Commit Graph

4648 Commits

Author SHA1 Message Date
thorpej
895f21a679 Prototype freebsd_syscall_intern() in freebsd_machdep.h so that
freebsd_exec.c has access to it.
2000-12-18 23:29:46 +00:00
jdolecek
02f7f5ef0e make compilable within LKM 2000-12-18 20:49:17 +00:00
jdolecek
6701524af4 put back the rval[1] microoptimization, this is no longer used by FreeBSD emulation 2000-12-18 20:40:25 +00:00
jdolecek
2db538c869 separate the freebsd syscall code similarily to other emulations 2000-12-18 20:38:27 +00:00
jdolecek
214db5ca6b Store edx to rval[1], so that it won't get zeroed spuriously - this made
shared FreeBSD binaries coredump.
This solves port-i386/11708 by Atsushi Onoe.

XXX it may be worth it to split FreeBSD syscall code off syscall.c similar way
XXX as other emulations
2000-12-18 11:37:08 +00:00
jmc
3926121843 Comment out the isa wss driver.
GENERIC is the kernel used on new installs and too many things overlap the
resources (irq mostly) the default isa version may be using which causes
panics on reboot.
2000-12-17 22:53:52 +00:00
thorpej
0a8ec4baf3 Add an at isapnp. 2000-12-17 20:35:45 +00:00
jdolecek
cacec11253 delete obsolete comment 2000-12-17 15:52:39 +00:00
jmc
d1c6f1123f Add pnpbios_getiosize 2000-12-17 07:59:14 +00:00
jmc
50445038df Change the attach routine so it handles all the known cases that pnpbios
returns for floppy controllers i/o ports. i.e.

1. 1 range of length 4 (which conveniently "forgets" the ctl i/o port),
2. 2 ranges: 1 of length 4 and the ctl i/o port.
3. 1 range of length 6 which goes to the end of the ctl i/o.

Make this line up with the MI fdc code by mapping in a range of 4 on the
base i/o and then either direct or submapping the ctl i/o. In the one case
where the BIOS lies and says it's not there just map it in anyways. (but note
the fact to the end user).
2000-12-17 07:58:37 +00:00
thorpej
cd414d81ab Glue in ALTQ. 2000-12-14 23:55:31 +00:00
mycroft
e06ad67fa5 Remove ALTENTRY(). It's dangerous and should not be used. 2000-12-14 20:15:26 +00:00
mycroft
33d5c07c1b Clean up #includes. 2000-12-14 18:34:59 +00:00
onoe
a7125e435c Add PCI version of Aironet, not tested at all.
ISA/ISApnp version won't come since I don't know about isa...
2000-12-14 04:11:25 +00:00
mycroft
716c484ada Remove unnecessary #includes. 2000-12-13 01:24:46 +00:00
he
4ce4648c24 Add the an(4) driver to these configurations as well. 2000-12-12 21:14:51 +00:00
mycroft
a7889f8229 Whoops. Previous change was not correct. 2000-12-12 20:30:12 +00:00
mycroft
2c7eaf2b50 Minor code rearrangement. 2000-12-12 20:22:49 +00:00
mycroft
0175db6e40 #ifdef KTRACE some stuff. 2000-12-12 03:33:16 +00:00
onoe
9c25e9208a Port 'an' driver for Aironet PC4500/PC4800 IEEE802.11 card from FreeBSD.
This is very adhoc work for IETF meeting.
- Since it seems that 'an' and 'wi' have similar hardware, low level
  functions should be shared.
- There are PCI/ISA cards of Aironet but not supported yet.
- The wiconfig interface is changed so that wiconfig cannot be used.
- 'ancontrol' of FreeBSD is not ported.
- Only infrastructure mode is tested.
- WEP is not supported.

