I/O port access is implemented on Itanium by reading and writing to a
special region in memory. To hide details and avoid misaligned memory
accesses, a process did I/O port reads and writes by making a MD system
call. There's one fatal problem with this approach: unprivileged access
was not being prevented. /dev/io serves that purpose on amd64/i386, so
employ it on ia64 as well. Use an ioctl for doing the actual I/O and
remove the sysarch(2) interface.
Backward compatibility is not being considered. The sysarch(2) approach
was added to support X11, but support for FreeBSD/ia64 was never fully
implemented in X11. Thus, nothing gets broken that didn't need more work
to begin with.
MFC after: 1 week
o Optimize for memory mapped I/O by making all I/O port acceses function
calls and marking the test for the IA64_BUS_SPACE_IO tag with
__predict_false(). Implement the I/O port access functions in a new
file, called bus_machdep.c.
o Change the bus_space_handle_t for memory mapped I/O to the virtual
address rather than the physical address. This eliminates the PA->VA
translation for every I/O access. The handle for I/O port access is
still the port number.
o Move inb(), outb(), inw(), outw(), inl(), outl(), and their string
variants from cpufunc.h and define them in bus.h. On ia64 these are
not CPU functions at all. In bus.h they are merely aliases for the
new I/O port access functions defined in bus_machdep.h.
o Handle the ACPI resource bug in nexus_set_resource(). There we can
do it once so that we don't have to worry about it whenever we need
to write to an I/O port that is really a memory mapped address.
The upshot of this change is that the KBI is better defined and that I/O
port access always involves a function call, allowing us to change the
actual implementation without breaking the KBI. For memory mapped I/O the
virtual address is abstracted, so that we can change the VA->PA mapping
in the kernel without causing an KBI breakage. The exception at this time
is for bus_space_map() and bus_space_unmap().
MFC after: 1 week.
The frequencies are in MHz (i.e. a value of 1000 represents 1GHz). The
frequencies are rounded to the nearest whole MHz.
While here, rename and re-type bus_frequency, processor_frequency and
itc_frequency to bus_freq, cpu_freq and itc_freq and make them static.
As unsigned integers, the hw.freq.cpu sysctl can more easily be made
generic (across all architectures) making porting easier.
MFC after: 3 days
excluded, as it's used by MI code) and mode the sysctl variables from
pcpu_stats to pcpu_md.
Adjust all references accordingly.
While nearby, change the PCPU sysctl tree so that they match the CPU
device sysctl tree -- they are now children of a static node called
"machdep.cpu" and are named only with their cpu ID.
allocating MAXCPU VHPTs up-front. This allows us to max-out MAXCPU
without memory waste -- MAXCPU is now 32 for SMP kernels.
This change also eliminates the VHPT scaling based in the total
memory in the system. It's the workload that determines the best size
of the VHPT. The workload can be affected by the amount of memory,
but not necessarily. For example, there's no performance difference
between VHPT sizes of 256KB, 512KB and 1MB when building the LINT
kernel. This was observed with a system that has 8GB of memory.
By default the kernel will allocate a 1MB VHPT. The user can tune the
system with the "machdep.vhpt.log2size" tunable.
Memory accesses are posted in program order by virtue of the
uncacheable memory attribute.
Since GCC, by default, adds acquire and release semantics to
volatile memory loads and stores, we need to use inline assembly
to guarantee it. With inline assembly, we don't need volatile
pointers anymore.
Itanium does not support semaphore instructions to uncacheable
memory.
adding statistics counters to the PCPU structure. Export the counters
through sysctl by giving each PCPU structure its own sysctl context.
While here, fix cnt.v_intr by not just having it count clock interrupts,
but every interrupt and add more counters for each interrupt source.
the kernel stack at all. The new USB stack simply caused a change
in timing that triggered a firmware bug more often. The addition
of PRINTF_BUFR_SIZE apparently triggered the same firmware bug
even more reliably.
But even with KSTACK_PAGES=5, one instance of the firmware bug
remained: booting with a CD inserted. This problem was run into
by accident after installing Debian and having to boot FreeBSD
to fixup the GPT partitioning (Thanks... not). After bumping
KSTACK_PAGES to 5, it was pretty unbelievable that the stack was
still being too small.
After updating the firmware we could boot with a CD inserted and
KSTACK_PAGES could be lowered back to 4 pages without problems.
Note: It is believed to be a timing related firmware bug, because
the machine check information showed access to the serial console
on one CPU and access to the EHCI HCD on the other CPU. Since
both are devices on the management unit and thus virtualized in
some way, any execution trace that does not include concurrent
access to the BMC from both CPUs is fine.
Note also that it's not understood exactly how increasing the
kernel stack avoided hitting the firmware bug. A change in page
faults does change timing, but it's not known if that's what's
happening here.
In any case: the problem is being monitored. Reverting back to
4 pages for the kernel stack is preferred, because it makes it
easier to switch to 16K pages (double the page size) without
wasting too much memory by not being able to half the number of
pages...
o Move all code into a single file for easier maintenance.
o Use a single global lock to avoid having to handle either
multiple locks or race conditions.
o Make sure to disable the high FP registers after saving
or dropping them.
o use msleep() to wait for the other CPU to save the high
FP registers.
