Commit Graph

45256 Commits

Author SHA1 Message Date
xtraeme
699c85af09 Make the Enhanced Speedstep driver available for i386 and amd64.
To use it on EM64T CPUs supporting the EST CPUID feature. Note that
some CPUs still don't work with this driver, like Xeon or Pentium 4.

Move the p[34]_get_bus_clock functions into its own file,
intel_busclock.c and remove this code from i386/identcpu.c.

Tested on i386 by myself and amd64 by Tonerre.
2007-06-03 19:32:26 +00:00
pavel
4d6ee26f48 Attempt to fix PR kern/36431 by including process_read_fpregs() and
process_read_regs() when COREDUMP is defined. These function are used
in core_elf32.c .
2007-06-03 09:38:16 +00:00
dogcow
9ab52b1142 Move isp* from INSTALL to INSTALL_LARGE; isp is gigantic enough that it
makes a noticable impact in the size of the kernel. (build release now works
again.)
2007-06-03 03:39:53 +00:00
uwe
c75253c460 Nuke traces of PCI. Minor cosmetic clean ups. 2007-06-03 01:52:54 +00:00
uwe
dcaa9bf80c Include igsfb_ofbusvar.h by its full official name. 2007-06-03 01:31:31 +00:00
uwe
217bc0b26b Make the name of the multiple-inclusion guard symbol match the file location. 2007-06-03 01:28:54 +00:00
tsutsui
04776a47dd Remove another #ifdef PCB_CMAP2 portion. 2007-06-02 23:51:37 +00:00
njoly
df2c7d05e6 Revert fs/gs changes for now.
This make clone(2) work again under NPTL compat linux emulation.

ok by fvdl.
2007-06-02 22:59:03 +00:00
uwe
cc652c9405 Add COMPAT_40. While here, make COMPAT_30 comment match other COMPAT_*. 2007-06-02 17:16:12 +00:00
tsutsui
0264c643f5 Rather than attaching fpu as a usual device during cpu_configure(9),
initialize it explicitly in identifycpu() (before cpu_configure(9))
because after yamt-idlelwp merge savectx() is called before
cpu_configure(9) and it requires fputype initialized properly,
otherwise a kernel fails in savectx() if a machine doesn't have an FPU.
2007-06-02 12:07:03 +00:00
tsutsui
5044d1542d Fix bus_dmamap_sync(9) to do more proper cache flush operations.
Mostly taken from mvme68k.

Tested on (not yet integrated) mac68k SONIC Ethernet with MI dp83932.c.

XXX: should we also enable _pmap_set_page_cache{able,inhibit}()
XXX: functions in m68k/pmap_motorola.c for BUS_DMA_COHERENT support?
2007-06-02 11:13:45 +00:00
tsutsui
9955baccd9 Add #define __NO_STRICT_ALIGNMENT. 2007-06-02 06:34:21 +00:00
tsutsui
59a3d07b9a Don't try to clear BSS if bootinfo is passed
(i.e. running kernel is loaded by the native bootloader).
2007-06-02 06:30:17 +00:00
tsutsui
f4ba139f1a Use some more ANSI character escape sequences in keymaps. 2007-06-02 06:27:12 +00:00
tsutsui
b1102490c5 Sort some keymap entries. 2007-06-02 06:24:56 +00:00
tsutsui
c0f042edea - Don't refer fputype before FPU is probed.
- Call m68k_make_fpu_idle_frame() from initfpu().
2007-06-02 06:23:56 +00:00
nisimura
404200bf24 - add MPC8245 to cpuname table. 2007-06-02 02:41:41 +00:00
uwe
4a747421e0 Rename j6x0tp_set_enable to j6x0tp_enable_child which better reflects
what this function does and is way less confusing too.
2007-06-01 18:23:46 +00:00
uwe
aa1315c604 Cosmetic. Use device_xname. Don't compile debugging support. Drop
out-of-date comment for j6x0tp_attach (we haven't been using submatch
for quite a while).
2007-06-01 17:44:46 +00:00
uwe
6542efa4c8 Add COMPAT_40. 2007-06-01 14:30:32 +00:00
blymn
12de420234 Add smsc device 2007-06-01 14:26:36 +00:00
nisimura
c966ccb8d8 - add booted_device and booted_kernel sysctl nodes to let userland know
about runtime environment.
2007-06-01 14:23:50 +00:00
blymn
3099e9ed7a Added smsc device. 2007-06-01 14:14:16 +00:00
uwe
0390c5df8f Add COMPAT_40. 2007-06-01 13:20:56 +00:00
skrll
322ba9d3c4 Another missed changed for acorn26 idlelwp. Provide IRQenable/IRQdisable
macros.
2007-06-01 11:31:01 +00:00
skrll
4868ebda4e want_resched is no longer required here. 2007-06-01 07:21:51 +00:00
skrll
afcc3d424c Missed this in previous commit.
Attempt to adacpt acorn26 to idlelwp. This is untested.

