as it could be and can do with some more cleanup. Currently its under
options LAZY_SWITCH. What this does is avoid %cr3 reloads for short
context switches that do not involve another user process. ie: we can
take an interrupt, switch to a kthread and return to the user without
explicitly flushing the tlb. However, this isn't as exciting as it could
be, the interrupt overhead is still high and too much blocks on Giant
still. There are some debug sysctls, for stats and for an on/off switch.
The main problem with doing this has been "what if the process that you're
running on exits while we're borrowing its address space?" - in this case
we use an IPI to give it a kick when we're about to reclaim the pmap.
Its not compiled in unless you add the LAZY_SWITCH option. I want to fix a
few more things and get some more feedback before turning it on by default.
This is NOT a replacement for Bosko's lazy interrupt stuff. This was more
meant for the kthread case, while his was for interrupts. Mine helps a
little for interrupts, but his helps a lot more.
The stats are enabled with options SWTCH_OPTIM_STATS - this has been a
pseudo-option for years, I just added a bunch of stuff to it.
One non-trivial change was to select a new thread before calling
cpu_switch() in the first place. This allows us to catch the silly
case of doing a cpu_switch() to the current process. This happens
uncomfortably often. This simplifies a bit of the asm code in cpu_switch
(no longer have to call choosethread() in the middle). This has been
implemented on i386 and (thanks to jake) sparc64. The others will come
soon. This is actually seperate to the lazy switch stuff.
Glanced at by: jake, jhb
statclock based on profhz when profiling is enabled MD, since most platforms
don't use this anyway. This removes the need for statclock_process, whose
only purpose was to subdivide profhz, and gets the profiling clock running
outside of sched_lock on platforms that implement suswintr.
Also changed the interface for starting and stopping the profiling clock to
do just that, instead of changing the rate of statclock, since they can now
be separate.
Reviewed by: jhb, tmm
Tested on: i386, sparc64
- It actually works this time, honest!
- Fine grained TLB shootdowns for SMP on i386. IPI's are very expensive,
so try and optimize things where possible.
- Introduce ranged shootdowns that can be done as a single IPI.
- PG_G support for i386
- Specific-cpu targeted shootdowns. For example, there is no sense in
globally purging the TLB cache for where we are stealing a page from
the local unshared process on the local cpu. Use pm_active to track
this.
- Add some instrumentation for the tlb shootdown code.
- Rip out SMP code from <machine/cpufunc.h>
- Try and fix some very bogus PG_G and PG_PS interactions that were bad
enough to cause vm86 bios calls to break. vm86 depended on our existing
bugs and this was the cause of the VESA panics last time.
- Fix the silly one-line error that caused the 'panic: bad pte' last time.
- Fix a couple of other silly one-line errors that should have caused more
pain than they did.
Some more work is needed:
- pmap_{zero,copy}_page[_idle]. These can be done without IPI's if we
have a hook in cpu_switch.
- The IPI handlers need some cleanup. I have a bogus %ds load that can
be avoided.
- APTD handling is rather bogus and appears to be a large source of
global TLB IPI shootdowns for no really good reason.
I see speedups of between 1.5% and ~4% on buildworlds in a while 1 loop.
I expect to see a bigger difference when there is significant pageout
activity or the system otherwise has memory shortages.
I have backed out a few optimizations that I had been using over the last
few days in order to be a little more conservative. I'll revisit these
again over the next few days as the dust settles.
New option: DISABLE_PG_G - In case I missed something.
and cpu_critical_exit() and moves associated critical prototypes into their
own header file, <arch>/<arch>/critical.h, which is only included by the
three MI source files that need it.
Backout and re-apply improperly comitted syntactical cleanups made to files
that were still under active development. Backout improperly comitted program
structure changes that moved localized declarations to the top of two
procedures. Partially re-apply one of the program structure changes to
move 'mask' into an intermediate block rather then in three separate
sub-blocks to make the code more readable. Re-integrate bug fixes that Jake
made to the sparc64 code.
Note: In general, developers should not gratuitously move declarations out
of sub-blocks. They are where they are for reasons of structure, grouping,
readability, compiler-localizability, and to avoid developer-introduced bugs
similar to several found in recent years in the VFS and VM code.
Reviewed by: jake
disablement assumptions in kern_fork.c by adding another API call,
cpu_critical_fork_exit(). Cleanup the td_savecrit field by moving it
from MI to MD. Temporarily move cpu_critical*() from <arch>/include/cpufunc.h
to <arch>/<arch>/critical.c (stage-2 will clean this up).