This change fixes the high FP inconsistency panics.
A single global lock typically serializes too much, which may
be noticable when a lot of threads use the high FP registers,
but in that case it's probably better to switch the high FP
context synchronuously. Put differently: cpu_switch() should
switch the high FP registers if the incoming and outgoing
threads both use the high FP registers.
more stack hungry as compared to the old one that my RX2660 gets
a machine check and spontaneously reboots at the time the USB DVD
drive is found and attached to CAM as a mass storage device. This
doesn't happen always, but definitely varies per kernel build.
Likewise when using a 128-byte printf buffer. The additional 128
bytes that printf needs seems to be enough to have the memory stack
and register stack collide and causing a machine check.
Thus: Bump KSTACK_PAGES from 4 to 5.
by looking at the bases used for non-relocatable executables by gnu ld(1),
and adjusting it slightly.
Discussed with: bz
Reviewed by: kan
Tested by: bz (i386, amd64), bsam (linux)
MFC after: some time
has proven to have a good effect when entering KDB by using a NMI,
but it completely violates all the good rules about interrupts
disabled while holding a spinlock in other occasions. This can be the
cause of deadlocks on events where a normal IPI_STOP is expected.
* Adds an new IPI called IPI_STOP_HARD on all the supported architectures.
This IPI is responsible for sending a stop message among CPUs using a
privileged channel when disponible. In other cases it just does match a
normal IPI_STOP.
Right now the IPI_STOP_HARD functionality uses a NMI on ia32 and amd64
architectures, while on the other has a normal IPI_STOP effect. It is
responsibility of maintainers to eventually implement an hard stop
when necessary and possible.
* Use the new IPI facility in order to implement a new userend SMP kernel
function called stop_cpus_hard(). That is specular to stop_cpu() but
it does use the privileged channel for the stopping facility.
* Let KDB use the newly introduced function stop_cpus_hard() and leave
stop_cpus() for all the other cases
* Disable interrupts on CPU0 when starting the process of APs suspension.
* Style cleanup and comments adding
This patch should fix the reboot/shutdown deadlocks many users are
constantly reporting on mailing lists.
Please don't forget to update your config file with the STOP_NMI
option removal
Reviewed by: jhb
Tested by: pho, bz, rink
Approved by: re (kib)
dependent memory attributes:
Rename vm_cache_mode_t to vm_memattr_t. The new name reflects the
fact that there are machine-dependent memory attributes that have
nothing to do with controlling the cache's behavior.
Introduce vm_object_set_memattr() for setting the default memory
attributes that will be given to an object's pages.
Introduce and use pmap_page_{get,set}_memattr() for getting and
setting a page's machine-dependent memory attributes. Add full
support for these functions on amd64 and i386 and stubs for them on
the other architectures. The function pmap_page_set_memattr() is also
responsible for any other machine-dependent aspects of changing a
page's memory attributes, such as flushing the cache or updating the
direct map. The uses include kmem_alloc_contig(), vm_page_alloc(),
and the device pager:
kmem_alloc_contig() can now be used to allocate kernel memory with
non-default memory attributes on amd64 and i386.
vm_page_alloc() and the device pager will set the memory attributes
for the real or fictitious page according to the object's default
memory attributes.
Update the various pmap functions on amd64 and i386 that map pages to
incorporate each page's memory attributes in the mapping.
Notes: (1) Inherent to this design are safety features that prevent
the specification of inconsistent memory attributes by different
mappings on amd64 and i386. In addition, the device pager provides a
warning when a device driver creates a fictitious page with memory
attributes that are inconsistent with the real page that the
fictitious page is an alias for. (2) Storing the machine-dependent
memory attributes for amd64 and i386 as a dedicated "int" in "struct
md_page" represents a compromise between space efficiency and the ease
of MFCing these changes to RELENG_7.
In collaboration with: jhb
Approved by: re (kib)
o add to platforms where it was missing (arm, i386, powerpc, sparc64, sun4v)
o define as "1" on amd64 and i386 where there is no restriction
o make the type returned consistent with ALIGN
o remove _ALIGNED_POINTER
o make associated comments consistent
Reviewed by: bde, imp, marcel
Approved by: re (kensmith)
required by video card drivers. Specifically, this change introduces
vm_cache_mode_t with an appropriate VM_CACHE_DEFAULT definition on all
architectures. In addition, this changes adds a vm_cache_mode_t parameter
to kmem_alloc_contig() and vm_phys_alloc_contig(). These will be the
interfaces for allocating mapped kernel memory and physical memory,
respectively, with non-default cache modes.
In collaboration with: jhb
a fair number of static data structures, making this an unlikely
option to try to change without also changing source code. [1]
Change default cache line size on ia64, sparc64, and sun4v to 128
bytes, as this was what rtld-elf was already using on those
platforms. [2]
Suggested by: bde [1], jhb [2]
MFC after: 2 weeks
CACHE_LINE_SIZE constant. These constants are intended to
over-estimate the cache line size, and be used at compile-time
when a run-time tuning alternative isn't appropriate or
available.
Defaults for all architectures are 64 bytes, except powerpc
where it is 128 bytes (used on G5 systems).
MFC after: 2 weeks
Discussed on: arch@
to the full path of the image that is being executed.