OK'd by Ben Harris
2007-06-01 07:21:26 +00:00
skrll
fb901b95c7 __HAVE_MD_RUNQUEUE is no longer relevant. 2007-06-01 07:18:31 +00:00
skrll
a6c4dd7bcd Attempt to adacpt acorn26 to idlelwp. This is untested.
OK'd by Ben Harris
2007-06-01 07:04:53 +00:00
mhitch
fb64759fb7 Get rid of one more incompatibility with pmap_motorola.c from pmap_bootstrap(). 2007-06-01 04:05:05 +00:00
mhitch
d9ac73c21c PTE for Sysptmap was not getting set correctly. 2007-05-31 22:16:48 +00:00
christos
3fe1c0adb5 lock_stubs.S depends on assym.h 2007-05-30 17:16:55 +00:00
christos
3ce718d4da Add compat 4.0 2007-05-29 23:21:52 +00:00
christos
cb17d4ac11 Add COMPAT_40 for struct ifreq changes. 2007-05-29 21:37:07 +00:00
macallan
b0698fecd5 cpu_info.ci_idle_pcb is gone so don't try to set it
probably needs more work but it works fine like that at least on my PB3400c
2007-05-29 13:26:39 +00:00
tsutsui
82e46da13f Use the kernel symbol table to see the beginning of the current
subroutine to get more proper backtrace on ddb(4).

In the previous code it scans backwards from the current PC
for the end of the previous subroutine and checks "jr ra" or
"jr k0" instructions, but it often fails because gcc is
so aggressive nowadays as to reorder instruction blocks
to create efficient code path by branch predict etc. and
"jr ra" is not always located at the end of subroutines.

No objection on port-mips.
2007-05-29 12:03:45 +00:00
mrg
565d4d6ca6 disable mutex_spin_enter() & mutex_spin_exit() as they are currently
broken and cause MULTIPROCESSOR kernels to hang while entering
userland boot up.
2007-05-28 22:56:19 +00:00
mhitch
82970a259f Got rid of too much - need to allocate virtual space for msgbuf. 2007-05-28 22:36:15 +00:00
pavel
c9cec95fb7 Move ixpide from INSTALL to INSTALL_LARGE. This controller is apparently
found only in southbridges of relatively recent motherboards.
2007-05-28 21:54:38 +00:00
mrg
b4ff11c2dd fix some NOTDEF_DEBUG code. 2007-05-28 21:52:49 +00:00
pavel
e09c8c416e Move puc, ahcisata, nfe, svwsata from INSTALL to INSTALL_LARGE to help fit
INSTALL to a 2.88 MB floppy image. Suggested by Manuel Bouyer.
2007-05-28 21:45:29 +00:00
mrg
839b86dc5e cpu_attach(): in the MP case, fix up idle lwp info as well as lwp0.l_cpu.
for LOCKDEBUG, mutex_destroy() the cpu0 spc_lwplock with the global
cpuinfo VA, and re-mutex_init() it with the per-cpu address that is only
now available.  for non-boot cpus, be sure to also initialise curlwp to
the idle lwp.

xcall(), pmap_quiet_check(), pmap_pmap_pool_ctor(), pmap_pmap_pool_dtor(),
and pmap_enu4m(): don't care about cpus that have not been attached yet.
2007-05-28 21:24:17 +00:00
mrg
8f8cea5c68 clean up cpu_switchto() a little - no need to flushw or membar
excessively.  optimise an rdpr.  tested by tnn.
2007-05-28 20:09:50 +00:00
mhitch
5a0ae3b2c5 Oops. I thought I had caught this before commiting the CMAP[12] removal,
but obviously I didn't.  Non-DEBUG kernels were unbootable because I left
some code inside an #ifdef DEBUG group.
2007-05-27 21:33:55 +00:00
uwe
19eb8fc1bc Need to use atop() when adjusting kernend for ksyms.
XXX: This code has obviously bit-rotted.

Pointed out by Anders Gavare on port-sh3.
2007-05-27 18:30:01 +00:00
tsutsui
2fc470a315 Reorder IPL_SOFTSERIAL and IPL_SOFTCLOCK properly. 2007-05-27 14:22:36 +00:00
tsutsui
cd02a491e7 - Don't clear BSS in mach_init() if a kernel is loaded
by our native bootloader.
- Restore lwp0.l_cpu (which is required for curcpu()) and
  cpu_info_store.ci_curlwp after clearing BSS in case
  a kernel is loaded by the firmware directly since these
  values are in BSS and initilized before mach_init() is called.
  (actually they are restored in mips_machdep.c:mips_vector_init()
   but we use curcpu() earlier than that point)

Fixes silent hang right after boot on cobalt.
2007-05-27 14:10:49 +00:00
tsutsui
5964c0cd2d Pass correct args to lwp_startup() in lwp_trampoline().
Should fix "panic: TLB out of universe" on MIPS3 machines.
2007-05-27 13:57:30 +00:00
uwe
269ff6629e In sh3_tlb_update we need to ensure we are not interrupted while
meesing with the TLB.  That would usually show up as misterious kernel
crashes under heavy load.

Prettify the code while here so that traversal of the memory mapped
TLB arrays looks congruent in all functions.

Fixes PR 34706
2007-05-27 12:21:24 +00:00
he
e0ce1262e3 Bump SYMTAB_SPACE so that it fits again. 2007-05-27 11:47:11 +00:00