4013 Commits

Author SHA1 Message Date
bde
f11d5307d3 Adjusted FAKE_MCOUNT()s for amd64. This is needed for both ordinary
and high resolution profiling of interrupt handlers.  The adjustments
are routine once the magic stack offset 13*4 is decoded to be TF_RIP
(there were originally more types of stack frames so using TF_EIP for
one of them wouldn't have been much simpler).

Removed garbage comments attached to some of the FAKE_MCOUNT()s.
2004-05-23 16:23:29 +00:00
bde
e3c5341a54 Spell "retq" as "ret" in pagezero() like it is everywhere, else so
that the usual macro for "ret" hides the detail of calling .mexitcount
before returning.

Fixed missing call to .mexitcount in lgdt().  This was missing on
i386's, mainly because lgdt() uses lret[q] insted of ret.  This is
very unimportant since lgdt() is not (normally?) called until after
profiling is initialized.
2004-05-23 14:56:02 +00:00
bde
37bf942f68 MFi386 (1.103 and 1.104: fixed some problems in high resolution profiling
and improved some comments).  Also, made the documented {f,s}uword()
functions the standard entry points and the undocumented {f,s}uword64()
functions alternative entry points, like {f,s}uword32() for i386's.  The
bitrot in the comments was a little larger here -- there are new undocumented
32-bit sub-word functions, not just renaming of 16-bit functions from
documented ones to undocumented ones.
2004-05-21 16:50:57 +00:00
bde
5e189e571a MFi386 (1.37: GUPROF calibration macros; only routine adjustments needed). 2004-05-20 16:22:57 +00:00
peter
acd31bc19e Like on i386, clear the last three entries in the pml4 page when doing a
pmap_release(), and put it the free queue marked as already zeroed.
2004-05-19 21:55:37 +00:00
bde
a128646d2f Fixed the type of fptrdiff_t. It needs to be 64 bits in theory, and in
practice too since kernel addresses are almost 2^64 higher than most
user addresses.
2004-05-19 16:19:11 +00:00
bde
1808931bc7 Fixed some style bugs (mainly misalignment of backslashes). 2004-05-19 16:04:26 +00:00
bde
f309a749f5 Moved most of the "MI" definitions and declarations from <machine/profile.h>
to <sys/gmon.h>.  Cleaned them up a little by not attempting to ifdef
for incomplete and out of date support for GUPROF in userland, as in
the sparc64 version.
2004-05-19 15:41:26 +00:00
peter
fecf8b1849 Unbreak builds without DDB. Bad Bruce! No cookie! :-) 2004-05-19 01:23:48 +00:00
peter
18fd75c1c3 The 'call mcount' hooks that gcc inserts when profiling are in a place that
cannot handle the scratch registers being trashed.  So we have to preserve
them ourselves.
2004-05-18 22:52:32 +00:00
stefanf
e784f59f15 <stdint.h> should define WINT_M{AX,IN} independent from whether WCHAR_MIN is
defined.  Otherwise first including <wchar.h> and then <stdint.h> leads to no
WINT_M{AX,IN} at all.

PR:		64956
Approved by:	das (mentor)
2004-05-18 16:04:57 +00:00
bde
28faba2e27 Fixed DDB_NOKLDSYM on amd64's:
machdep.c:
Initialize the symbol table pointers, not quite like for other arches.

db_elf.c:
Don't claim to be an i486 in the fake ELF header.
2004-05-18 05:30:06 +00:00
peter
4070a4a8d4 Turn on modules for amd64. Fear. 2004-05-17 22:13:14 +00:00
peter
d0605f2d9e Deal with REL records that have the addend embedded variable sized targets
rather than the RELA table.  I dont know if bintutils will ever generate
REL records, but just in case.....
2004-05-17 21:16:49 +00:00
peter
8b77ed77bd Checkpoint some of what I was starting to tinker with for having some
different context support for 32 vs 64 bit processes.  This simply omits
the save/restore of the segment selector registers for non 32 bit
processes.  This avoids the rdmsr/rwmsr juggling when restoring %gs
clobbers the kernel msr that holds the gsbase.