Implement interrupt deferral for i386 that allows interrupts to remain
enabled inside critical sections. This also fixes an IPI interlock bug,
and requires uses of icu_lock to be enclosed in a true interrupt disablement.
This is the stage-1 commit. Stage-2 will occur after stage-1 has stabilized,
and will move cpu_critical*() into its own header file(s) + other things.
This commit may break non-i386 architectures in trivial ways. This should
be temporary.
Reviewed by: core
Approved by: core
There is some unresolved badness that has been eluding me, particularly
affecting uniprocessor kernels. Turning off PG_G helped (which is a bad
sign) but didn't solve it entirely. Userland programs still crashed.
enabled in critical sections and streamline critical_enter() and
critical_exit().
This commit allows an architecture to leave interrupts enabled inside
critical sections if it so wishes. Architectures that do not wish to do
this are not effected by this change.
This commit implements the feature for the I386 architecture and provides
a sysctl, debug.critical_mode, which defaults to 1 (use the feature). For
now you can turn the sysctl on and off at any time in order to test the
architectural changes or track down bugs.
This commit is just the first stage. Some areas of the code, specifically
the MACHINE_CRITICAL_ENTER #ifdef'd code, is strictly temporary and will
be cleaned up in the STAGE-2 commit when the critical_*() functions are
moved entirely into MD files.
The following changes have been made:
* critical_enter() and critical_exit() for I386 now simply increment
and decrement curthread->td_critnest. They no longer disable
hard interrupts. When critical_exit() decrements the counter to
0 it effectively calls a routine to deal with whatever interrupts
were deferred during the time the code was operating in a critical
section.
Other architectures are unaffected.
* fork_exit() has been conditionalized to remove MD assumptions for
the new code. Old code will still use the old MD assumptions
in regards to hard interrupt disablement. In STAGE-2 this will
be turned into a subroutine call into MD code rather then hardcoded
in MI code.
The new code places the burden of entering the critical section
in the trampoline code where it belongs.
* I386: interrupts are now enabled while we are in a critical section.
The interrupt vector code has been adjusted to deal with the fact.
If it detects that we are in a critical section it currently defers
the interrupt by adding the appropriate bit to an interrupt mask.
* In order to accomplish the deferral, icu_lock is required. This
is i386-specific. Thus icu_lock can only be obtained by mainline
i386 code while interrupts are hard disabled. This change has been
made.
* Because interrupts may or may not be hard disabled during a
context switch, cpu_switch() can no longer simply assume that
PSL_I will be in a consistent state. Therefore, it now saves and
restores eflags.
* FAST INTERRUPT PROVISION. Fast interrupts are currently deferred.
The intention is to eventually allow them to operate either while
we are in a critical section or, if we are able to restrict the
use of sched_lock, while we are not holding the sched_lock.
* ICU and APIC vector assembly for I386 cleaned up. The ICU code
has been cleaned up to match the APIC code in regards to format
and macro availability. Additionally, the code has been adjusted
to deal with deferred interrupts.
* Deferred interrupts use a per-cpu boolean int_pending, and
masks ipending, spending, and fpending. Being per-cpu variables
it is not currently necessary to lock; bus cycles modifying them.
Note that the same mechanism will enable preemption to be
incorporated as a true software interrupt without having to
further hack up the critical nesting code.
* Note: the old critical_enter() code in kern/kern_switch.c is
currently #ifdef to be compatible with both the old and new
methodology. In STAGE-2 it will be moved entirely to MD code.
Performance issues:
One of the purposes of this commit is to enhance critical section
performance, specifically to greatly reduce bus overhead to allow
the critical section code to be used to protect per-cpu caches.
These caches, such as Jeff's slab allocator work, can potentially
operate very quickly making the effective savings of the new
critical section code's performance very significant.
The second purpose of this commit is to allow architectures to
enable certain interrupts while in a critical section. Specifically,
the intention is to eventually allow certain FAST interrupts to
operate rather then defer.
The third purpose of this commit is to begin to clean up the
critical_enter()/critical_exit()/cpu_critical_enter()/
cpu_critical_exit() API which currently has serious cross pollution
in MI code (in fork_exit() and ast() for example).
The fourth purpose of this commit is to provide a framework that
allows kernel-preempting software interrupts to be implemented
cleanly. This is currently used for two forward interrupts in I386.
Other architectures will have the choice of using this infrastructure
or building the functionality directly into critical_enter()/
critical_exit().
