constructed L1 PT page, this saves us from having to copy the kernel
L1 PTEs into the user L1 PT page at fork time (it's already set up).
A simple test shows a 1 second improvement of a rapid fork/exit operation
10000 times on a 533MHz 21164A (12s to 11s).
multiprocessor support:
- Implement MP-safe halt.
- Make the FPU saving code more like Bill's on the i386 MP branch.
XXX This code will no doubt be revisited again.
- Pass the cpu_info and trapframe to IPI handlers, saving some work
in the handlers themselves, and also making it possible for the
"pause" handler to reference register state for DDB.
- Add "machine cpu" to DDB, making it possible to reference other
CPUs registers (and thus get e.g. a traceback) from whichever
CPU is actually running the debugger.
- Garbage-collect "machine halt" and "machine reboot" DDB commands.
They don't have a prayer of working properly in multiprocessor
kernels, and didn't really work all that well in uniprocessor kernels.
* move all exec-type specific information from struct emul to execsw[] and
provide single struct emul per emulation
* elf:
- kern/exec_elf32.c:probe_funcs[] is gone, execsw[] how has one entry
per emulation and contains pointer to respective probe function
- interp is allocated via MALLOC() rather than on stack
- elf_args structure is allocated via MALLOC() rather than malloc()
* ecoff: the per-emulation hooks moved from alpha and mips specific code
to OSF1 and Ultrix compat code as appropriate, execsw[] has one entry per
emulation supporting ecoff with appropriate probe function
* the makecmds/probe functions don't set emulation, pointer to emulation is
part of appropriate execsw[] entry
* constify couple of structures
as normal device interrupts. Because of this, we won't get IPIs while
servicing such interrupts. This can lead to the following deadlock
scenario as reported by Bill Sommerfeld:
- Process runs on cpu1, but has FP state on cpu0.
- Process executes FP-using insn, causing an FP trap, which causes
the kernel lock to be acquired.
- At roughly the same time, cpu0 receives a device interrupt, and attempts
to acquire the kernel lock, which blocks since cpu1 already has it.
- cpu1 sends cpu0 a SYNCH FPU IPI, and waits for cpu0 to release its
FP state.
- Since cpu0 cannot notice the IPI until it has processed the device
interrupt, which it cannot do because it cannot acquire the kernel
lock, we have deadlock.
Solve the problem by adding a spinlock interlock release hook which
checks for pending IPIs and processes them.
Idea from Bill Sommerfeld.
rather than assigning to the whole field, set or clear individual flags,
which implies that the B_BUSY and B_INVAL flags will remain set.
this allows us to make the assertion in brelse() that B_BUSY is set,
which is the purpose of all this.
CPU to handle interrupts, but prevents a whole slew of things
from working properly on the second CPU. It is no longer
necessary now that several other bugs have been fixed.
on it, and make sure to deactivate the old process's address space,
as this will not otherwise be done by cpu_switch() later once we're
on the idle PCB.
Also, update some comments related to when s0 (i.e. the old process)
is NULL in the context switching code.
Drochner's work made in nisimura-pmax-wscons branch. Still a little
to do before useful for DECstation. MI softintr required.
- allow wildcard matchs to have zstty/zskbd/zsms in default hardware
configuration.
- abandon to check zs_ioasic_cnattach() return value; it doesn't fail.
- have zs_ioasic_cnattach() serial line parameters hardcoded inside, in
symmetry with zs_ioasic_lk201_cnattach().
pseudo-device pty 2 # pseudo-terminals (Sysinst needs two)
(Some installers may not be using sysinst, in which case this just reduces
the number of ptys from 16 that are not used to 2 that are not used)
For i386 conf files, no change other than comments.
Check the first partition type in devopen(), and if it is of type
FS_RAID, add 64 to blkdev_part_offset.