However, I suspect it might be better to conditionally do this at
user<->kernel transition where we wouldn't need to do the juggling in the
first place.  Or have per-thread extended context save/restore hooks.
2004-05-16 22:43:57 +00:00
peter
be73fb5080 Kill the LAZYPMAP ifdefs. While they worked, they didn't do anything
to help the AMD cpus (which have a hardware tlb flush filter).  I held
off to see what the 64 bit Intel cpus did, but it doesn't seem to help
much there either.  Oh well, store it in the Attic.
2004-05-16 22:11:50 +00:00
peter
a3df9de5e8 Converge some more with i386. 2004-05-16 21:27:29 +00:00
peter
f0d8d38564 MFi386: add rue and twa 2004-05-16 20:57:01 +00:00
peter
cd0ccfc778 MFi386: avoid partial register references, for what its worth. 2004-05-16 20:46:13 +00:00
peter
172c58b998 For consistency with i386, have pmap_kenter_temporary() take a vm_paddr_t
argument.  It is actually the same type on amd64 (vm_paddr_t = vm_offset_t)
but this reduces the i386<->amd64 diffs a little.
2004-05-16 20:44:41 +00:00
peter
c6a708cab1 MFi386: numerous interrupt and acpi updates 2004-05-16 20:30:47 +00:00
peter
ea4215c521 Make a small revision to the api between the elf linker core and the
elf_reloc() backends for two reasons.  First, to support the possibility
of there being two elf linkers in the kernel (eg: amd64), and second, to
pass the relocbase explicitly (for relocating .o format kld files).
2004-05-16 20:00:28 +00:00
njl
3d06d54b9d Make unnecessary globals static and remove unused includes.
Pointed out by:	cscout
2004-05-06 02:18:58 +00:00
njl
f6bbab06a9 Add an MI implementation of the ACPI global lock routines and retire the
individual asm versions.  The global lock is shared between the BIOS and
OS and thus cannot use our mutexes.  It is defined in section 5.2.9.1 of
the ACPI specification.

Reviewed by:	marcel, bde, jhb
2004-05-05 20:04:14 +00:00
jhb
9f40cdcc06 Add a simple mini-driver for the ELCR register. Originally, the ELCR
register controlled the trigger mode and polarity of EISA interrupts.
However, it appears that most (all?) PCI systems use the ELCR to manage
the trigger mode and polarity of ISA interrupts as well since ISA IRQs used
to route PCI interrupts need to be level triggered with active low
polarity.  We check to see if the ELCR exists by sanity checking the value
we get back ensuring that IRQS 0 (8254), 1 (atkbd), 2 (the link from the
slave PIC), and 8 (RTC) are all clear indicating edge trigger and active
high polarity.

This mini-driver will be used by the atpic driver to manage the trigger and
polarity of ISA IRQs.  Also, the mptable parsing code will use this mini
driver rather than examining the ELCR directly.
2004-05-04 20:07:46 +00:00
marcel
381147f827 Add option GEOM_GPT. This brings the ability to have a large number of
partitions on a single disk.
2004-05-02 20:40:19 +00:00
obrien
3937b2c0ab Spell Ethernet correctly. 2004-05-02 18:57:29 +00:00
das
0d58ef0153 Hide FLT_EVAL_METHOD and DECIMAL_DIG in pre-C99 compilation
environments.

PR:		63935
Submitted by:	Stefan Farfeleder <stefan@fafoe.narf.at>
2004-04-25 02:36:29 +00:00
njl
d3289512e2 Don't check for NULL, device_get_softc() always succeeds. 2004-04-21 02:10:58 +00:00
alc
f2547e07b8 Simplify the sf_buf implementation. In short, make it a trivial veneer
over the direct virtual-to-physical mapping.
2004-04-18 05:36:37 +00:00
alc
c4d18d39a4 Set the "global" attribute on the page table entries for the kernel and
direct mappings.  This shaves a few seconds off of my buildworld times.