Increase AT_COUNT.
Remove no longer true comment about types used in Linux ELF binaries,
listed types contain FreeBSD-specific entries.
Reviewed by: kan
while holding a spin mutex. Instead, it now shoves the machine check
records onto a queue that is later drained to add sysctl nodes for each
record. While a routine to drain the queue is present, it is not currently
called.
Reviewed by: marcel
and ifnet functions
- add memory barriers to <machine/atomic.h>
- update drivers to only conditionally define their own
- add lockless producer / consumer ring buffer
- remove ring buffer implementation from cxgb and update its callers
- add if_transmit(struct ifnet *ifp, struct mbuf *m) to ifnet to
allow drivers to efficiently manage multiple hardware queues
(i.e. not serialize all packets through one ifq)
- expose if_qflush to allow drivers to flush any driver managed queues
This work was supported by Bitgravity Inc. and Chelsio Inc.
all to date and the latter also is only used in ia64 and powerpc
code which no longer serves a real purpose after bring-up and just
can be removed as well. Note that architectures like sun4u also
provide no means of implementing IPI'ing a CPU itself natively
in the first place.
Suggested by: jhb
Reviewed by: arch, grehan, jhb
Use the new inline function in ia64_invalidate_icache().
While there, add proper synchronization so that we know
the fc.i instructions have taken effect when we return.
Now that st_rdev is being automatically generated by the kernel, there
is no need to define static major/minor numbers for the iodev and
memdev. We still need the minor numbers for the memdev, however, to
distinguish between /dev/mem and /dev/kmem.
Approved by: philip (mentor)
inlining resulted in constant propagation to the extend that cmpval
was known to the compiler to be URWLOCK_WRITE_OWNER (= 0x80000000U).
Unfortunately, instead of zero-extending the unsigned constant, it
was sign-extended. As such, the cmpxchg instruction was comparing
0x0000000080000000LU to 0xffffffff80000000LU and obviously didn't
perform the exchange.
But, since the value returned by cmpxhg equalled cmpval (when zero-
extended), the _thr_rtld_lock_release() function thought the exchange
did happen and as such returned as if having released the lock. This
was not the case. Subsequent locking requests found rw_state non-zero
and the thread in question entered the kernel and block indefinitely.
The work-around is to zero-extend by casting to uint64_t.
was a kluge. This implementation matches the behaviour on powerpc
and sparc64.
While on the subject, make sure to invalidate the I-cache after
loading a kernel module.
MFC after: 2 weeks
these days, so de-generalize the acquire_timer/release_timer api
to just deal with speakers.
The new (optional) MD functions are:
timer_spkr_acquire()
timer_spkr_release()
and
timer_spkr_setfreq()
the last of which configures the timer to generate a tone of a given
frequency, in Hz instead of 1/1193182th of seconds.
Drop entirely timer2 on pc98, it is not used anywhere at all.
Move sysbeep() to kern/tty_cons.c and use the timer_spkr*() if
they exist, and do nothing otherwise.
Remove prototypes and empty acquire-/release-timer() and sysbeep()
functions from the non-beeping archs.
This eliminate the need for the speaker driver to know about
i8254frequency at all. In theory this makes the speaker driver MI,
contingent on the timer_spkr_*() functions existing but the driver
does not know this yet and still attaches to the ISA bus.
Syscons is more tricky, in one function, sc_tone(), it knows the hz
and things are just fine.
In the other function, sc_bell() it seems to get the period from
the KDMKTONE ioctl in terms if 1/1193182th second, so we hardcode
the 1193182 and leave it at that. It's probably not important.
Change a few other sysbeep() uses which obviously knew that the
argument was in terms of i8254 frequency, and leave alone those
that look like people thought sysbeep() took frequency in hertz.
This eliminates the knowledge of i8254_freq from all but the actual
clock.c code and the prof_machdep.c on amd64 and i386, where I think
it would be smart to ask for help from the timecounters anyway [TBD].
While the KSE project was quite successful in bringing threading to
FreeBSD, the M:N approach taken by the kse library was never developed
to its full potential. Backwards compatibility will be provided via
libmap.conf for dynamically linked binaries and static binaries will
be broken.
ways:
(1) Cached pages are no longer kept in the object's resident page
splay tree and memq. Instead, they are kept in a separate per-object
splay tree of cached pages. However, access to this new per-object
splay tree is synchronized by the _free_ page queues lock, not to be
confused with the heavily contended page queues lock. Consequently, a
cached page can be reclaimed by vm_page_alloc(9) without acquiring the
object's lock or the page queues lock.
This solves a problem independently reported by tegge@ and Isilon.
Specifically, they observed the page daemon consuming a great deal of
CPU time because of pages bouncing back and forth between the cache
queue (PQ_CACHE) and the inactive queue (PQ_INACTIVE). The source of
this problem turned out to be a deadlock avoidance strategy employed
when selecting a cached page to reclaim in vm_page_select_cache().
However, the root cause was really that reclaiming a cached page
required the acquisition of an object lock while the page queues lock
was already held. Thus, this change addresses the problem at its
root, by eliminating the need to acquire the object's lock.