Finally, this commit is designed to greatly improve the flexibility
of various architectures to manage critical section handling,
software interrupts, preemption, and other highly integrated
architecture-specific details.
on for a while:
- fine grained TLB shootdown for SMP on i386
- ranged TLB shootdowns.. eg: specify a range of pages to shoot down with
a single IPI, since the IPI is very expensive. Adjust some callers
that used to trigger this inside tight loops to do a ranged shootdown
at the end instead.
- PG_G support for SMP on i386 (options ENABLE_PG_G)
- defer PG_G activation till after we decide what we are going to do with
PSE and the 4MB pages at the start of the kernel. This should solve
some rumored strangeness about stale PG_G entries getting stuck
underneath the 4MB pages.
- add some instrumentation for the fine TLB shootdown
- convert some asm instruction wrappers from functions to inlines. gcc
seems to do a fair bit better with this.
- [temporarily!] pessimize the tlb shootdown IPI handlers. I will fix
this again shortly.
This has been working fairly well for me for a while, but I have tweaked
it again prior to commit since my last major testing round. The only
outstanding problem that I know of is PG_G related, which is why there
is an option for it (not on by default for SMP). I have seen a world
speedups by a few percent (as much as 4 or 5% in one case) but I have
*not* accurately measured this - I am a bit sceptical of these numbers.
mutex releases to not require flags for the cases when preemption is
not allowed:
The purpose of the MTX_NOSWITCH and SWI_NOSWITCH flags is to prevent
switching to a higher priority thread on mutex releease and swi schedule,
respectively when that switch is not safe. Now that the critical section
API maintains a per-thread nesting count, the kernel can easily check
whether or not it should switch without relying on flags from the
programmer. This fixes a few bugs in that all current callers of
swi_sched() used SWI_NOSWITCH, when in fact, only the ones called from
fast interrupt handlers and the swi_sched of softclock needed this flag.
Note that to ensure that swi_sched()'s in clock and fast interrupt
handlers do not switch, these handlers have to be explicitly wrapped
in critical_enter/exit pairs. Presently, just wrapping the handlers is
sufficient, but in the future with the fully preemptive kernel, the
interrupt must be EOI'd before critical_exit() is called. (critical_exit()
can switch due to a deferred preemption in a fully preemptive kernel.)
I've tested the changes to the interrupt code on i386 and alpha. I have
not tested ia64, but the interrupt code is almost identical to the alpha
code, so I expect it will work fine. PowerPC and ARM do not yet have
interrupt code in the tree so they shouldn't be broken. Sparc64 is
broken, but that's been ok'd by jake and tmm who will be fixing the
interrupt code for sparc64 shortly.
Reviewed by: peter
Tested on: i386, alpha
and it's associated state variables: icu_lock with the name "icu". This
renames the imen_mtx for x86 SMP, but also uses the lock to protect
access to the 8259 PIC on x86 UP. This also adds an appropriate lock to
the various Alpha chipsets which fixes problems with Alpha SMP machines
dropping interrupts with an SMP kernel.
code in ipl.s and icu_ipl.s that used them was removed when the
interrupt thread system was committed. Debuggers also knew about
Xresume* because these labels hide the real names of the interrupt
handlers (Xintr*), and debuggers need to special-case interrupt
handlers to get the interrupt frame.
Both gdb and ddb will now use the Xintr* and Xfastintr* symbols to
detect interrupt frames. Fast interrupt frames were never identified
correctly before, so this fixes the problem of the running stack
frame getting lost in a ddb or gdb trace generated from a fast
interrupt - e.g. when debugging a simple infinite loop in the kernel
using a serial console, the frame containing the loop would never
appear in a gdb or ddb trace.
Reviewed by: jhb, bde
Note ALL MODULES MUST BE RECOMPILED
make the kernel aware that there are smaller units of scheduling than the
process. (but only allow one thread per process at this time).
This is functionally equivalent to teh previousl -current except
that there is a thread associated with each process.
Sorry john! (your next MFC will be a doosie!)
Reviewed by: peter@freebsd.org, dillon@freebsd.org
X-MFC after: ha ha ha ha
been made machine independent and various other adjustments have been made
to support Alpha SMP.
