Commit Graph

9585 Commits

Author SHA1 Message Date
Bruce Evans
4e72678dee Fixed panics in apic interrupt handlers if kernel profiling is turned
on.  MCOUNT and FAKE_MCOUNT() may clobber all the call-used registers,
and one FAKE_MCOUNT() was placed so that an active %eax was clobbered.
The fix is to move this FAKE_MCOUNT() earlier where it should have
been anyway.

Fixed 3 layers of bitrot in the comment about why this FAKE_MCOUNT()
was where it was by removing the comment.  (mcount() should be called
as early as possible after entering a new level, but an implementation
detail got in the way until 3 layers of changes ago.)

Kernel profiling still gives wrong results because the new interrupt
code rearranged object files too much.  mcount() depends on trap,
syscall and interrupt handlers being between certain magic labels with
interrupt handlers last, and on nothing else being there.  Splitting
up exception.o moved the magic labels to effectively random places
relative to what they are supposed to delimit.  This mainly broke the
call graph; the flat profile is still usable.
2003-12-02 15:02:12 +00:00
Bruce Evans
933a354fe5 Fixed breakage of the pci case of the cy driver by the new interrupt
code.  Both the driver and the new code were wrong.  Driver interrupt
handlers are supposed to take "void *vsc" arg, but some including all
COMPAT_ISA drivers and the pci part of the cy driver want an "int unit"
arg.  They got this using bogus casts of function pointers which should
have kept working despite their bogusness.  However, the new interrupt
code doesn't honor requests to pass an arg of ((void *)0), so things
are very broken if the arg is actually a representation of unit 0.

The fix is to use a normal "void *vsc" arg for the pci case and a
wrapper for the COMPAT_ISA case (of the cy driver).  This cleans up
new-busification of the pci case but takes the COMPAT_ISA case a little
further from new-bus.  The corresponding bug for the COMPAT_ISA case
has already been fixed similarly using a wrapper in compat_isa.c and
we need another wrapper just to undo that.

Fixed some directly related style bugs (mainly by removing compatibility
cruft).

cy.c:
Fixed an indirectly related old bug in cyattach_common().  A wrong status
was returned in the unlikely event that malloc() failed.

Approved by:	re (scottl)
2003-12-02 12:36:00 +00:00
Poul-Henning Kamp
173c971bd6 Refactor AMD Elan 520 CPU support.
Make it possible to configure GPIO pins as led(4) devices, PPS inputs
and PPS-echo outputs with a sysctl.  Led(4) and PPS-echo can be configured
for active-high or active-low.

Be more complete in initialization of timecounter hardware.

Approved by:	re@
2003-11-27 20:27:29 +00:00
John Baldwin
798a45964d - Split cpu_mp_probe() into two parts. cpu_mp_setmaxid() is still called
very early (SI_SUB_TUNABLES - 1) and is responsible for setting mp_maxid.
  cpu_mp_probe() is now called at SI_SUB_CPU and determines if SMP is
  actually present and sets mp_ncpus and all_cpus.  Splitting these up
  allows an architecture to probe CPUs later than SI_SUB_TUNABLES by just
  setting mp_maxid to MAXCPU in cpu_mp_setmaxid().  This could allow the
  CPU probing code to live in a module, for example, since modules
  sysinit's in modules cannot be invoked prior to SI_SUB_KLD.  This is
  needed to re-enable the ACPI module on i386.
- For the alpha SMP probing code, use LOCATE_PCS() instead of duplicating
  its contents in a few places.  Also, add a smp_cpu_enabled() function
  to avoid duplicating some code.  There is room for further code
  reduction later since much of this code is also present in cpu_mp_start().
- All archs besides i386 still set mp_maxid to the same values they set it
  to before this change.  i386 now sets mp_maxid to MAXCPU.

Tested on:	alpha, amd64, i386, ia64, sparc64
Approved by:	re (scottl)
2003-11-21 22:23:26 +00:00
Peter Wemm
c287f723c4 MFamd64: use a less compiler-intensive MD implementation of 'curthread'
so that the compiler doesn't have to do so much work.

Approved by:  re (jhb)
2003-11-20 23:23:22 +00:00
John Baldwin
6006b8f4c6 Add a special check for a stray IRQ 7 or IRQ 15 to see if it is actually
a spurious interrupt from one of the 8259As.  If so, don't log it as a
stray IRQ, but just silently ignore it.

