Commit Graph

4067 Commits

Author SHA1 Message Date
Nate Lawson
37c55a039a Revert last change. If acpi is loaded or compiled into the kernel, its
devclass will be present even if the driver was disabled by a hint.  Using
device_get_softc() provides the right info even if it's overkill.

Explained by:	jhb
2004-06-17 17:27:37 +00:00
Alan Cox
d45f21f31a Do not preset PG_BUSY on VM_ALLOC_NOOBJ pages. Such pages are not
accessible through an object.  Thus, PG_BUSY serves no purpose.
2004-06-17 06:16:58 +00:00
Poul-Henning Kamp
89c9c53da0 Do the dreaded s/dev_t/struct cdev */
Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
Alan Cox
d7c3bd4918 Add some lock assertions. Lock a small part of pmap_enter(). 2004-06-16 07:51:19 +00:00
Alan Cox
82d8e6f5a0 Correct an error in the implementation of pmap_is_prefaultable(). When I
introduced this function in revision 1.441, I inverted one of the
comparisons.
2004-06-16 03:11:24 +00:00
Alan Cox
7b9d474460 Remove a stale comment. 2004-06-15 19:28:40 +00:00
Alan Cox
0dec7f69a6 Add pmap locking to pmap_extract(), pmap_mincore(), and pmap_remove(). 2004-06-15 07:41:44 +00:00
Nate Lawson
345281bc43 We only need the devclass_find() result, not the softc. 2004-06-15 02:12:12 +00:00
Alan Cox
50f91a9445 Introduce pmap locking to many of the pmap functions. There is more to
come later.
2004-06-14 01:17:50 +00:00
David E. O'Brien
2a79761d77 The majority of FreeBSD/amd64 machines are SMP, so use ADAPTIVE_MUTEXES
by default to improve performance.
2004-06-13 23:03:57 +00:00
Alan Cox
7881f95056 Prevent the loss of a PG_M bit through an SMP race in pmap_ts_referenced(). 2004-06-13 21:59:42 +00:00
Alan Cox
b34ec165b4 Remove dead or unneeded code, e.g., spl calls. 2004-06-13 19:48:38 +00:00
Alan Cox
8559e0a291 - Remove an unused declaration.
- Move a definition inside the scope of a #ifdef _KERNEL.
2004-06-13 03:44:11 +00:00
Alan Cox
2d0dc0fcd6 In a multiprocessor, the PG_W bit in the pte must be changed atomically.
Otherwise, the setting of the PG_M bit by one processor could be lost if
another processor is simultaneously changing the PG_W bit.

Reviewed by:	tegge@
2004-06-12 20:01:48 +00:00
Poul-Henning Kamp
1930e303cf Deorbit COMPAT_SUNOS.
We inherited this from the sparc32 port of BSD4.4-Lite1.  We have neither
a sparc32 port nor a SunOS4.x compatibility desire these days.
2004-06-11 11:16:26 +00:00
Peter Wemm
a520047095 Argh. Add the mini-stack-frame back in for mcount's benefit for syscall
stubs.
2004-06-10 22:02:26 +00:00
Peter Wemm
6d05d7c75a Make profiling work for varargs functions.. %al is an additional argument
which indicates the number of xmm registers used in the varargs.  This
stops the explosion that happened when profiling printf() etc.
2004-06-10 22:00:58 +00:00
Peter Wemm
11b000253e Insta-MFi386: ignore disabled cpu apic id's entirely 2004-06-10 21:30:08 +00:00
John Baldwin
bad4ce7d91 - Use the correct devclass name ("acpi" vs "ACPI") to detect if acpi0 is
present and thus that the PnPBIOS probe should be skipped instead of
  having ACPI zero out the PnPBIOStable pointer.
- Make the PnPBIOStable pointer static to i386/i386/bios.c now that that is
  the only place it is used.
2004-06-10 20:43:04 +00:00
John Baldwin
092a5c4530 Remove atdevbase and replace it's remaining uses with direct references to
KERNBASE instead.
2004-06-10 20:31:00 +00:00
Peter Wemm
591506d322 In pmap_extract_and_hold(), there is no need to mask off PG_FRAME because
pmap_extract() already does it.
In pmap_enter(), opa has already been masked so don't do it again.
Wrap a long line (recent transgression).
Use trunc_page() in pmap_mapdev() instead of anding with PG_FRAME, since
that is what we really meant.