- It splits the per-process portions of hardclock() and statclock() off
into hardclock_process() and statclock_process() respectively. hardclock()
and statclock() call the *_process() functions for the current process so
that UP systems will run as before. For SMP systems, it is simply necessary
to ensure that all other processors execute the *_process() functions when the
main clock functions are triggered on one CPU by an interrupt. For the alpha
4100, clock interrupts are delievered in a staggered broadcast fashion, so
we simply call hardclock/statclock on the boot CPU and call the *_process()
functions on the secondaries. For x86, we call statclock and hardclock as
usual and then call forward_hardclock/statclock in the MD code to send an IPI
to cause the AP's to execute forwared_hardclock/statclock which then call the
*_process() functions.
- forward_signal() and forward_roundrobin() have been reworked to be MI and to
involve less hackery. Now the cpu doing the forward sets any flags, etc. and
sends a very simple IPI_AST to the other cpu(s). AST IPIs now just basically
return so that they can execute ast() and don't bother with setting the
astpending or needresched flags themselves. This also removes the loop in
forward_signal() as sched_lock closes the race condition that the loop worked
around.
- need_resched(), resched_wanted() and clear_resched() have been changed to take
a process to act on rather than assuming curproc so that they can be used to
implement forward_roundrobin() as described above.
- Various other SMP variables have been moved to a MI subr_smp.c and a new
header sys/smp.h declares MI SMP variables and API's. The IPI API's from
machine/ipl.h have moved to machine/smp.h which is included by sys/smp.h.
- The globaldata_register() and globaldata_find() functions as well as the
SLIST of globaldata structures has become MI and moved into subr_smp.c.
Also, the globaldata list is only available if SMP support is compiled in.
Reviewed by: jake, peter
Looked over by: eivind
and used in C or vice versa. The elf compiler uses the same names
for both. Remove asnames.h with great prejudice; it has served its
purpose.
Note that this does not affect the ability to generate an aout kernel
due to gcc's -mno-underscores option.
moral support from: peter, jhb
attributes. This is needed for AST's to be properly posted in a preemptive
kernel. They are backed by two new flags in p_sflag: PS_ASTPENDING and
PS_NEEDRESCHED. They are still accesssed by their old macros:
aston(), astoff(), etc. For completeness, an astpending() macro has been
added to check for a pending AST, and clear_resched() has been added to
clear need_resched().
- Rename syscall2() on the x86 back to syscall() to be consistent with
other architectures.
interrupt threads to run with it always >= 1, so that malloc can
detect M_WAITOK from "interrupt" context. This is also necessary
in order to context switch from sched_ithd() directly.
Reviewed By: peter
variables from i386 assembly language. The syntax is PCPU(member)
where member is the capitalized name of the per-cpu variable, without
the gd_ prefix. Example: movl %eax,PCPU(CURPROC). The capitalization
is due to using the offsets generated by genassym rather than the symbols
provided by linking with globals.o. asmacros.h is the wrong place for
this but it seemed as good a place as any for now. The old implementation
in asnames.h has not been removed because it is still used to de-mangle
the symbols used by the C variables for the UP case.
Also, while here, run up to 32 interrupt sources on APIC systems.
Normalize INTREN/INTRDIS so they are the same on both UP and SMP systems
rather than sometimes a macro, and sometimes a function.
Reviewed by: jhb, jakeb
return through doreti to handle ast's. This is necessary for the
clock interrupts to work properly.
- Change the clock interrupts on the x86 to be fast instead of threaded.
This is needed because both hardclock() and statclock() need to run in
the context of the current process, not in a separate thread context.
- Kill the prevproc hack as it is no longer needed.
- We really need Giant when we call psignal(), but we don't want to block
during the clock interrupt. Instead, use two p_flag's in the proc struct
to mark the current process as having a pending SIGVTALRM or a SIGPROF
and let them be delivered during ast() when hardclock() has finished
running.
- Remove CLKF_BASEPRI, which was #ifdef'd out on the x86 anyways. It was
broken on the x86 if it was turned on since cpl is gone. It's only use
was to bogusly run softclock() directly during hardclock() rather than
scheduling an SWI.
- Remove the COM_LOCK simplelock and replace it with a clock_lock spin
mutex. Since the spin mutex already handles disabling/restoring
interrupts appropriately, this also lets us axe all the *_intr() fu.
- Back out the hacks in the APIC_IO x86 cpu_initclocks() code to use
temporary fast interrupts for the APIC trial.
- Add two new process flags P_ALRMPEND and P_PROFPEND to mark the pending
signals in hardclock() that are to be delivered in ast().
Submitted by: jakeb (making statclock safe in a fast interrupt)
Submitted by: cp (concept of delaying signals until ast())
- Make softinterrupts (SWI's) almost completely MI, and divorce them
completely from the x86 hardware interrupt code.
