freebsd-skq/sys/amd64/include
John Baldwin 215e7c161a Rework how we wire up interrupt sources to CPUs:
- Throw out all of the logical APIC ID stuff.  The Intel docs are somewhat
  ambiguous, but it seems that the "flat" cluster model we are currently
  using is only supported on Pentium and P6 family CPUs.  The other
  "hierarchy" cluster model that is supported on all Intel CPUs with
  local APICs is severely underdocumented.  For example, it's not clear
  if the OS needs to glean the topology of the APIC hierarchy from
  somewhere (neither ACPI nor MP Table include it) and setup the logical
  clusters based on the physical hierarchy or not.  Not only that, but on
  certain Intel chipsets, even though there were 4 CPUs in a logical
  cluster, all the interrupts were only sent to one CPU anyway.
- We now bind interrupts to individual CPUs using physical addressing via
  the local APIC IDs.  This code has also moved out of the ioapic PIC
  driver and into the common interrupt source code so that it can be
  shared with MSI interrupt sources since MSI is addressed to APICs the
  same way that I/O APIC pins are.
- Interrupt source classes grow a new method pic_assign_cpu() to bind an
  interrupt source to a specific local APIC ID.
- The SMP code now tells the interrupt code which CPUs are avaiable to
  handle interrupts in a simpler and more intuitive manner.  For one thing,
  it means we could now choose to not route interrupts to HT cores if we
  wanted to (this code is currently in place in fact, but under an #if 0
  for now).
- For now we simply do static round-robin of IRQs to CPUs when the first
  interrupt handler just as before, with the change that IRQs are now
  bound to individual CPUs rather than groups of up to 4 CPUs.
- Because the IRQ to CPU mapping has now been moved up a layer, it would
  be easier to manage this mapping from higher levels.  For example, we
  could allow drivers to specify a CPU affinity map for their interrupts,
  or we could allow a userland tool to bind IRQs to specific CPUs.

The MFC is tentative, but I want to see if this fixes problems some folks
had with UP APIC kernels on 6.0 on SMP machines (an SMP kernel would work
fine, but a UP APIC kernel (such as GENERIC in RELENG_6) would lose
interrupts).

MFC after:	1 week
2006-02-28 22:24:55 +00:00
..
pc Make the facility for recognizing BIOS-signatures more general 2005-07-21 09:48:37 +00:00
_bus.h Break out the definition of bus_space_{tag,handle}_t and a few other types 2005-04-18 21:45:34 +00:00
_inttypes.h
_limits.h Move MINSIGSTKSZ from <machine/signal.h> to <machine/_limits.h> and rename 2005-08-20 16:44:41 +00:00
_stdint.h <stdint.h> should define WINT_M{AX,IN} independent from whether WCHAR_MIN is 2004-05-18 16:04:57 +00:00
_types.h By popular demand, move __HAVE_ACPI and __PCI_REROUTE_INTERRUPT into 2006-01-09 06:05:57 +00:00
acpica_machdep.h Move the code for halting the CPU (acpi_cpu_c1) into machdep files. 2004-10-11 05:39:15 +00:00
apicreg.h JumboMFi386: use bitmapped IPI handler. Update elcr and default mptable 2005-01-21 06:01:20 +00:00
apicvar.h Rework how we wire up interrupt sources to CPUs: 2006-02-28 22:24:55 +00:00
asm.h Align the entry point to assembly language functions to a 16-byte boundary. 2005-04-10 20:49:21 +00:00
asmacros.h MFi386: 2005-12-08 18:33:30 +00:00
atomic.h Add a new atomic_fetchadd() primitive that atomically adds a value to a 2005-09-27 17:39:11 +00:00
bus_dma.h Fix $FreeBSD$. 2005-07-22 04:03:25 +00:00
bus.h Remove bus_{mem,p}io.h and related code for a micro-optimization on i386 2005-05-29 04:42:30 +00:00
clock.h Tweak how the MD code calls the fooclock() methods some. Instead of 2005-12-22 22:16:09 +00:00
cpu.h Tweak how the MD code calls the fooclock() methods some. Instead of 2005-12-22 22:16:09 +00:00
cpufunc.h Add a knob for disabling/enabling HTT, "machdep.hyperthreading_allowed". 2005-05-13 00:10:56 +00:00
cputypes.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
db_machdep.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
elf.h Add style(9) foolishness. 2004-08-03 08:21:48 +00:00
endian.h Match i386 rev 1.38 with __cplusplus support 2005-03-11 21:46:01 +00:00
exec.h Use a common multi-inclusion protection, and add such a 2005-02-19 21:16:48 +00:00