Approved by:	re (rwatson)
2003-11-19 15:40:23 +00:00
John Baldwin
8c770ed571 - Add counts to the ATPIC interrupt sources and point the ATPIC interrupt
source count pointers at them so that intr_execute_handlers() won't
  choke when it tries to handle an unregisterd ATPIC interrupt source.
- Install the low-level ATPIC interrupt handlers when we first program the
  ATPIC in atpic_startup() rather than at SI_SUB_INTR.  This is only
  necessary to work around buggy code that enables interrupts too early
  in the boot process (namely, the vm86 code).

Approved by:	re (rwatson)
2003-11-19 15:38:56 +00:00
Tim J. Robbins
7e1985fcd8 Replace the dangerous strcpy() call with strlcpy(), instead of the safe one
that was incorrectly changed in rev. 1.61.

Approved by:	re
2003-11-18 14:21:34 +00:00
Alan Cox
0543fa5398 - Change the i386's sf_buf implementation so that it never allocates
more than one sf_buf for one vm_page.  To accomplish this, we add
   a global hash table mapping vm_pages to sf_bufs and a reference
   count to each sf_buf.  (This is similar to the patches for RELENG_4
   at http://www.cs.princeton.edu/~yruan/debox/.)

   For the uninitiated, an sf_buf is nothing more than a kernel virtual
   address that is used for temporary virtual-to-physical mappings by
   sendfile(2) and zero-copy sockets.  As such, there is no reason for
   one vm_page to have several sf_bufs mapping it.  In fact, using more
   than one sf_buf for a single vm_page increases the likelihood that
   sendfile(2) blocks, hurting throughput.
   (See http://www.cs.princeton.edu/~yruan/debox/.)
2003-11-17 18:22:24 +00:00
Peter Wemm
0bfbe7b935 Widen the enable/disable helper function's argument in line with the
ithread_create() changes etc.  This should be mostly a NOP.
2003-11-17 06:10:15 +00:00
Bruce Evans
6a0d1abc5d Fixed pedantic warnings for statement-expressions using __extension__
and by not using a statement-expression for the non-expression
__PCPU_SET().
2003-11-17 04:40:58 +00:00
Bruce Evans
81bbee5996 Fixed a pedantic syntax error (a stray semicolon at the end of
PCPU_MD_FIELDS).
2003-11-17 03:40:41 +00:00
Bruce Evans
6d800f89c4 Fixed pedantic syntax errors. Many macros didn't permit a semicolon after
their invocation in the !KLD_MODULE case, but a semicolon is provided after
all invocations and is required in the KLD_MODULE case.
2003-11-17 02:55:25 +00:00
Bruce Evans
ed9800b4de Avoid a warning for compiling with `gcc -Wbad-function cast'. (This
is the warning that points to the bug in `(char *)malloc(...)' where
malloc() is implicitly declared as returning int.  We do similar things
here, but they work because u_int is the same as uintptr_t on i386's.)
2003-11-17 02:11:13 +00:00
Alan Cox
0ec3db3072 - Remove unnecessary synchronization from sf_buf_init(). (There is only
one active CPU when sf_buf_init() is performed.)
2003-11-16 23:40:06 +00:00
Maxim Sobolev
d09c47acd9 Pull latest changes from OpenBSD:
- improve sysinfo(2) syscall;
- add dummy fadvise64(2) syscall;
- add dummy *xattr(2) family of syscalls;
- add protos for the syscalls 222-225, 238-249 and 253-267;
- add exit_group(2) syscall, which is currently just wired to exit(2).