Though I only have an Aironet card, Cisco card should be expected to work.
2000-12-11 23:16:50 +00:00
mycroft
9a8fbe25bc Handle the VM86-mode syscall problem by changing the syscall vector when
jumping in or out of VM86 mode.
2000-12-11 17:36:03 +00:00
mycroft
1ef2795be9 Separate the syscall handlers into two versions -- one which does ktrace, and
one which doesn't.  Dynamically switch.
There's more to do on this, but I have to go to work.
2000-12-11 16:49:15 +00:00
tron
d0de1ec852 Protect against multiple inclusion. 2000-12-11 11:39:06 +00:00
tron
c6cdb692f3 Include "machine/frame.h" to avoid lint errors while building "libc". 2000-12-11 10:22:55 +00:00
mycroft
e6d33b69ec Nuke reference to trace.h. 2000-12-11 05:41:06 +00:00
mycroft
0f17615da7 There is no Berkeley code in any of these files. 2000-12-11 05:37:01 +00:00
mycroft
f495cd9dc6 Introduce 2 new flags in types.h:
* __HAVE_SYSCALL_INTERN.  If this is defined, e_syscall is replaced by
  e_syscall_intern, which is called at key places in the kernel.  This can be
  used to set a MD syscall handler pointer.  This obsoletes and replaces the
  *_HAS_SEPARATED_SYSCALL flags.
* __HAVE_MINIMAL_EMUL.  If this is defined, certain (deprecated) elements in
  struct emul are omitted.
2000-12-11 05:28:59 +00:00
mycroft
6a4dfe5f24 Slight tweak. 2000-12-11 02:48:07 +00:00
mycroft
66610a4779 Introduce PROC_PC(), which is used to get a process's user PC. If this is
defined, call addupc_intr() directly from statclock() in the system time case,
using the same P_OWEUPC path if the copyin/copyout fails.
Use this in i386 to remove profiling code from the normal userret() path.
2000-12-10 19:29:30 +00:00
jdolecek
8ca533ee5b since this is now standalone and not called via syscall(), we need to
update statistics and set p->p_md.md_regs accordingly here
do the errno translation uncoditionally
2000-12-10 12:23:50 +00:00
jdolecek
603884ff22 syscall(): no emulation using this needs errno mapping, G/C the code 2000-12-10 12:09:59 +00:00
jdolecek
44d7fef20a make locore.s call emulation-specific syscall code directly
this should improve speed of emulation syscall path (avoids one function call,
and emulations syscall can use the trapframe trick)
Idea: Charles Hannum
2000-12-09 13:20:04 +00:00
jdolecek
6f0fceeb1d the call to emulation-specific syscall function will be handled differently,
don't do it explicitly
2000-12-09 12:57:17 +00:00
jdolecek
87a6bee3a9 prototype ibcs2_syscall() in compat/ibcs2/ibcs2_exec.c 2000-12-09 12:45:44 +00:00
jdolecek
8f9b01eba1 prototype svr4_syscall() (only) in compat/svr4/svr4_exec.c 2000-12-09 12:44:26 +00:00
jdolecek
45144cbd17 put userret() into header file machine/userret.h and pull it where needed 2000-12-09 11:21:52 +00:00
mycroft
b2052e5951 GC garbage in userret() and *_syscall(). 2000-12-09 06:34:07 +00:00
mycroft
afc49e36eb Set md_regs before call the alternate syscall handler.
Remove the EMUL_HAS_SYS___syscall test, because the handler is no longer shared
with SVR4.
Rather than comparing with e_nsysent, just mask the value.  This is only done
to protect us from malicious programs anyway.
2000-12-09 06:33:15 +00:00
mycroft
3340782619 Add a separate svr4_syscall(). 2000-12-09 06:30:54 +00:00
mycroft
7a70c438c8 Couple more tweaks to eliminate unnecessary cycles. 2000-12-09 02:46:17 +00:00
mycroft
57fdd67be5 Do the resched check *only* on an AST.
Also, there is no reason to save the original PC; there are no restartable
syscalls that change the PC -- and even if there were, they wouldn't do it
in the restart case.
2000-12-09 02:18:16 +00:00
mycroft
5415a63e13 Rework the way the nested task and trace flags are handled when entering the
old (lcall) syscall gate.  This speeds up that path a little.
2000-12-08 23:14:04 +00:00
mycroft
985a81f23b GC unneeded code. 2000-12-08 22:32:09 +00:00
thorpej
55ecab37a0 Fix a silly MP race condition introduced in the previous. 2000-12-07 21:53:46 +00:00
thorpej
ac083d0b74 Use a pool cache for PDPs. This results in a 20% improvement in a
loop of 10000 fork/exit operations on a 400MHz Celeron.
2000-12-07 20:19:05 +00:00
thorpej
0abaa8b70d Nothing calls pmap_release() except for pmap_destroy(). Merge the two. 2000-12-07 17:12:21 +00:00
thorpej
b4b01fe8e6 Nothing calls pmap_pinit() except for pmap_create(), so merge the two. 2000-12-07 17:09:26 +00:00
chs
4fa4734937 go back to the first version of the pte/pde dance now that kvtopte()
has been changed to consisently return a pointer into the recursive
page table mapping.
2000-12-06 03:20:35 +00:00
chs
2a95fc3517 in kvtopte(), return the mapping of the pde in the recursive page-table
mapping space rather than the mapping in the kernel address space.
2000-12-06 03:13:47 +00:00
jdolecek
94e024dc14 constify 2000-12-05 10:45:09 +00:00