float.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
floatingpoint.h The file machine/ieeefp.h needs sys/cdefs.h on amd64 and i386 after my 2005-04-02 17:31:42 +00:00
fpu.h Remove advertising clause from University of California Regent's license, 2004-04-05 21:29:41 +00:00
frame.h Tweak how the MD code calls the fooclock() methods some. Instead of 2005-12-22 22:16:09 +00:00
gdb_machdep.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
ieeefp.h The memory operands to fldcw and ldmxcsr are inputs, not outputs. 2005-04-12 23:12:00 +00:00
in_cksum.h Remove diffs to i386 version that came in via the compiler support ifdefs. 2005-03-11 22:16:09 +00:00
intr_machdep.h Rework how we wire up interrupt sources to CPUs: 2006-02-28 22:24:55 +00:00
iodev.h Break out the MI part of the /dev/[k]mem and /dev/io drivers into 2004-08-01 11:40:54 +00:00
kdb.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
legacyvar.h MFi386: pci attribute allocation fixes. 2005-09-18 01:42:43 +00:00
limits.h netchild's mega-patch to isolate compiler dependencies into a central 2005-03-02 21:33:29 +00:00
md_var.h - Print number of physical/logical cores and more CPUID info. 2005-10-14 22:52:01 +00:00
memdev.h Break out the MI part of the /dev/[k]mem and /dev/io drivers into 2004-08-01 11:40:54 +00:00
metadata.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
mp_watchdog.h Add an "options MP_WATCHDOG" to i386. This option allows one of the 2004-08-15 18:02:09 +00:00
mptable.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
mutex.h Kill the ASM versions of the mtx_lock_spin and friends. They were never 2003-12-06 23:17:18 +00:00
param.h By popular demand, move __HAVE_ACPI and __PCI_REROUTE_INTERRUPT into 2006-01-09 06:05:57 +00:00
pcb_ext.h
pcb.h I believe the stack underflows during early development that caused me to 2005-09-27 21:11:35 +00:00
pci_cfgreg.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
pcpu.h Remove diffs to i386 version that came in via the compiler support ifdefs. 2005-03-11 22:16:09 +00:00
pmap.h - Cleanup whitespace and extra ()s in vtophys() macros. 2005-12-06 21:09:01 +00:00
pmc_mdep.h MFP4: 2005-06-09 19:45:09 +00:00
ppireg.h - Move timerreg.h to <arch>/include and split i8253 specific defines into 2005-05-14 09:10:02 +00:00
proc.h Divorce critical sections from spinlocks. Critical sections as denoted by 2005-04-04 21:53:56 +00:00
profile.h Remove diffs to i386 version that came in via the compiler support ifdefs. 2005-03-11 22:16:09 +00:00
psl.h Remove advertising clause from University of California Regent's license, 2004-04-05 21:29:41 +00:00
ptrace.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
reg.h Remove advertising clause from University of California Regent's license, 2004-04-05 23:55:14 +00:00
reloc.h Remove advertising clause from University of California Regent's license, 2004-04-05 21:29:41 +00:00
resource.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
runq.h Begin all license/copyright comments with /*- 2005-01-05 20:17:21 +00:00
segments.h Remove advertising clause from University of California Regent's license, 2004-04-05 21:29:41 +00:00
setjmp.h
sf_buf.h MFi386: Update alc's copyright notice 2005-02-28 23:38:15 +00:00
sigframe.h
signal.h Move MINSIGSTKSZ from <machine/signal.h> to <machine/_limits.h> and rename 2005-08-20 16:44:41 +00:00
smp.h Tweak how the MD code calls the fooclock() methods some. Instead of 2005-12-22 22:16:09 +00:00
specialreg.h Correct few MSR addresses. 2005-10-15 00:44:56 +00:00
stdarg.h Remove diffs to i386 version that came in via the compiler support ifdefs. 2005-03-11 22:16:09 +00:00
sysarch.h Implement 32-bit compatable fsbase/gsbase methods so that we can run 2005-04-14 16:57:58 +00:00
timerreg.h Change the spkr_set_pitch() function to a macro to fix low level profiling. 2005-05-28 13:40:27 +00:00
trap.h Remove advertising clause from University of California Regent's license, 2004-04-05 21:29:41 +00:00
tss.h Show that I can actually count. 2005-04-15 18:39:31 +00:00
ucontext.h Update the graffiti. 2003-11-08 04:39:22 +00:00
varargs.h Remove diffs to i386 version that came in via the compiler support ifdefs. 2005-03-11 22:16:09 +00:00
vmparam.h Raise MAXDSIZ from 8G to 32G. The old limit was just an arbitary choice 2004-10-27 17:21:15 +00:00