work properly with single-stepping in a kernel debugger. Specifically,
these routines have always disabled interrupts before increasing the nesting
count and restored the prior state of interrupts after decreasing the nesting
count to avoid problems with a nested interrupt not disabling interrupts
when acquiring a spin lock. However, trap interrupts for single-stepping
can still occur even when interrupts are disabled. Now the saved state of
interrupts is not saved in the thread until after interrupts have been
disabled and the nesting count has been increased. Similarly, the saved
state from the thread cannot be read once the nesting count has been
decreased to zero. To fix this, use temporary variables to store interrupt
state and shuffle it between the thread's MD area and the appropriate
registers.
In cooperation with: bde
MFC after: 1 month
only, and should be protected with an ifdef, and the no-execute bit in
32-bit set_user_sr() should be set before the comparison, not after, or
it will never match.
set_user_sr() itself caches the user segment VSID, there is no need for
cpu_switch() to do it again. This change also unifies the 32 and 64-bit
code paths for kernel faults on user pages and remaps the user SLB slot
on 64-bit systems when taking a syscall to avoid some unnecessary segment
exception traps.
concurrency bug. Since all SLB/SR entries were invalidated during an
exception, a decrementer exception could cause the user segment to be
invalidated during a copyin()/copyout() without a thread switch that
would cause it to be restored from the PCB, potentially causing the
operation to continue on invalid memory. This is now handled by explicit
restoration of segment 12 from the PCB on 32-bit systems and a check in
the Data Segment Exception handler on 64-bit.
While here, cause copyin()/copyout() to check whether the requested
user segment is already installed, saving some pipeline flushes, and
fix the synchronization primitives around the mtsr and slbmte
instructions to prevent accessing stale segments.
MFC after: 2 weeks
values to zero. A correct solution would involve emulating vector
operations on denormalized values, but this has little effect on accuracy
and is much less complicated for now.
MFC after: 2 weeks
by just caching the mode for later use by pmap_enter(), following amd64.
While here, correct some mismerges from mmu_oea64 -> mmu_oea and clean
up some dead code found while fixing the fictitious page behavior.
which are similar to the previous ones, and one for user maps, which
are arrays of pointers into the SLB tree. This changes makes user SLB
updates atomic, closing a window for memory corruption. While here,
rearrange the allocation functions to make context switches faster.
hardware with a lockless sparse tree design. This marginally improves
the performance of PMAP and allows copyin()/copyout() to run without
acquiring locks when used on wired mappings.
Submitted by: mdf
include/mmuvar.h - Change the MMU_DEF macro to also create the class
definition as well as define the DATA_SET. Add a macro, MMU_DEF_INHERIT,
which has an extra parameter specifying the MMU class to inherit methods
from. Update the comments at the start of the header file to describe the
new macros.
booke/pmap.c
aim/mmu_oea.c
aim/mmu_oea64.c - Collapse mmu_def_t declaration into updated MMU_DEF macro
The MMU_DEF_INHERIT macro will be used in the PS3 MMU implementation to
allow it to inherit the stock powerpc64 MMU methods.
Reviewed by: nwhitehorn
The main goal of this is to generate timer interrupts only when there is
some work to do. When CPU is busy interrupts are generating at full rate
of hz + stathz to fullfill scheduler and timekeeping requirements. But
when CPU is idle, only minimum set of interrupts (down to 8 interrupts per
second per CPU now), needed to handle scheduled callouts is executed.
This allows significantly increase idle CPU sleep time, increasing effect
of static power-saving technologies. Also it should reduce host CPU load
on virtualized systems, when guest system is idle.
There is set of tunables, also available as writable sysctls, allowing to
control wanted event timer subsystem behavior:
kern.eventtimer.timer - allows to choose event timer hardware to use.
On x86 there is up to 4 different kinds of timers. Depending on whether
chosen timer is per-CPU, behavior of other options slightly differs.
kern.eventtimer.periodic - allows to choose periodic and one-shot
operation mode. In periodic mode, current timer hardware taken as the only
source of time for time events. This mode is quite alike to previous kernel
behavior. One-shot mode instead uses currently selected time counter
hardware to schedule all needed events one by one and program timer to
generate interrupt exactly in specified time. Default value depends of
chosen timer capabilities, but one-shot mode is preferred, until other is
forced by user or hardware.
kern.eventtimer.singlemul - in periodic mode specifies how much times
higher timer frequency should be, to not strictly alias hardclock() and
statclock() events. Default values are 2 and 4, but could be reduced to 1
if extra interrupts are unwanted.
kern.eventtimer.idletick - makes each CPU to receive every timer interrupt
independently of whether they busy or not. By default this options is
disabled. If chosen timer is per-CPU and runs in periodic mode, this option
has no effect - all interrupts are generating.