Discussed with:	peter@
2004-04-16 03:45:28 +00:00
alc
ec17c32ba1 - is_physical_memory()'s parameter, which is a physical address, should be
a vm_paddr_t not a vm_offset_t.
2004-04-11 04:26:58 +00:00
alc
c380417937 - pmap_kenter_temporary() is unused by machine-independent code. Therefore,
move its declaration to the machine-dependent header file on those
   machines that use it.  In principle, only i386 should have it.
   Alpha and AMD64 should use their direct virtual-to-physical mapping.
 - Remove pmap_kenter_temporary() from ia64.  It is unused.  Approved
   by: marcel@
2004-04-10 22:41:46 +00:00
peter
bc5252478b Update to include both the L1 and L2 TLB stats, as well as the seperate
2M/4M page TLB vs 4K page TLB stats.  This also applies to the i386
platform, as does the cpu features fixes.
2004-04-07 00:44:15 +00:00
peter
b035268c4a MFi386: move rss() from db_interface.c to cpufunc.h 2004-04-07 00:41:05 +00:00
imp
ceab5f79a3 Remove advertising clause from University of California Regent's license,
per letter dated July 22, 1999 and email from Peter Wemm.

Approved by: core, peter
2004-04-05 23:55:14 +00:00
imp
21f7c26600 Remove advertising clause from University of California Regent's license,
per letter dated July 22, 1999.

Approved by: core
2004-04-05 21:29:41 +00:00
imp
0f03c03f09 Remove advertising clause from University of California Regent's license,
per letter dated July 22, 1999 and permission from Alan Cox.

Approved by: core, alc@
2004-04-05 21:15:52 +00:00
alc
19e5eda309 Remove unused arguments from pmap_init(). 2004-04-05 00:37:50 +00:00
alc
d74988558f Remove ptmmap and ptvmmap. They are unused on amd64. 2004-04-03 20:53:50 +00:00
alc
1ec4d75266 In some cases, sf_buf_alloc() should sleep with pri PCATCH; in others, it
should not.  Add a new parameter so that the caller can specify which is
the case.

Reported by:	dillon
2004-04-03 09:16:27 +00:00
alc
f18821928b Microoptimize pagezero() based upon something that I learned writing the
optimized pagecopy().  This also has the virtual of making these two
functions more similar in style.
2004-04-03 05:33:10 +00:00
alc
8294bebe0a - Add an optimized page copy function for use by pmap_copy_page(). It is
roughly four times faster than bcopy() for uncached pages.
 - Sort the function prototypes in md_var.h.
2004-03-31 02:03:49 +00:00
peter
09ff2f9fe5 Finish tidying up a couple of leftovers from the KSTACK_PAGES stuff. Some
files still #included the opt_ file.  powerpc hadn't been updated yet.
2004-03-29 19:38:05 +00:00
peter
b920a97333 MFi386: correctly calculate the top-of-stack when a kthread is created
with a larger kernel stack.
2004-03-25 03:39:38 +00:00
peter
021cb09919 Run print_AMD_features() for both AuthenticAMD and GenuineIntel cpus.
Report the %ecx bits in cpuid function 1.  This is a hack.
When reporting AMD Features, only mask off the common bits.  Otherwise
the SEP bit masks off SYSCALL etc in the report.
2004-03-25 03:38:31 +00:00
obrien
596137082c Add NTFS since many may want to dual-boot MS-Win64 w/FreeBSD. 2004-03-25 03:19:28 +00:00
alc
293aebf5cc - Add uiomove_fromphys() implementations to alpha and ia64. These only
differ trivially from amd64.
 - Correct a spelling error in a comment.
2004-03-20 21:06:20 +00:00
marcel
7e4265388d Introduce the cpumask_t type. The purpose of the type is to create a
level of abstraction for any and all CPU mask and CPU bitmap variables
so that platforms have the ability to break free from the hard limit
of 32 CPUs, simply because we don't have more bits in an u_int. Note
that the type is not supposed to solve massive parallelism, where
the number of CPUs can be larger than the width of the widest integral
type. As such, cpumask_t is not supposed to be a compound type. If
such would be necessary in the future, we can deal with the issues
then and there. For now, it can be assumed that the type is integral
and unsigned.

With this commit, all MD definitions start off as u_int. This allows
us to phase-in cpumask_t at our leasure without breaking anything.
Once cpumask_t is used consistently, platforms can switch to wider
(or smaller) types if such would be beneficial (or not; whatever :-)

Compile-tested on: i386
2004-03-20 20:41:40 +00:00
alc
2490c09470 Introduce uiomove_fromphys(). This is a variant of uiomove() that takes
a collection of physical pages as the source.  On amd64 it is implemented
using the direct virtual-to-physical map.
2004-03-20 19:36:29 +00:00