NOTE: This brings the size of the alpha first-stage bootblocks up to
close to the maximum. RAID1 support is controlled by the
BOOTXX_RAID1_SUPPORT define, and is easy to disable if size
becomes an issue.
based on it working already for macppc.
Also add commented out:
#options VNODE_OP_NOINLINE # Don't inline vnode op calls
#options NFS_V2_ONLY # Exclude NFS3 and NQNFS code
as suggestions for additional savings
maps standard boot flags to corresponding RB_* values
use BOOT_FLAG() in port's MD code as appropriate
as discussed on tech-kern, add new boot flags -v, -q for booting
verbosely or quietly, and corresponding AB_VERBOSE/AB_QUIET
boot flags; also add FreeBSD-compatible bootverbose macro and
NetBSD-specific bootquiet macro
for hpcmips, use new bootverbose instead of it's own hpcmips_verbose
Tested on i386, and to limited extend (compile of affected files) also for
mvme68k, hp300, luna68k, sun3.
indeed tagged with ASM and the ASN, "per the Alpha architecture".
It is therefore safe to cancel any pending lazy I-sync on a
given CPU when a new ASN is assigned on that CPU.
w/ Chuq Silvers. Fixes a panic when a program with wired pages that
has run for a long time when the system is under heavy memory load
exits (specific case was ntpd, reported by Simon Burge).
routine. Works similarly fto pmap_prefer(), but allows callers
to specify a minimum power-of-two alignment of the region.
How we ever got along without this for so long is beyond me.
in the non-MULTIPROCESSOR case (LOCKDEBUG requires it). Scheduler
lock is held upon entry to mi_switch() and cpu_switch(), and
cpu_switch() releases the lock before returning.
Largely from Bill Sommerfeld, with some minor bug fixes and
machine-dependent code hacking from me.
I-sync in pmap_remove_mapping() if the old mapping had PG_EXEC, and
kick curcpu (IMB) or other CPUs (via an IPI) only if the pmap was
the kernel pmap or active on other CPUs (curcpu is handled in userret()).
- Use lazy I-sync everywhere, (hopefully) eliminating the last of the
I-sync issues for multiprocessor support.
- Eliminate some memory barriers added in a couple of previous revisions,
after some discussion on port-alpha/tech-smp.
Still some lazy I-sync optimization possibilites:
- pmap_changebit() does not need to I-sync when only write-protecting
a page.
- pmap_asn_alloc() may be able to cancel a pending lazy I-sync when a
new ASN is allocated. Need to double check against Green Book or
Brown Book.
- pmap_zero_page() and pmap_copy_page(): if MULTIPROCESSOR, issue
a memory barrier after we zero/copy the page, to ensure that
other CPUs see the correct data.
- XXX Should we use MB, or is WMB good enough?
Also, bzero -> memset, bcopy -> memcpy.
- Make sure to do an MB after a PTE is set to a new value, so that
other processors see it.
- Use lazy I-sync in two pmap_page_protect() and in pmap_changebit(),
so that it is MP-safe. XXX Two more places where IMB is used in
the raw, but they're not in the common path.
There'll be some more lazy I-sync cleanups soon.
- MB_LEN_MAX is increased to 32.
- To ensure binary compatibility for old executables
under multibyte locale, versioned setlocale is added.
- __mb_len_cur definision is added in setlocale.c
and enable it in stdlib.h .
It is also important for multibyte locale stuffs,
but I just forgot.
move the vers.c depend/build goo to Makefile.bootprogs and remove
explicit rules in other Makefiles
sync the message in */version files with other ports using newvers_stand.sh
XXX the new depend rules were tested to limited extend (also with obj dirs)
XXX on i386 and should be ok; the changes should not otherwise influence build
have an ISA chipset present before the PCI-EISA bridge has been
attached (because the STDIO module has an ISA DMA-using device,
the floppy controller, connected to it).
device and the id string is not found in option ROM string, but
ULTRIX/OSF1 call it so anyway.