As soon as this patch modifies cpu_idle() on some platforms, I have also
refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions
(if supported) under high sleep/wakeup rate, as fast alternative to other
methods. It allows SMP scheduler to wake up sleeping CPUs much faster
without using IPI, significantly increasing performance on some highly
task-switching loads.
Tested by: many (on i386, amd64, sparc64 and powerc)
H/W donated by: Gheorghe Ardelean
Sponsored by: iXsystems, Inc.
This reflects actual type used to store and compare child device orders.
Change is mostly done via a Coccinelle (soon to be devel/coccinelle)
semantic patch.
Verified by LINT+modules kernel builds.
Followup to: r212213
MFC after: 10 days
wrong element of the VSID bitmap array to be examined after a collision,
leading to reallocation of in-use VSIDs under some circumstances, with
attendant memory corruption. Also add an assert to check for this kind of
problem in the future.
MFC after: 4 days
PowerPC CPUs running a 32-bit kernel. This bug could cause in-use VSIDs
to be allocated again to another process, causing memory space overlaps
and corruption.
Reported by: linimon
the existing code was very platform specific, and broken for SMP systems
trying to reboot from KDB.
- Add a new PLATFORM_RESET() method to the platform KOBJ interface, and
migrate existing reset functions into platform modules.
- Modify the OF_reboot() routine to submit the request by hand to avoid
the IPIs involved in the regular openfirmware() routine. This fixes
reboot from KDB on SMP machines.
- Move non-KDB reset and poweroff functions on the Powermac platform
into the relevant power control drivers (cuda, pmu, smu), instead of
using them through the Open Firmware backdoor.
- Rename platform_chrp to platform_powermac since it has become
increasingly Powermac specific. When we gain support for IBM systems,
we will grow a new platform_chrp.
Kernel sources for 64-bit PowerPC, along with build-system changes to keep
32-bit kernels compiling (build system changes for 64-bit kernels are
coming later). Existing 32-bit PowerPC kernel configurations must be
updated after this change to specify their architecture.
(exec_setregs, etc.) in order to simplify the addition of 64-bit support,
and possible future extension of the Book-E code to handle hard floating
point and Altivec.
MFC after: 1 month
CPUs by default, and provide a functional version of BUS_BIND_INTR().
While here, fix some potential concurrency problems in the interrupt
handling code.
IBAT entry in early boot in order to prevent possible faults from races
between the instruction cache and the MMU.
PR: powerpc/148003
MFC after: 3 days
allow pmap_enter() to be performed on an unmanaged page that doesn't have
VPO_BUSY set. Having VPO_BUSY set really only matters for managed pages.
(See, for example, pmap_remove_write().)
PG_REFERENCED changes in vm_pageout_object_deactivate_pages().
Simplify this function's inner loop using TAILQ_FOREACH(), and shorten
some of its overly long lines. Update a stale comment.
Assert that PG_REFERENCED may be cleared only if the object containing
the page is locked. Add a comment documenting this.
Assert that a caller to vm_page_requeue() holds the page queues lock,
and assert that the page is on a page queue.
Push down the page queues lock into pmap_ts_referenced() and
pmap_page_exists_quick(). (As of now, there are no longer any pmap
functions that expect to be called with the page queues lock held.)
Neither pmap_ts_referenced() nor pmap_page_exists_quick() should ever
be passed an unmanaged page. Assert this rather than returning "0"
and "FALSE" respectively.
ARM:
Simplify pmap_page_exists_quick() by switching to TAILQ_FOREACH().
Push down the page queues lock inside of pmap_clearbit(), simplifying
pmap_clear_modify(), pmap_clear_reference(), and pmap_remove_write().
Additionally, this allows for avoiding the acquisition of the page
queues lock in some cases.
PowerPC/AIM:
moea*_page_exits_quick() and moea*_page_wired_mappings() will never be
called before pmap initialization is complete. Therefore, the check
for moea_initialized can be eliminated.