Moreover, keeping cached pages in the object's primary splay tree and
memq was, in effect, optimizing for the uncommon case. Cached pages
are reclaimed far, far more often than they are reactivated. Instead,
this change makes reclamation cheaper, especially in terms of
synchronization overhead, and reactivation more expensive, because
reactivated pages will have to be reentered into the object's primary
splay tree and memq.
(2) Cached pages are now stored alongside free pages in the physical
memory allocator's buddy queues, increasing the likelihood that large
allocations of contiguous physical memory (i.e., superpages) will
succeed.
Finally, as a result of this change long-standing restrictions on when
and where a cached page can be reclaimed and returned by
vm_page_alloc(9) are eliminated. Specifically, calls to
vm_page_alloc(9) specifying VM_ALLOC_INTERRUPT can now reclaim and
return a formerly cached page. Consequently, a call to malloc(9)
specifying M_NOWAIT is less likely to fail.
Discussed with: many over the course of the summer, including jeff@,
Justin Husted @ Isilon, peter@, tegge@
Tested by: an earlier version by kris@
Approved by: re (kensmith)
There's no advantage in allowing nested external interrupts.
In fact, it leads to a potential stack overrun.
While here, put the interrupt vector in the trapframe, so as
to compensate for the 36 cycle latency of reading cr.ivr.
Further simplify assembly code by dealing with ASTs from C.
Approved by: re (blanket)
us to do the data serializations once after writing multiple
region registers, as is done in pmap_switch(). All existing
calls to ia64_set_rr() are followed with calls to ia64_srlz_d().
Approved by: re (blanket)
(INTR_FILTER). This includes:
o Save a pointer to the sapic structure and IRQ for every vector,
so that we can quickly EOI, mask and unmask the interrupt.
o Add locking to the sapic code now that we can reprogram a
sapic on multiple CPUs at the same time.
o Use u_int for the vector and IRQ. We only have 256 vectors, so
using a 64-bit type for it is rather excessive.
o Properly handle concurrent registration of a handler for the
same vector.
Since vectors have a corresponding priority, we should not map
IRQs to vectors in a linear fashion, but rather pick a vector
that has a priority in line with the interrupt type. This is left
for later. The vector/IRQ interchange has been untangled as much
as possible to make this easier.
Approved by: re (blacket)
merely lucky that the VHPT was mapped as a side-effect of
mapping the kernel, but when there's enough physical memory,
this may not at all be the case.
Approved by: re (blanket)
is really a memory mapped I/O address. The bug is in the GAS that
describes the address and in particular the SpaceId field. The field
should not say the address is an I/O port when it clearly is not.
With an additional check for the IA64_BUS_SPACE_IO case in the bus
access functions, and the fact that I/O ports pretty much not used
in general on ia64, make the calculation of the I/O port address a
function. This avoids inlining the work-around into every driver,
and also helps reduce overall code bloat.
caches with data caches after writing to memory. This typically
is required to make breakpoints work on ia64 and powerpc. For
those architectures the function is implemented.
- Rename PCPU_LAZY_INC into PCPU_INC
- Add the PCPU_ADD interface which just does an add on the pcpu member
given a specific value.
Note that for most architectures PCPU_INC and PCPU_ADD are not safe.
This is a point that needs some discussions/work in the next days.
Reviewed by: alc, bde
Approved by: jeff (mentor)
speculative loads. This at least makes control speculative loads
work. In the future we should analyze which faults/exceptions
we want to handle rather than defer to avoid having to call the
recovery code when it's not strictly necessary.
VM_PHYSSEG_SPARSE depending on whether the physical address space is
densely or sparsely populated with memory. The effect of this
definition is to determine which of two implementations of
vm_page_array and PHYS_TO_VM_PAGE() is used. The legacy
implementation is obtained by defining VM_PHYSSEG_DENSE, and a new
implementation that trades off time for space is obtained by defining
VM_PHYSSEG_SPARSE. For now, all architectures except for ia64 and
sparc64 define VM_PHYSSEG_DENSE. Defining VM_PHYSSEG_SPARSE on ia64
allows the entirety of my Itanium 2's memory to be used. Previously,
only the first 1 GB could be used. Defining VM_PHYSSEG_SPARSE on
sparc64 allows USIIIi-based systems to boot without crashing.
This change is a combination of Nathan Whitehorn's patch and my own
work in perforce.
Discussed with: kmacy, marius, Nathan Whitehorn
PR: 112194
vm.kmem_size_min. Useful when using ZFS to make sure that vm.kmem size will
be at least 256mb (for example) without forcing a particular value via vm.kmem_size.
Approved by: njl (mentor)
Reviewed by: alc
Change the pc_pcb field from a pointer to struct pcb to struct pcb
so that sizeof(struct pcb) includes the PCB we use for IPI_STOP.
Statically declare early_pcb so that we don't have to allocate the
PCB for thread0. This way we can setup the PCPU before cninit()
and thus before we use printf().
Submitted by:
Reviewed by:
Approved by:
Obtained from:
MFC after:
Security:
Move the relocation definitions to the common elf header so that DTrace
can use them on one architecture targeted to a different one.
Add the additional ELF types defines in Sun's "Linker and Libraries"
manual.