- Squash IOASIC DMA buffer allocation into device attach() sequence.
* put #includes of opt headers and headers to get protos used by
net/netisr_dispatch.h in net/netisr.h (if !defined(_LOCORE)) (rather than
in netisr_dispatch.h itself, and potentially nowhere, respectively).
* require netisr.h to be included before netisr_dispatch.h.
* minor additional cleanup of both netisr.h and netisr_dispatch.h.
* clean up uses to remove now-unnecessary header file inclusions, and
local prototypes of the fns.
* convert netisr dispatch implementations which didn't use
netisr_dispatch.h (pc532) to use it.
<vm/pglist.h> -> <uvm/uvm_pglist.h>
<vm/vm_inherit.h> -> <uvm/uvm_inherit.h>
<vm/vm_kern.h> -> into <uvm/uvm_extern.h>
<vm/vm_object.h> -> nothing
<vm/vm_pager.h> -> into <uvm/uvm_pager.h>
also includes a bunch of <vm/vm_page.h> include removals (due to redudancy
with <vm/vm.h>), and a scattering of other similar headers.
"off_t" and the return value is a "paddr_t" to allow mappings
at offsets past 2^31 bytes. Somewhat inspired by FreeBSD, which
only changed the offset to a "vm_offset_t".
Includes updates for the i386, pc532 and sh3 mmmmap from Jason Thorpe.
the `get window' method ends up with the wrong physical address to
pass onto userspace (which wants to mmap the space).
Compensate by adding a CHIP_PHYSADDR() macro which un-hacks the address
suitably for mapping with other-than-KSEG.
support module, make consistent with other systems which options
you must enable for it to work. Treat ``Lynx'' (2100A_A500) like
``Sable'' (2100_A500).
"return (v0);" where v0 is an unsigned long and the function return
type is int. Stops lint warnings when this file is included.
Fix from Jason Thorpe.
and scsipi_device respectively, with size reduction of ncr53c9x_softc.
Specifying NULL instructs the driver to use default adapter and default
device codes. Every target port has ncr53c9x_attach(sc, NULL, NULL) anyway.
macro in the MULTIPROCESSOR case (hardclock() wants it).
- Implement __GENERIC_SOFT_INTERRUPTS, and redefine the legacy
software interrupts in terms of it. Garbage-collect setsoftserial().
marked SONPROC.
- Always make curproc, fpcurproc, astpending, and want_resched per-CPU
variables in struct cpu_info. Restructure code accordingly, and trim
a few instructions from a few spots in various places in locore.
doing a cpu_set_kpc(), just pass the entry point and argument all
the way down the fork path starting with fork1(). In order to
avoid special-casing the normal fork in every cpu_fork(), MI code
passes down child_return() and the child process pointer explicitly.
This fixes a race condition on multiprocessor systems; a CPU could
grab the newly created processes (which has been placed on a run queue)
before cpu_set_kpc() would be performed.
- Change ktrace interface to pass in the current process, rather than
p->p_tracep, since the various ktr* function need curproc anyway.
- Add curproc as a parameter to mi_switch() since all callers had it
handy anyway.
- Add a second proc argument for inferior() since callers all had
curproc handy.
Also, miscellaneous cleanups in ktrace:
- ktrace now always uses file-based, rather than vnode-based I/O
(simplifies, increases type safety); eliminate KTRFLAG_FD & KTRFAC_FD.
Do non-blocking I/O, and yield a finite number of times when receiving
EWOULDBLOCK before giving up.
- move code duplicated between sys_fktrace and sys_ktrace into ktrace_common.
- simplify interface to ktrwrite()
state into global and per-CPU scheduler state:
- Global state: sched_qs (run queues), sched_whichqs (bitmap
of non-empty run queues), sched_slpque (sleep queues).
NOTE: These may collectively move into a struct schedstate
at some point in the future.