Obtained from:  OpenBSD
MFC after:      2 weeks
2003-11-16 15:07:10 +00:00
Bruce Evans
095ab45bde Restored the call to schedsofttty() (now spelled swi_sched(...)) again.
Its restoration in rev.1.102 was mistranslated to the equivalent of
setsofttty() in rev.1.105.  This increased overheads by causing a
context switch to the SWI handler after almost every interrupt.  The
increase was approx. 50% on a Celeron 366 (from 23 usec to 34 usec
per interrupt).
2003-11-16 13:31:45 +00:00
Alan Cox
e45db9b837 - Modify alpha's sf_buf implementation to use the direct virtual-to-
physical mapping.
 - Move the sf_buf API to its own header file; make struct sf_buf's
   definition machine dependent.  In this commit, we remove an
   unnecessary field from struct sf_buf on the alpha, amd64, and ia64.
   Ultimately, we may eliminate struct sf_buf on those architecures
   except as an opaque pointer that references a vm page.
2003-11-16 06:11:26 +00:00
Bruce Evans
416ab90e6b Localized the cy driver's locking. 2003-11-16 00:55:54 +00:00
Peter Wemm
5333eee414 Minor source sync with amd64. Use int as the type for the width
field of %.*s rather than size_t.
2003-11-14 22:29:21 +00:00
Peter Wemm
ad641f0fd5 Minor source sync with amd64. For %.*s printf formats, pass in an
int rather than a size_t.  cast the ioapicaddress variable via
uintptr_t before going to void *.
2003-11-14 22:26:29 +00:00
Peter Wemm
7b66b81ee4 Convert a couple of pointers to integers for source compatability with
amd64.
2003-11-14 22:23:30 +00:00
Peter Wemm
40e3826a9f Whitespace nit (sorry, couldn't help it) 2003-11-14 22:21:30 +00:00
John Baldwin
64bb257f0b Always install IDT entries for ATPIC interrupt sources. The APIC no
longer uses these interrupt vectors for its ISA interrupt pins, so these
entries will not be overwritten.  If we get a spurious interrupt from the
ATPIC when using the APIC, it will be treated as a stray interrupt instead
of causing a panic.
2003-11-14 21:02:49 +00:00
John Baldwin
43d63d12fa If an interrupt source doesn't have an ithread, treat it as a stray
interrupt.  This can only happen if an unregistered interrupt source
triggers an interrupt.
2003-11-14 21:00:32 +00:00
Peter Wemm
db049820e1 basemem is in K, not bytes. I think I tricked jhb into making the same
mistake I did and then committing it to cvs.
2003-11-14 20:51:07 +00:00
Peter Wemm
2edfe38b10 "opt_auto_eoi.h" is not used here anymore. See atpic.c. 2003-11-14 20:06:24 +00:00
John Baldwin
f082493f10 Replace magic numbers with macros for i8259A register constants. Still
need the ICW4 bits for PC98 though.
2003-11-14 19:13:06 +00:00
John Baldwin
be06bd496f - Bring in constants for 8259A registers from amd64 with some updated
comments from NetBSD's dev/ic/i8259A.h.  These bits really belong in
  a file of the same name as well, but this will do for now.
- Axe unused HWI_MASK.
2003-11-14 19:12:25 +00:00
John Baldwin
3ab2ba59f4 Shuffle the APIC interrupt vectors around a bit:
- Move the IPI and local APIC interrupt vectors up into the 0xf0 - 0xff
  range.  The pmap lazyfix IPI was reordered down next to the TLB
  shootdowns to avoid conflicting with the spurious interrupt vector.
- Move the base of APIC interrupts up 16 so that the first 16 APIC
  interrupts do not overlap the vectors used by the ATPIC.
- Remove bogus interrupt vector reservations for LINT[01].
- Now that 0xc0 - 0xef are available, use them for device interrupts.
  This increases the number of APIC device interrupts to 191.
- Increase the system-wide number of global interrupts to 191 to catch up
  to more APIC interrupts.

Requested by:	peter (2)
2003-11-14 19:10:13 +00:00
John Baldwin
274461c944 Fix a typo. We need opt_acpi.h not opt_apic.h for DEV_ACPI. 2003-11-14 19:00:02 +00:00
John Baldwin
dced4b0c31 Whitespace. 2003-11-13 18:16:37 +00:00
John Baldwin
69487322d8 Fix a typo. 2003-11-13 16:41:07 +00:00
Poul-Henning Kamp
daf9092bcd Don't disable the TSC with statclock_disable. 2003-11-13 10:02:12 +00:00
John Baldwin
0926d72773 Garbage collect unused values. 2003-11-12 18:14:34 +00:00
John Baldwin
bd9cd7e3f7 - Move manipulation of td_intr_nesting_level out of assembly interrupt
vector stubs and into the C functions they call.
- Move disabling and EOIing of interrupt sources out of PIC driver entry
  points and into intr_execute_handlers().  Intr_execute_handlers() only
  disables a source for an interrupt if it is a stray interrupt or has
  threaded handlers.  Sources with fast handlers no longer disable (mask)
  the source while executing the handlers.
- Move the setting of clkintr_pending into intr_execute_handlers() and set
  the variable for any interrupt source with a vector of 0.  (Should only
  be true for IRQ 0.)  This fixes clkintr_pending in the NO_MIXED_MODE
  case.
- Implement lapic_eoi() and use it to implement ioapic_eoi_source().
- Rename atpic_sched_ithd() to atpic_handle_intr() since it is used to
  handle all atpic interrupts and not just threaded ones.

Inspired by:	peter's changes to amd64 in p4 (1)
Requested by:	bde (2)
2003-11-12 18:13:57 +00:00
John Baldwin
fd7d14d30b Don't probe busses in the MP Table for the MP Table PCI bridge drivers
if the bus number doesn't correspond to a PCI bus in the MP Table.