WB (write-back) on x86 via control bits in PTEs and PDEs (including making
use of the PAT MSR). Changes include:
- A new pmap_mapdev_attr() function for amd64 and i386 which takes an
additional parameter (relative to pmap_mapdev()) specifying the cache
mode for this mapping. Note that on amd64 only WB mappings are done with
the direct map, all other modes result in a private mapping.
- pmap_mapdev() on i386 and amd64 now defaults to using UC (uncached)
mappings rather than WB. Previously we relied on the BIOS setting up
MTRR's to enforce memio regions being treated as UC. This might make
hw.cbb_start_memory unnecessary in some cases now for example.
- A new pmap_mapbios()/pmap_unmapbios() API has been added to allow places
that used pmap_mapdev() to map non-device memory (such as ACPI tables)
to do so using WB as before.
- A new pmap_change_attr() function for amd64 and i386 that changes the
caching mode for a range of KVA.
Reviewed by: alc
PCB in which the context of stopped CPUs is stored. To access this
PCB from KDB, we introduce a new define, called KDB_STOPPEDPCB. The
definition, when present, lives in <machine/kdb.h> and abstracts
where MD code saves the context. Define KDB_STOPPEDPCB on i386,
amd64, alpha and sparc64 in accordance to previous code.
param.h. Per request, I've placed these just after the
_NO_NAMESPACE_POLLUTION ifndef. I've not renamed anything yet, but
may since we don't need the __.
Submitted by: bde, jhb, scottl, many others.
passing a pointer to an opaque clockframe structure and requiring the
MD code to supply CLKF_FOO() macros to extract needed values out of the
opaque structure, just pass the needed values directly. In practice this
means passing the pair (usermode, pc) to hardclock() and profclock() and
passing the boolean (usermode) to hardclock_cpu() and hardclock_process().
Other details:
- Axe clockframe and CLKF_FOO() macros on all architectures. Basically,
all the archs were taking a trapframe and converting it into a clockframe
one way or another. Now they can just extract the PC and usermode values
directly out of the trapframe and pass it to fooclock().
- Renamed hardclock_process() to hardclock_cpu() as the latter is more
accurate.
- On Alpha, we now run profclock() at hz (profhz == hz) rather than at
the slower stathz.
- On Alpha, for the TurboLaser machines that don't have an 8254
timecounter, call hardclock() directly. This removes an extra
conditional check from every clock interrupt on Alpha on the BSP.
There is probably room for even further pruning here by changing Alpha
to use the simplified timecounter we use on x86 with the lapic timer
since we don't get interrupts from the 8254 on Alpha anyway.
- On x86, clkintr() shouldn't ever be called now unless using_lapic_timer
is false, so add a KASSERT() to that affect and remove a condition
to slightly optimize the non-lapic case.
- Change prototypeof arm_handler_execute() so that it's first arg is a
trapframe pointer rather than a void pointer for clarity.
- Use KCOUNT macro in profclock() to lookup the kernel profiling bucket.
Tested on: alpha, amd64, arm, i386, ia64, sparc64
Reviewed by: bde (mostly)
- Move vtophys() macros next to vtopte() where vtopte() exists to match
comments above vtopte().
- Remove references to the alternate address space in the comment above
vtopte(). amd64 never had the alternate address space, and i386 lost it
prior to PAE support being added.
- s/entires/entries/ in comments.
Reviewed by: alc
MACHINE_ARCH and MACHINE). Their purpose was to be able to test
in cpp(1), but cpp(1) only understands integer type expressions.
Using such unsupported expressions introduced a number of subtle
bugs, which were discovered by compiling with -Wundef.
variable and returns the previous value of the variable.
Tested on: i386, alpha, sparc64, arm (cognet)
Reviewed by: arch@
Submitted by: cognet (arm)
MFC after: 1 week
o s/vhpt_base/pmap_vhpt_base/g
o s/vhpt_bucket/pmap_vhpt_bucket/g
o Declare the above in <machine/pmap.h>
o Move the vm.stats.vhpt.* sysctls to machdep.vhpt.*
o Create a tunable machdep.vhpt.log2size, with corresponding sysctl.
The tunable allows the user to specify the VHPT size from the loader.
o Don't keep track of the number of PTEs in the VHPT. Calculate the
population when necessary by iterating the buckets and summing up
the length of the buckets.
o Don't perform the tpa instruction with a bucket lock held. The
instruction can (theoretically) fault and locking is not needed.
it to __MINSIGSTKSZ. Define MINSIGSTKSZ in <sys/signal.h>.
This is done in order to use MINSIGSTKSZ for the macro PTHREAD_STACK_MIN
in <pthread.h> (soon <limits.h>) without having to include the whole
<sys/signal.h> header.
Discussed with: bde
o Allocate a VHPT per CPU. The VHPT is a hash table that the CPU
uses to look up translations it can't find in the TLB. As such,
the VHPT serves as a level 1 cache (the TLB being a level 0 cache)
and best results are obtained when it's not shared between CPUs.