Submitted by:  alc (first item)
2004-06-08 02:20:40 +00:00
Peter Wemm
71e0fe3abc Fix my silly typo in asm statement in previous commit. 2004-06-08 01:35:48 +00:00
Peter Wemm
576bb07aaa Argh. Remove stray number that slipped into the previous commit. 2004-06-08 01:20:37 +00:00
Peter Wemm
96a7759e99 Reapply rev 1.151 after enable sse/fpuinit order fixed in mp_machdep.c
Obtained from:  das
2004-06-08 01:14:39 +00:00
Peter Wemm
18154cd6f8 Set up the fpu *after* enabling SSE mode on AP's
Submitted by: (argh, I can't find the email)
2004-06-08 01:07:51 +00:00
Peter Wemm
430e272c7e Initial PG_NX support (no-execute page bit)
- export the rest of the cpu features (and amd's features).
- turn on EFER_NXE, depending on the NX amd feature bit
- reorg the identcpu stuff a bit in order to stop treating the
  amd features as second class features (since it is now a primary feature
  bit set) and make it easier to export.
2004-06-08 01:02:52 +00:00
Peter Wemm
7d95d34bb7 Mask pte's with PG_FRAME before passing it to PHYS_TO_VM_PAGE().. PG_NX
lives in the top 12 'available' bits.  atop() in the PHYS_TO_VM_PAGE()
macro only masks off the lower bits (by accident) and the upper bits
in the 64 bit ptes turn into "interesting" index values.
2004-06-08 00:29:42 +00:00
Peter Wemm
74567f8f85 Use trunc_page(va) when we mean it rather than anding it with PG_FRAME
(which doesn't work all that well when there are bits at the top that are
 masked by PG_FRAME)
2004-06-08 00:11:32 +00:00
Peter Wemm
c2490f2dfb Fix a serious problem that manifested during swap, and a few other times.
pmap_remove() would be called with a huge range and we'd stride across
it in only 2MB chunks.  This would manifest as massive cpu time and a
largely unresponsive system during hard swap.  Instead, check the higher
page directories which means we can run pmap_remove() in just a few
hundred loop iterations instead of millions since we can process
address space in chunks of 512GB and 1GB as well as 2MB.

Eternal thanks to:  tmm
2004-06-07 23:51:20 +00:00
Peter Wemm
b8168edefc Be a little more consistent in the naming of the PML4 defines. 2004-06-07 23:47:59 +00:00
David Schultz
8c2267ec9a Back out revision 1.150, since dwmalone reports that it causes a panic
upon startup on his machine.
2004-06-06 09:16:02 +00:00
David Schultz
ad070467cd Initialize the MXCSR to the appropriate default value at startup.
Tested on:	tjr
2004-06-05 03:13:39 +00:00
Poul-Henning Kamp
79005bbdbe Add new bios_string() which will hunt for a string inside a given range
of the BIOS.  This can be used for finding arbitrary magic in the BIOS
in order to recognize particular platforms.
2004-06-03 22:36:24 +00:00
Peter Wemm
74cfa96999 MFi386: add ixgp device 2004-06-03 21:40:41 +00:00
Peter Wemm
ea10166e8a MFi386: apic intpin programming updates etc. 2004-06-03 20:25:05 +00:00
Peter Wemm
bfe14b3edc MFi386: remove debug printf 2004-06-03 20:22:48 +00:00
Peter Wemm
b5eb4e5196 Move module.h include to the same place as on i386 for diff reduction. 2004-06-03 20:21:30 +00:00
Peter Wemm
9248fc7bc0 MFi386: move cpu_nameclass struct next to its only consumer 2004-06-03 20:18:15 +00:00
Tim J. Robbins
cc05397ffc Remove checks for curthread == NULL - it can't happen. 2004-06-03 10:22:47 +00:00
Poul-Henning Kamp
fd360128ff Add missing <sys/module.h> instances which were shadowed by the nested
include in <sys/kernel.h>
2004-06-03 05:58:30 +00:00
Tim J. Robbins
fa2a4d0595 Move TDF_DEADLKTREAT into td_pflags (and rename it accordingly) to avoid
having to acquire sched_lock when manipulating it in lockmgr(), uiomove(),
and uiomove_fromphys().