Reported by:	jhay
2003-11-11 21:19:43 +00:00
John Baldwin
d3c01334ce Some motherboards like to remap the SCI (normally IRQ 9) up to a PCI
interrupt such as IRQ 22 or 19.  However, the ACPI BIOS still routes
interrupts from some PCI devices to the same intpin calling the pin
IRQ 22.  Thus, ACPI expects to address a single interrupt source via two
different names.  To work around this, if the SCI is remapped to a non-ISA
interrupt (i.e., greater than 15), then we use
acpi_OverrideInterruptLevel() function to tell ACPI to use IRQ 22 or 19
rather than IRQ 9 for the SCI.

Previously we would change IRQ 22 or 19's name to IRQ 9 when we encountered
such an Interrupt Source Override entry in the MADT which routed the SCI
properly but left PCI devices mapped to IRQ 22 or 19 w/o a routable
interrupt.

Tested by:	sos
2003-11-11 18:20:10 +00:00
John Baldwin
da17811e64 Enable HTT CPUs by default instead of halting them by default. Users
should now only have HTT CPUs if they have explicitly asked for them
either by enabling HyperThreading in the BIOS or by using the
MPTABLE_FORCE_HTT kernel option.
2003-11-11 17:16:15 +00:00
John Baldwin
f9dbba5c4e Disable probing of HTT CPUs by default for the MP Table case. HTT CPUs
should only be used if they are enabled in the BIOS.  Now that we support
enumerating CPUs using the ACPI MADT, any HTT machine using ACPI should
respect the BIOS setting.  For HTT machines with ACPI disabled in the
kernel, the MPTABLE_FORCE_HTT kernel option can be used to try to probe HTT
CPUs like have done in the past for the MP Table case.  This option should
only be enabled if HTT is enabled in the BIOS.
2003-11-11 17:14:26 +00:00
John Baldwin
d1ccde5545 - Remove empty rogue SMP hardware section.
- Add some additional comments about 'device apic' to note that it can be
  used in both UP and SMP kernels but is required for SMP kernels.
2003-11-11 15:52:31 +00:00
John Baldwin
7e42a90bf6 Use the same style of paragraph indention that the rest of NOTES uses in
the SMP section.
2003-11-11 15:49:37 +00:00
John Baldwin
20694d64a2 Axe rotted comment about MP Tables and PCI cards with built in bridges.
Now that we properly route PCI interrupts for the apic case, these cards
are no longer a problem.
2003-11-11 15:47:44 +00:00
John Baldwin
4e478d4521 MFamd64 (via P4, not in CVS yet):
- Use the static boot_address variable directly rather than passing it
  around to several functions.
- Clean up a couple of magic numbers.
2003-11-10 21:24:34 +00:00
John Baldwin
95020215db Bump APIC ID limits up to 32 since a machine with 16 CPUs will have APIC
IDs for the I/O APICs that are greater than 16.

Reported by:	John Cagle <john.cagle@hp.com>
2003-11-10 19:52:58 +00:00
John Baldwin
4c76854fad Update a comment related to SMP and describe the NO_MIXED_MODE kernel
option.

Requested by:	bde
2003-11-10 15:54:32 +00:00
John Baldwin
c68ec2c20b Update a comment.
Requested by:	bde
2003-11-10 15:48:30 +00:00
Marcel Moolenaar
fcaa2925a9 Change the clear_ret argument of get_mcontext() to be a flags argument.
Since all callers either passed 0 or 1 for clear_ret, define bit 0 in
the flags for use as clear_ret. Reserve bits 1, 2 and 3 for use by MI
code for possible (but unlikely) future use. The remaining bits are for
use by MD code.

This change is triggered by a need on ia64 to have another knob for
get_mcontext().
2003-11-09 20:31:04 +00:00
Seigo Tanimura
512824f8f7 - Implement selwakeuppri() which allows raising the priority of a
thread being waken up.  The thread waken up can run at a priority as
  high as after tsleep().

- Replace selwakeup()s with selwakeuppri()s and pass appropriate
  priorities.

- Add cv_broadcastpri() which raises the priority of the broadcast
  threads.  Used by selwakeuppri() if collision occurs.

Not objected in:	-arch, -current
2003-11-09 09:17:26 +00:00
Alan Cox
5e4a2fc9aa - Similar to post-PAE RELENG_4 split pmap_pte_quick() into two cases,
pmap_pte() and pmap_pte_quick().  The distinction being based upon the
   locks that are held by the caller.  When the given pmap is not the
   current pmap, pmap_pte() should be used when Giant is held and
   pmap_pte_quick() should be used when the vm page queues lock is held.
 - When assigning to PMAP1 or PMAP2, include PG_A anf PG_M.
 - Reenable the inlining of pmap_is_current().

In collaboration with:	tegge
2003-11-08 03:01:26 +00:00