- The ihandlers array is now gone. Instead, there is a MI shandlers array
that just contains SWI handlers.
- Most of the former machine/ipl.h files have moved to a new sys/ipl.h.
- Stub out all the spl*() functions on all architectures.
Submitted by: dfr
include:
* Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The
alpha port is still in transition and currently uses both.)
* Per-CPU idle processes.
* Interrupts are run in their own separate kernel threads and can be
preempted (i386 only).
Partially contributed by: BSDi (BSD/OS)
Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
have their own lock and do not need the MP lock. The SMP cleanup was
a little too conservative in MP locking fast interrupts but at least
it's trivial to fix. MFC soon.
Submitted by: bde
operands. `movw' could be used, but instead let the assembler decide
the right instruction to use.
2. AT&T asm syntax requires a leading '*' in front of the operand for
indirect calls and jumps.
removed unncessary MPLOCKED and 'lock' prefixes from the interrupt
nesting level, since (A) the MP lock is held at the time, and (B) since
the neting level is restored prior to return any interrupted code
will see a consistent value.
syscall path inward. A system call may select whether it needs the MP
lock or not (the default being that it does need it).
A great deal of conditional SMP code for various deadended experiments
has been removed. 'cil' and 'cml' have been removed entirely, and the
locking around the cpl has been removed. The conditional
separately-locked fast-interrupt code has been removed, meaning that
interrupts must hold the CPL now (but they pretty much had to anyway).
Another reason for doing this is that the original separate-lock for
interrupts just doesn't apply to the interrupt thread mechanism being
contemplated.
Modifications to the cpl may now ONLY occur while holding the MP
lock. For example, if an otherwise MP safe syscall needs to mess with
the cpl, it must hold the MP lock for the duration and must (as usual)
save/restore the cpl in a nested fashion.
This is precursor work for the real meat coming later: avoiding having
to hold the MP lock for common syscalls and I/O's and interrupt threads.
It is expected that the spl mechanisms and new interrupt threading
mechanisms will be able to run in tandem, allowing a slow piecemeal
transition to occur.
This patch should result in a moderate performance improvement due to
the considerable amount of code that has been removed from the critical
path, especially the simplification of the spl*() calls. The real
performance gains will come later.
Approved by: jkh
Reviewed by: current, bde (exception.s)
Some work taken from: luoqi's patch
to use a locked cmpexg when unlocking a lock that we already hold, since
nobody else can touch the lock while we hold it. Second, it is not
necessary to use a locked cmpexg when locking a lock that we already
hold, for the same reason. These changes will allow MP locks to be used
recursively without impacting performance.
Modify two procedures that are called only by assembly and are already
NOPROF entries to pass a critical argument in %edx instead of on the
stack, removing a significant amount of code from the critical path
as a consequence.
Reviewed by: Alfred Perlstein <bright@wintelcom.net>, Peter Wemm <peter@netplex.com.au>
the caller to specify a function to be guarded between an entry and exit
barrier, as well as pre- and post-barrier functions.
The primary use for this function is synchronised update of per-cpu private
data. The implementation is almost (but not quite) MI; with a better
mechanism for masking per-CPU interrupts it could probably be hoisted.
Reviewed by: peter (partially)
The old version only worked right when the time was read strictly
more often than every 1/HZ seconds, but we only guarantee reading
it every (1/HZ + epsilon) seconds. Part of rev.1.126-1.127 attempted
to fix this but didn't succeed. Detect counter rollover using the
heuristic from the old version of microtime() with additional
complications for supporting calls from fast interrupt handlers.
This works provided i8254 interrupts are not delayed by more than
1/(2*HZ) seconds.
This needs more comments, and cleanups for the SMP case, and more
testing of the SMP case before it is merged into RELENG_3.
Tested by: jhay
- %fs register is added to trapframe and saved/restored upon kernel entry/exit.
- Per-cpu pages are no longer mapped at the same virtual address.
- Each cpu now has a separate gdt selector table. A new segment selector
is added to point to per-cpu pages, per-cpu global variables are now
accessed through this new selector (%fs). The selectors in gdt table are
rearranged for cache line optimization.
- fask_vfork is now on as default for both UP and SMP.
- Some aio code cleanup.
Reviewed by: Alan Cox <alc@cs.rice.edu>
John Dyson <dyson@iquest.net>
Julian Elischer <julian@whistel.com>
Bruce Evans <bde@zeta.org.au>
David Greenman <dg@root.com>