- Per-CPU state, struct schedstate_percpu: spc_runtime
(time process on this CPU started running), spc_flags
(replaces struct proc's p_schedflags), and
spc_curpriority (usrpri of processes on this CPU).
- Every platform must now supply a struct cpu_info and
a curcpu() macro. Simplify existing cpu_info declarations
where appropriate.
- All references to per-CPU scheduler state now made through
curcpu(). NOTE: this will likely be adjusted in the future
after further changes to struct proc are made.
Tested on i386 and Alpha. Changes are mostly mechanical, but apologies
in advance if it doesn't compile on a particular platform.
from db_stack_trace_cmd() to db_stack_trace_print(),
and add an additional argument, a function pointer for an
output routine (i.e. printf() or db_printf()).
Add db_stack_trace_cmd() in db_command.[ch], calling
db_stack_trace_print() with db_printf() as the printer.
Move count==-1 special handling from db_stack_trace_print() [nee
db_stack_trace_cmd()] to db_stack_trace_cmd() [nascent here].
Again, I'm unable to test compilation on all affected platforms,
so advance apologies for potential brokenness.
which indicates that the process is actually running on a
processor. Test against SONPROC as appropriate rather than
combinations of SRUN and curproc. Update all context switch code
to properly set SONPROC when the process becomes the current
process on the CPU.
pointer indicating how to print the symbol. This allows db_printsym()
to called in places where db_printf() is not an appropriate output
function.
While straightforward, apologies in advance if I've introduced any minor
syntax errors; I was unable to test compilation this on all the affected
platforms.
change these from bp->b_un.b_addr to bp->b_data, as well. This also
allows us more flexibility to experiment with other data buffer types
hung off of struct buf.
it to determine the boot device: mvme68k, pc532, macppc, ofppc. Those
platforms should be changed to use device_register(). In the mean time,
those ports defined __BROKEN_DK_ESTABLISH.
contains the values __SIMPLELOCK_LOCKED and __SIMPLELOCK_UNLOCKED, which
replace the old SIMPLELOCK_LOCKED and SIMPLELOCK_UNLOCKED. These files
are also required to supply inline functions __cpu_simple_lock(),
__cpu_simple_lock_try(), and __cpu_simple_unlock() if locking is to be
supported on that platform (i.e. if MULTIPROCESSOR is defined in the
_KERNEL case). Change these functions to take an int * (&alp->lock_data)
rather than the struct simplelock * itself.
These changes make it possible for userland to use the locking primitives
by including <machine/lock.h>.
to be called `rl' and it conflict with RL vax disks, canonical and
historical unix driver name.
This changes are minimal: it only changes the name of RealTek driver.
The filename of source code and a lot of the letter `rl' in source
files should be changed shortly.
uvm_page_init() has completed, add a boolean uvm.page_init_done,
and test against that. Use this same boolean (rather than
pmap_initialized) in pmap_growkernel() to determine if we are
being called via uvm_page_init() to grow the kernel address space.
This fixes a problem on some i386 configurations where pmap_init()
itself was needing to have the kernel page table grown, and since
pmap_initialized was not yet set to TRUE, pmap_growkernel() was
choosing the wrong code path.
Fix tested by Havard Eidnes.
an extern for 'struct platform platform' further down this file anyway,
and this would have declared a 'platform' variable in any userland code
including this file.
* Remove the casts to vaddr_t from the round_page() and trunc_page() macros to
make them type-generic, which is necessary i.e. to operate on file offsets
without truncating them.
* In due course, cast pointer arguments to these macros to an appropriate
integral type (paddr_t, vaddr_t).
Originally done by Chuck Silvers, updated by myself.
timeout()/untimeout() API:
- Clients supply callout handle storage, thus eliminating problems of
resource allocation.
- Insertion and removal of callouts is constant time, important as
this facility is used quite a lot in the kernel.
The old timeout()/untimeout() API has been removed from the kernel.
mycroft