Commit Graph

184 Commits

Author SHA1 Message Date
joerg 823959a6b2 Refactor interrupt dispatching. 2007-12-13 16:41:59 +00:00
ad 9e9c09c30a Fix _atomic_cas_64. Noted by bouyer@. 2007-12-10 14:13:12 +00:00
ad 164992f1a5 Add missing strong aliases. 2007-12-09 17:38:51 +00:00
ad 6260baac6e Add missing strong aliases; sure I did this before? 2007-12-09 17:33:56 +00:00
ad 1206a8cb5c Add some goo to make this work in userspace. 2007-12-08 22:53:33 +00:00
ad 96f5ce86b9 objdump -d to the rescue. I think this should work now. 2007-12-08 22:42:46 +00:00
joerg e982f87297 Expect that src/common/include is in include path and use x86emu/* for
the includes.
2007-12-04 17:32:22 +00:00
he 1e962f25f7 Correct what must obviously be a cut+paste-induced set of typos. 2007-12-04 01:13:38 +00:00
wiz 7b371a343f Typo fixes. 2007-12-02 18:57:56 +00:00
joerg 00487473b9 Move x86emu's include files to src/common/include/x86emu. 2007-12-01 20:14:09 +00:00
ad 234aa6237e Memory barriers for MIPS. 2007-11-30 22:50:44 +00:00
joerg 22ac6fc0fb Add x86emu for in-kernel use. This is a 386 real mode emulator that can
be used to run ROM code inside the kernel without having to setup vm86
and without having a x86 CPU at hand.

This code was originally written by SciTech and released under MIT/X11
license for XFree86. It has been refactored be less than half of the
original size, the compiled code being around a third of the former
size. The interface was changed to be a real library, e.g. no global
variables are used.
2007-11-30 20:02:49 +00:00
ad 4d460dc055 This is atomic_dec, not atomic_inc! 2007-11-30 17:11:48 +00:00
ad f377e7f907 Fix ia32 -> amd64 thinko. 2007-11-29 21:59:47 +00:00
ad 7243d328e1 Atomic ops for sh3. 2007-11-29 17:33:38 +00:00
ad 6f3a041727 atomic_add_* takes signed integers, the others take unsigned. 2007-11-29 17:29:40 +00:00
ad 29d1edb61b Apparently pointers are also returned in %a0. 2007-11-29 17:12:21 +00:00
ad 935eacf351 Pull in m68k atomic ops from the thorpej-atomic branch. 2007-11-29 17:09:31 +00:00
ad 6b0ff77038 Atomic ops for hppa. 2007-11-29 16:14:56 +00:00
ad 881b1b699e Use the CAS-based inc/dec variants, since these CPUs don't have atomic
add in hardware (does arm?).
2007-11-29 15:29:46 +00:00
ad fe803ab0ed Add inc/dec variants that use atomic_cas in preference to atomic_add. 2007-11-29 15:27:57 +00:00
ad 5c5043b187 membar_read/write are gone. 2007-11-29 15:20:43 +00:00
ad 11f83d35dd Atomic ops for arm. 2007-11-29 15:19:54 +00:00
ad 2da5f5559a Fix the aliases. 2007-11-29 02:16:00 +00:00
ad ebe806aa11 sparc atomic ops 2007-11-29 02:01:22 +00:00
ad c69c0e0f91 Make the 64-bit operations available when possible. 2007-11-29 01:04:57 +00:00
ad 695c2bc225 Only build the 64-bit versions if __HAVE_ATOMIC64_OPS. 2007-11-29 01:02:40 +00:00
ad 7b3b1e0427 Atomic ops for MIPS. Use the CAS functions already provided by the kernel,
and use the generic C code to provide the rest. Unfortunatley the C code
assembles up pretty badly on MIPS but at least it will work.
2007-11-29 00:16:23 +00:00
ad 2f827ecd6a ppc64 atomic ops. 2007-11-28 21:47:00 +00:00
ad bc7bf35c36 Atomic ops for powerpc. Please review. 2007-11-28 21:39:48 +00:00
ad 44fc0a18e5 A lock prefix on xchg is meaningless. 2007-11-28 18:02:29 +00:00
ad fb900ec918 Pull in alpha atomic ops from the thorpej-atomic branch. 2007-11-28 17:39:06 +00:00
ad f1d10eab90 Pull in sparc64 atomic ops from the thorpej-atomic branch. 2007-11-28 17:29:31 +00:00
ad 77ca26a240 Memory barriers for sparc64. XXX Can the 'membar' go in retl delay slot? 2007-11-28 17:29:13 +00:00
ad 0df165c04d Use xchgb, bswapl and not rotate/shift. 2007-11-28 17:03:35 +00:00
ad 13c86772e9 Memory barrier in respect of the CAS isn't necessary. 2007-11-28 16:59:09 +00:00
ad 79085586b9 Support files from the thorpej-atomic branch. 2007-11-28 16:54:57 +00:00
ad db42e14f44 Hook the atomic ops in. Right now they are only built into libkern, and
only if available.
2007-11-28 15:25:44 +00:00
ad b47fc5f5e0 Fix up a few minor problems. 2007-11-28 15:24:38 +00:00
ad 65728ed419 x86 atomic ops. 2007-11-28 01:33:47 +00:00
ad 51a2159851 Don't unconditionally clear the direction flag. The ABI says it must always
be clear when making a function call, and 'cld' takes about 50 clock cyles
on the P4.
2007-11-12 18:41:59 +00:00
mjf c05cd035e3 Fix typo 2007-11-06 11:41:35 +00:00
christos 2f1af36850 use the ffs32 from bitops.h which is faster 2007-11-02 21:05:06 +00:00
xtraeme d7ef1da684 Fix a typo that has been there forever. prop_dictionary_set:
the object will be retained by the dictionary not array.
2007-10-11 03:10:28 +00:00
he 2a0ce8a57f Improve consistency between <sys/bswap.h> and what's actually
defined by the C library for the various m68k ports by borrowing
some #ifs from the i386 port.