Reviewed by:	jhb
2004-06-03 01:47:37 +00:00
Tim J. Robbins
aa0aa7a113 Move TDF_SA from td_flags to td_pflags (and rename it accordingly)
so that it is no longer necessary to hold sched_lock while
manipulating it.

Reviewed by:	davidxu
2004-06-02 07:52:36 +00:00
Poul-Henning Kamp
41ee9f1c69 Add some missing <sys/module.h> includes which are masked by the
one on death-row in <sys/kernel.h>
2004-05-30 17:57:46 +00:00
Alan Cox
b59f545aa2 MFi386 revision 1.6
Reenable ithread preemption for interrupts that occur while executing in
 the kernel.
2004-05-30 04:49:39 +00:00
Tim J. Robbins
402705521a Implement __bb_init_func. This is a fairly straightforward conversion
of the i386 version.
2004-05-29 01:13:28 +00:00
Alan Cox
662d471da6 Remove a broken micro-optimization from pmap_enter(). The ill effect
of this micro-optimization occurs when we call pmap_enter() to wire an
already mapped page.  Because of the micro-optimization, we fail to
mark the PTE as wired.  Later, on teardown of the address space,
pmap_remove_pages() destroys the PTE before vm_fault_unwire() has
unwired the page.  (pmap_remove_pages() is not supposed to destroy
wired PTEs.  They are destroyed by a later call to pmap_remove().)
Thus, the page becomes lost.

Note: The page is not lost if the application called munlock(2), only
if it relies on teardown of the address space to unwire its pages.

For the historically inclined, this bug was introduced by a
megacommit, revision 1.182, roughly six years ago.

Leak observed by: green@ and dillon independently
Patch submitted by: dillon at backplane dot com
Reviewed by: tegge@
MFC after: 1 week
2004-05-28 19:42:02 +00:00
Thomas Moestl
65e29c4822 Retire cpu_sched_exit(); it is not used any more. 2004-05-26 12:09:39 +00:00
Bruce Evans
026afdcc05 Quick fix for overflow when tsc_freq >= 2^31. "int profrate" in struct
gmon and struct gmonhdr was originally just to represent the kernel
(profiling) clock frequency and it remains poorly suited to representing
the frequencies of fast counters like the TSC.  It broke a year or two
ago.  This quick fix keeps it working for another year or month or two
until TSC frequencies can exceed 2^32, by dividing the frequency by 2.
Dividing the frequency by 4 would work for a little longer but would
lose a little too much precision.
2004-05-26 09:43:38 +00:00
Bruce Evans
d7d197e0f9 Oops, ".align 4" for the data section in the previous commit should
have been ".p2align 4".  This bug is cosmetic since the data section
happens to be empty.
2004-05-24 12:42:16 +00:00
Bruce Evans
003d5d66b1 Fixed profiling of trap, syscall and interrupt handlers and some
ordinary functions, essentially by backing out half of rev.1.115 of
amd64/exception.S.  The handlers must be between certain labels for
the purposes of profiling, and this was broken by scattering them in
separately compiled .S files, especially for ordinary functions that
ended up between the labels.  Merge the files by #including them as
before, except with different pathnames and better comments and
organization.  Changes to the scattered files are minimal -- just
move the labels to the file that does the #includes.

This also partly fixes profiling of IPIs -- all IPI handlers are now
correctly classified as interrupt handlers, but many are still missing
mcount calls.
2004-05-24 12:08:56 +00:00