The collision chain (i.e. the hash bucket) is shared between CPUs,
as all buckets together constitute our collection of PTEs. To
achieve this, the collision chain does not point to the first PTE
in the list anymore, but to a hash bucket head structure. The
head structure contains the pointer to the first PTE in the list,
as well as a mutex to lock the bucket. Thus, each bucket is locked
independently of each other. With at least 1024 buckets in the VHPT,
this provides for sufficiently finei-grained locking to make the
ssolution scalable to large SMP machines.
o Add synchronisation to the lazy FP context switching. We do this
with a seperate per-thread lock. On SMP machines the lazy high FP
context switching without synchronisation caused inconsistent
state, which resulted in a panic. Since the use of the high FP
registers is not common, it's possible that races exist. The ia64
package build has proven to be a good stress test, so this will
get plenty of exercise in the near future.
o Don't use the local ID of the processor we want to send the IPI to
as the argument to ipi_send(). use the struct pcpu pointer instead.
The reason for this is that IPI delivery is unreliable. It has been
observed that sending an IPI to a CPU causes it to receive a stray
external interrupt. As such, we need a way to make the delivery
reliable. The intended solution is to queue requests in the target
CPU's per-CPU structure and use a single IPI to inform the CPU that
there's a new entry in the queue. If that IPI gets lost, the CPU
can check it's queue at any convenient time (such as for each
clock interrupt). This also allows us to send requests to a CPU
without interrupting it, if such would be beneficial.
With these changes SMP is almost working. There are still some random
process crashes and the machine can hang due to having the IPI lost
that deals with the high FP context switch.
The overhead of introducing the hash bucket head structure results
in a performance degradation of about 1% for UP (extra pointer
indirection). This is surprisingly small and is offset by gaining
reasonably/good scalable SMP support.
allocating a VHPT per CPU. Since we don't yet know how many CPUs
are actually in the system at the time we need to allocate the
VHPTs, we allocate for MAXCPU processors. This can result in a
lot of wasted space for 2-way machines. So, for now, limit MAXCPU
to something smaller until we have something more dynamic.
variables rather than void * variables. This makes it easier and simpler
to get asm constraints and volatile keywords correct.
MFC after: 3 days
Tested on: i386, alpha, sparc64
Compiled on: ia64, powerpc, amd64
Kernel toolchain busted on: arm
This case is triggered with ptrace(2) and the PT_SETREGS function.
Change the return type of the function to int so that errors can be
passed on to the caller.
Approved by: re (scottl)
pointer doesn't point to the first instruction of that function, but
rather to a descriptor. The descriptor has the address of the first
instruction, as well as the value of the global pointer. The symbol
table doesn't know anything about descriptors, so if you lookup the
name of a function you get the address of the first instruction. The
cast from the address, which is the result of the symbol lookup, to a
function pointer as is done in db_fncall is therefore invalid.
Abstract this detail behind the DB_CALL macro. By default DB_CALL is
defined as db_fncall_generic, which yields the old behaviour. On ia64
the macro is defined as db_fncall_ia64, in which a descriptor is
constructed to yield a valid function pointer.
While here, introduce DB_MAXARGS. DB_MAXARGS replaces the existing
(local) MAXARGS. The DB_MAXARGS macro can be defined by platforms to
create a convenient maximum. By default this will be the legacy 10.
On ia64 we define this macro to be 8, for 8 is the maximum number of
arguments that can be passed in registers. This avoids having to
implement spilling of arguments on the memory stack.
Approved by: re (dwhite)
- Implement sampling modes and logging support in hwpmc(4).
- Separate MI and MD parts of hwpmc(4) and allow sharing of
PMC implementations across different architectures.
Add support for P4 (EMT64) style PMCs to the amd64 code.
- New pmcstat(8) options: -E (exit time counts) -W (counts
every context switch), -R (print log file).
- pmc(3) API changes, improve our ability to keep ABI compatibility
in the future. Add more 'alias' names for commonly used events.
- bug fixes & documentation.
o Remove the clock interface. Not only does it conflict with the MI
version when device genclock is added to the kernel, it was also
not possible to have more than 1 clock device. This of course would
have been a problem if we actually had more than 1 clock device.
In short: we don't need a clock interface and if we do eventually,
we should be using the MI one.
o Rewrite inittodr() and resettodr() to take into account that:
1) We use the EFI interface directly.
2) time_t is 64-bit and we do need to make sure we can determine
leap years from year 2100 and on. Add a nice explanation of
where leap years come from and why.
3) This rewrite happened in 2005 so any date prior to 1/1/2005
(either M/D/Y or D/M/Y) is bogus. Reprogram the EFI clock with
1/1/2005 in that case.
4) The EFI clock has a high probability of being correct, so
only (further) correct the EFI clock when the file system time
is larger. That should never happen in a time-synchronised world.
Complain when EFI lost 2 days or more.
Replace the copyright notice now that I (pretty much) rewrote all of
this file.
into _bus.h to help with name space polution from including all of bus.h.
In a few days, I'll commit changes to the MI code to take advantage of thse
sepration (after I've made sure that these changes don't break anything in
the main tree, I've tested in my trees, but you never know...).
Suggested by: bde (in 2002 or 2003 I think)
Reviewed in principle by: jhb
critical_enter() and critical_exit() are now solely a mechanism for
deferring kernel preemptions. They no longer have any affect on
interrupts. This means that standalone critical sections are now very
cheap as they are simply unlocked integer increments and decrements for the
common case.
