12:59 +00:00
Sam Leffler
a96756932a Assert GIANT_REQUIRED where sockets are manipulated. This is 
preparatory for MPSAFE network commits and ongoing socket
locking work.

Supported by:	FreeBSD Foundation
 2003-11-07 22:57:09 +00:00
Robert Watson
bea2b56b26 When allocation of a socket peer label fails, scrub what was 
successfully initialized in the label as a socket peer label, not a
socket label.  For current policy modules, this didn't make a
difference, but if a policy module had label data in the peer label
that was to be GC'd in a different way than the normal socket label,
it might have been a problem.

Obtained from:	TrustedBSD Project
Sponsored by:	DARPA, Network Associates Laboratories
 2003-11-07 22:31:27 +00:00
Peter Wemm
2331fb5799 Ahh, the joys of badge engineering. Tell the sk driver that the 
Linksys EG1032 is yet another variation.  It looks just like the 3c940
except it only has a Marvell logo and no 3com logo.
 2003-11-07 22:04:26 +00:00
John Baldwin
fab73bc221 Regen. 2003-11-07 21:36:35 +00:00
John Baldwin
572e11ac18 Sync up MP safe flags with global syscalls.master for the first time. This 
includes read(), write(), close(), linux_setuid16(), linux_getuid16(),
linux_pause(), linux_nice(), linux_kill(), dup(), linux_pipe(),
linux_setgid16(), linux_getgid16(), linux_signal(), linux_geteuid16(),
linux_getegid16(), acct(), setpgid(), umask(), dup2(), getppid(),
getpgrp(), setsid(), linux_sigaction(), linux_sgetmask(), linux_ssetmask(),
linux_setreuid16(), linux_setregid16(), linux_sigsuspend(), getrusage(),
gettimeofday(), linux_getgroups16(), linux_setgroups16(), getpriority(),
setpriority(), linux_sigreturn(), linux_clone(), linux_sigprocmask(),
linux_getsid(), mlock(), munlock(), mlockall(), munlockall(),
sched_setparam(), sched_getparam(), linux_sched_setscheduler(),
linux_sched_getscheduler(), linux_sched_get_priority_max(),
linux_sched_get_priority_min(), sched_rr_get_interval(),
linux_setresuid16(), linux_getresuid16(), linux_setresgid16(),
linux_getresgid16(), linux_rt_sigaction(), linux_rt_sigprocmask(),
linux_rt_sigsuspend(), geteuid(), getegid(), setreuid(), setregid(),
linux_getgroups(), linux_setgroups(), setresuid(), getresuid(),
setresgid(), getresgid(), setuid(), and setgid().
 2003-11-07 21:36:14 +00:00
Peter Wemm
71d6084373 Dont write to the stackgap directly in execve(). 2003-11-07 21:27:13 +00:00
Peter Wemm
5ed80e429e Increase the size of SPARE_USRSPACE. It is way too small by default 
for things like execve.
 2003-11-07 21:25:54 +00:00
John Baldwin
13b6c41b6c Regen. 2003-11-07 21:13:08 +00:00
John Baldwin
11d4f58fdd Sync MP safe flags with global syscalls.master for the first time. This 
includes read(), write(), close(), setuid(), getuid(), linux_ptrace(),
linux_kill(), setpgid(), dup(), pipe(), getgid(), osf1_sigprocmask(),
umask(), getpgrp(), linux_setgroups(), linux_getgroups(), dup2(),
setpriority(), osf1_sigreturn(), osf1_sigsuspend(), osf1_gettimeofday(),
setreuid(), setregid(), setgid(), setsid(), osf1_sigaction(), getpgid(),
linux_getsid(), osf1_sysinfo(), linux_clone(), mlock(), munlock(),
mlockall(), munlockall(), sched_setparam(), sched_getparam(),
linux_sched_setscheduler(), linux_sched_getscheduler(),
linux_sched_get_priority_max(), linux_sched_get_priority_min(),
setresuid(), getresuid(), linux_rt_sigaction(), linux_rt_sigprocmask(),
linux_rt_sigsuspend(), gettimeofday(), linux_getitimer(),
linux_setitimer(), getrusage(), setresgid(), getresgid().
 2003-11-07 21:09:19 +00:00
John Baldwin
a2f10724b5 Regen. 2003-11-07 20:56:54 +00:00
John Baldwin
bc7e0ea5ec Sync up with global syscalls.master for MP safe syscalls for the first 
time.  This includes read(), write(), close(), getpid(), osf1_setuid(),
getuid(), osf1_kill(), setpgid(), dup(), pipe(), getgid(),
osf1_sigprocmask(), getlogin(), setlogin(), osf1_sigpending(), umask(),
getpgrp(), getgroups(), setgroups(), osf1_setpgrp(), getdtablesize(),
dup2(), setpriority(), getpriority(), osf1_sigreturn(), osf1_sigsuspend(),
osf1_osigstack(), setreuid(), setregid(), osf1_setgid(), setsid(),
osf1_sigaction(), msgctl(), msgget(), msgrcv(), msgsnd(), __semctl(),
semget(), semop(), shmat(), shmctl(), shmdt(), shmget(), osf1_signal(),
getpgid(), getsid(), osf1_sigaltstack(), osf1_sysinfo().
 2003-11-07 20:56:31 +00:00
Hajimu UMEMOTO
ba3484d943 nuke obsoleted ipsec_gethist(). it just did panic to notify user 
that it was obsoleted.  it is better to fail than just hiding use
of ipsec_gethist() at build.

Sugessted by:	"Bjoern A. Zeeb" <bzeeb-lists@lists.zabbadoz.net>
 2003-11-07 20:38:45 +00:00
John Baldwin
dac33f12cc Regen. 2003-11-07 20:30:30 +00:00
John Baldwin
c2545316b0 Sync with global syscalls.master by marking ptrace(), dup(), pipe(), 
ktrace(), freebsd32_sigaltstack(), sysarch(), issetugid(), utrace(), and
freebsd32_sigaction() as MP safe.
 2003-11-07 20:29:53 +00:00
John Baldwin
a060e9b7ef Sync with global syscalls.master. ptrace(), dup(), pipe(), ktrace(), 
ia32_sigaltstack(), sysarch(), issetugid(), utrace(), and ia32_sigaction()
are MP safe.
 2003-11-07 20:27:16 +00:00
Hajimu UMEMOTO
aef3a65eb7 use ipsec_getnhist() instead of obsoleted ipsec_gethist(). 
Submitted by:	"Bjoern A. Zeeb" <bzeeb-lists@lists.zabbadoz.net>
Reviewed by:	Ari Suutari <ari@suutari.iki.fi> (ipfw@)
 2003-11-07 20:25:47 +00:00
John Baldwin
c055e5d412 Mark ptrace(), ktrace(), utrace(), sysarch(), and issetugid() as MP safe. 
The parts of these calls that are not yet MP safe acquire Giant explicitly.
 2003-11-07 20:23:23 +00:00
Bruce Evans
dd7b5bb30b Fixed insertion sort errors in the opt_cpu.h section. 
Removed banal comments about ELAN*.  Complain about ELAN* being misnamed
instead (so that these options are not obviously related to a CPU and
don't sort with CPU_ELAN).

Complain about CPU_DISABLE_CMPXCHG being in the wrong namespace.
 2003-11-07 15:14:53 +00:00
Hidetoshi Shimokawa
132e28bd75 Allow shared open of /dev/fwmem* for read-only. 2003-11-07 12:39:39 +00:00