Also, align sun2 with the other m68k ports as to whether they
define __bswap{16,32} or bswap{16,32} in their C library (all
now define the __ variants).

This should make the m68k ports build sys/rump again, except
for sun2 which hits another problem later on.
2007-09-19 20:31:34 +00:00
xtraeme ed504658f5 tyop: guaranty -> guarantee. 2007-08-30 19:12:32 +00:00
joerg 4deb593145 Avoid using unbound amount of stack frames in prop_object_equal
by using a dynamic stack as well. Reorder arguments for the internalizer
as the iteration is always present and should go before possibly
NULL arguments.

Reviewed by mjf@ and adrianp@
2007-08-30 12:23:53 +00:00
pavel c0f560c80e struct plistref does not reference other proplib data types, so split it
in its own header file to be included by dkio.h. Fixes breakage due to
pollution from proplib.h in programs which include ioctl.h. Tested and OK
by dogcow@.
2007-08-17 11:05:03 +00:00
joerg e835604c26 libprop is currently using a recursive parser. While this is fine for
userland, deeply nested arrays and dictionaries can easily overflow
the kernel stack and thereby force a panic.

Fix the internalizer and prop_object_release to use a separate call
stack and alter the dictionary and array handling to not recurse on
the C stack. The default stack has an inline depth of 16 elements,
which should keep the overhead reasonable.

This issue was found by Pavel Cahyna and Jachym Holecek.

Additionally add a limit for prop_object_copyin_ioctl to prevent user
programs from temporary allocating unbound amount of kernel memory.
Allow malloc to fail so that tight loops of userland processes can't
force panics by exhausting the kernel map.

Tested with the sample exploit of Jachym, his test suite and reviewed
by himself (initial patch), Christos Zoulas and Jason Thorpe.
2007-08-16 21:44:06 +00:00
thorpej 04377267cc boolean_t -> bool
TRUE -> true
FALSE -> false
2007-08-16 16:28:17 +00:00