Spin mutexes now use a separate KPI implemented in MD code: spinlock_enter()
and spinlock_exit(). This KPI is responsible for providing whatever MD
guarantees are needed to ensure that a thread holding a spin lock won't
be preempted by any other code that will try to lock the same lock. For
now all archs continue to block interrupts in a "spinlock section" as they
did formerly in all critical sections. Note that I've also taken this
opportunity to push a few things into MD code rather than MI. For example,
critical_fork_exit() no longer exists. Instead, MD code ensures that new
threads have the correct state when they are created. Also, we no longer
try to fixup the idlethreads for APs in MI code. Instead, each arch sets
the initial curthread and adjusts the state of the idle thread it borrows
in order to perform the initial context switch.
This change is largely a big NOP, but the cleaner separation it provides
will allow for more efficient alternative locking schemes in other parts
of the kernel (bare critical sections rather than per-CPU spin mutexes
for per-CPU data for example).
Reviewed by: grehan, cognet, arch@, others
Tested on: i386, alpha, sparc64, powerpc, arm, possibly more
sys/bus_dma.h instead of being copied in every single arch. This slightly
reorders a flag that was specific to AXP and thus changes the ABI there.
The interface still relies on bus_space definitions found in <machine/bus.h>
so it cannot be included on its own yet, but that will be fixed at a later
date. Add an MD <machine/bus_dma.h> for ever arch for consistency and to
allow for future MD augmentation of the API. sparc64 makes heavy use of
this right now due to its different bus_dma implemenation.
place.
This moves the dependency on GCC's and other compiler's features into
the central sys/cdefs.h file, while the individual source files can
then refer to #ifdef __COMPILER_FEATURE_FOO where they by now used to
refer to #if __GNUC__ > 3.1415 && __BARC__ <= 42.
By now, GCC and ICC (the Intel compiler) have been actively tested on
IA32 platforms by netchild. Extension to other compilers is supposed
to be possible, of course.
Submitted by: netchild
Reviewed by: various developers on arch@, some time ago
o implement double-extended and single precision loads and stores,
o implement double precision stores,
o replace the machdep.unaligned_print sysctl with debug.unaligned_print
and change the default value to 0,
o replace the machdep.unaligned_sigbus sysctl with debug.unaligned_test,
o Remmove the fillfd() function. The function is trvial enough for
inline assembly.
The debug.unaligned_test sysctl is used to test the emulation of
misaligned loads and stores. When PSR.ac is 0, the CPU will handle
misaligned memory accesses itselfi and we don't get an exception
for it. When PSR.ac is 1, the process needs to be signalled and we
should not emulate. The sysctl takes effect when PSR.ac is 1 and
tells us that we should emulate and not send a signal.
PR: 72268
MFC after: 1 week
specified register, but a pointer to the in-memory representation of
that value. The reason for this is twofold:
1. Not all registers can be represented by a register_t. In particular
FP registers fall in that category. Passing the new register value
by reference instead of by value makes this point moot.
2. When we receive a G or P packet, both are for writing a register,
the packet will have the register value in target-byte order and
in the memory representation (modulo the fact that bytes are sent
as 2 printable hexadecimal numbers of course). We only need to
decode the packet to have a pointer to the register value.
This change fixes the bug of extracting the register value of the P
packet as a hexadecimal number instead of as a bit array. The quick
(and dirty) fix to bswap the register value in gdb_cpu_setreg() as
it has been added on i386 and amd64 can therefore be removed and has
in fact been that.
Tested on: alpha, amd64, i386, ia64, sparc64
o Remove a bogus comment that relates to alpha.
o s/u_int64_t/uint64_t/g
o Add bi_spare2 to make the internal padding explicit.
o Move BOOTINFO_MAGIC after the field it applies to.
old or previous value instead of void. This is not as is documented
in atomic(9), but is API (and ABI) compatible and simply makes sense.
This feature will primarily be used for atomic PTE updates in PMAP/ng.
Completely remove the remaining EFI includes and add our own (type)
definitions instead. While here, abstract more of the internals by
providing interface functions.
EFI headers and put them all in <machine/fpu.h>. The Intel EFI headers
conflict with the Intel ACPI headers (duplicate type definitions), so
are being phased out in the kernel.
these two reasons:
1. On ia64 a function pointer does not hold the address of the first
instruction of a functions implementation. It holds the address
of a function descriptor. Hence the user(), btrap(), eintr() and
bintr() prototypes are wrong for getting the actual code address.
2. The logic forces interrupt, trap and exception entry points to
be layed-out contiguously. This can not be achieved on ia64 and is
generally just bad programming.
The MCOUNT_FROMPC_USER macro is used to set the frompc argument to
some kernel address which represents any frompc that falls outside
the kernel text range. The macro can expand to ~0U to bail out in
that case.
The MCOUNT_FROMPC_INTR macro is used to set the frompc argument to
some kernel address to represent a call to a trap or interrupt
handler. This to avoid that the trap or interrupt handler appear to
be called from everywhere in the call graph. The macro can expand
to ~0U to prevent adjusting frompc. Note that the argument is selfpc,
not frompc.
This commit defines the macros on all architectures equivalently to
the original code in sys/libkern/mcount.c. People can take it from
here...
Compile-tested on: alpha, amd64, i386, ia64 and sparc64
Boot-tested on: i386
of the MCOUNT_ENTER, MCOUNT_EXIT and MCOUNT_DECL defines. Also make
sure there's a prototype of _MCOUNT_DECL(). This allows us to build
a kernel. There are still unresolved symbols, so linking fails.
_mcount() stub when profiling is enabled. Emit this code sequence
for assembly routines as welli (MCOUNT definition in <machine/asm.h>.
We do not pass the GOT entry however as the 4th argument, because it's
not used. The _mcount() stub calls __mcount(), which does the actual
work. Define _MCOUNT_DECL to define __mcount. We do not have an
implementation of mcount(), so we define MCOUNT as empty, but have a
weak alias to _mcount() in _mcount.S.
Note that the _mcount() stub in the kernel is slightly different from
the stub in userland. This is because we do not have to worry about
nested routines in the kernel.
have been rush hour...
While here, move COMPAT_IA32 from opt_global.h to opt_compat.h like on
amd64. Consequently, it's unsafe to use the option in pcb.h. We now
unconditionally have the ia32 specific registers in the PCB.
This commit is untested.
to allow dumping per-thread machine specific notes. On ia64 we use this
function to flush the dirty registers onto the backingstore before we
write out the PRSTATUS notes.
Tested on: alpha, amd64, i386, ia64 & sparc64
Not tested on: arm, powerpc
The hardware always gives read access for privilege level 0, which
means that we cannot use the hardware access rights and privilege
level in the PTE to test whether there's a change in protection. So,
we save the original vm_prot_t in the PTE as well.
Add pmap_pte_prot() to set the proper access rights and privilege
level on the PTE given a pmap and the requested protection.
The above allows us to compare the protection in pmap_extract_and_hold()
which was missing. While in pmap_extract_and_hold(), add pmap locking.
While here, clean up most (i.e. all but one) PTE macros we inherited
from alpha. They were either unused, used inconsistently, badly named
or simply weren't beneficial. We save the wired and managed state of
the PTE in distinct (bit) fields.
While in pte.h, s/u_int64_t/uint64_t/g
pmap locking obtained from: alc@
feedback & review by: alc@
related to breakpoints and single stepping into SIGTRAP so gdb(1) knows
why the remote target has stopped. In particular, gdb(1) needs to know
if the reason is something of its own doing.
being defined, define and use a new MD macro, cpu_spinwait(). It only
expands to something on i386 and amd64, so the compiled code should be
identical.
Name of the macro found by: jhb
Reviewed by: jhb
their own directory and module, leaving the MD parts in the MD
area (the MD parts _are_ part of the modules). /dev/mem and /dev/io
are now loadable modules, thus taking us one step further towards
a kernel created entirely out of modules. Of course, there is nothing
preventing the kernel from having these statically compiled.
dereference curthread. It is called only from critical_{enter,exit}(),
which already dereferences curthread. This doesn't seem to affect SMP
performance in my benchmarks, but improves MySQL transaction throughput
by about 1% on UP on my Xeon.
Head nodding: jhb, bmilekic
Most of the changes are a direct result of adding thread awareness.
Typically, DDB_REGS is gone. All registers are taken from the
trapframe and backtraces use the PCB based contexts. DDB_REGS was
defined to be a trapframe on all platforms anyway.
Thread awareness introduces the following new commands:
thread X switch to thread X (where X is the TID),
show threads list all threads.
The backtrace code has been made more flexible so that one can
create backtraces for any thread by giving the thread ID as an
argument to trace.
With this change, ia64 has support for breakpoints.
o ksym_start and ksym_end changed type to vm_offset_t.
o Make debugging support conditional upon KDB instead of DDB.
o Call kdb_enter() instead of breakpoint().
o Remove implementation of Debugger().
o Call kdb_trap() according to the new world order.
unwinder:
o s/db_active/kdb_active/g
o Various s/ddb/kdb/g
o Add support for unwinding from the PCB as well as the trapframe.
Abuse a spare field in the special register set to flag whether
the PCB was actually constructed from a trapframe so that we can
make the necessary adjustments.
md_var.h:
o Add RSE convenience macros.
o Add ia64_bsp_adjust() to add or subtract from BSP while taking
NaT collections into account.
a PCB from a trapframe for purposes of unwinding the stack. The PCB
is used as the thread context and all but the thread that entered the
debugger has a valid PCB.
This function can also be used to create a context for the threads
running on the CPUs that have been stopped when the debugger got
entered. This however is not done at the time of this commit.
in which multiple (presumably different) debugger backends can be
configured and which provides basic services to those backends.
Besides providing services to backends, it also serves as the single
point of contact for any and all code that wants to make use of the
debugger functions, such as entering the debugger or handling of the
alternate break sequence. For this purpose, the frontend has been
made non-optional.
All debugger requests are forwarded or handed over to the current
backend, if applicable. Selection of the current backend is done by
the debug.kdb.current sysctl. A list of configured backends can be
obtained with the debug.kdb.available sysctl. One can enter the
debugger by writing to the debug.kdb.enter sysctl.
backend improves over the old GDB support in the following ways:
o Unified implementation with minimal MD code.
o A simple interface for devices to register themselves as debug
ports, ala consoles.
o Compression by using run-length encoding.
o Implements GDB threading support.