Push down the page queues lock inside of moea*_clear_bit(),
simplifying moea*_clear_modify() and moea*_clear_reference().
The last parameter to moea*_clear_bit() is never used. Eliminate it.
PowerPC/BookE:
Simplify mmu_booke_page_exists_quick()'s control flow.
Reviewed by: kib@
pmap_is_referenced(). Eliminate the corresponding page queues lock
acquisitions from vm_map_pmap_enter() and mincore(), respectively. In
mincore(), this allows some additional cases to complete without ever
acquiring the page queues lock.
Assert that the page is managed in pmap_is_referenced().
On powerpc/aim, push down the page queues lock acquisition from
moea*_is_modified() and moea*_is_referenced() into moea*_query_bit().
Again, this will allow some additional cases to complete without ever
acquiring the page queues lock.
Reorder a few statements in vm_page_dontneed() so that a race can't lead
to an old reference persisting. This scenario is described in detail by a
comment.
Correct a spelling error in vm_page_dontneed().
Assert that the object is locked in vm_page_clear_dirty(), and restrict the
page queues lock assertion to just those cases in which the page is
currently writeable.
Add object locking to vnode_pager_generic_putpages(). This was the one
and only place where vm_page_clear_dirty() was being called without the
object being locked.
Eliminate an unnecessary vm_page_lock() around vnode_pager_setsize()'s call
to vm_page_clear_dirty().
Change vnode_pager_generic_putpages() to the modern-style of function
definition. Also, change the name of one of the parameters to follow
virtual memory system naming conventions.
Reviewed by: kib
independent code. Move this code into mincore(), and eliminate the
page queues lock from pmap_mincore().
Push down the page queues lock into pmap_clear_modify(),
pmap_clear_reference(), and pmap_is_modified(). Assert that these
functions are never passed an unmanaged page.
Eliminate an inaccurate comment from powerpc/powerpc/mmu_if.m:
Contrary to what the comment says, pmap_mincore() is not simply an
optimization. Without a complete pmap_mincore() implementation,
mincore() cannot return either MINCORE_MODIFIED or MINCORE_REFERENCED
because only the pmap can provide this information.
Eliminate the page queues lock from vfs_setdirty_locked_object(),
vm_pageout_clean(), vm_object_page_collect_flush(), and
vm_object_page_clean(). Generally speaking, these are all accesses
to the page's dirty field, which are synchronized by the containing
vm object's lock.
Reduce the scope of the page queues lock in vm_object_madvise() and
vm_page_dontneed().
Reviewed by: kib (an earlier version)
Extend struct sysvec with three new elements:
sv_fetch_syscall_args - the method to fetch syscall arguments from
usermode into struct syscall_args. The structure is machine-depended
(this might be reconsidered after all architectures are converted).
sv_set_syscall_retval - the method to set a return value for usermode
from the syscall. It is a generalization of
cpu_set_syscall_retval(9) to allow ABIs to override the way to set a
return value.
sv_syscallnames - the table of syscall names.
Use sv_set_syscall_retval in kern_sigsuspend() instead of hardcoding
the call to cpu_set_syscall_retval().
The new functions syscallenter(9) and syscallret(9) are provided that
use sv_*syscall* pointers and contain the common repeated code from
the syscall() implementations for the architecture-specific syscall
trap handlers.
Syscallenter() fetches arguments, calls syscall implementation from
ABI sysent table, and set up return frame. The end of syscall
bookkeeping is done by syscallret().
Take advantage of single place for MI syscall handling code and
implement ptrace_lwpinfo pl_flags PL_FLAG_SCE, PL_FLAG_SCX and
PL_FLAG_EXEC. The SCE and SCX flags notify the debugger that the
thread is stopped at syscall entry or return point respectively. The
EXEC flag augments SCX and notifies debugger that the process address
space was changed by one of exec(2)-family syscalls.
The i386, amd64, sparc64, sun4v, powerpc and ia64 syscall()s are
changed to use syscallenter()/syscallret(). MIPS and arm are not
converted and use the mostly unchanged syscall() implementation.
Reviewed by: jhb, marcel, marius, nwhitehorn, stas
Tested by: marcel (ia64), marius (sparc64), nwhitehorn (powerpc),
stas (mips)
MFC after: 1 month
where running ofwdump could cause hangs by forcing all secondary CPUs
into a busy wait with interrupts off during the call.
Following section 8.4 of the Open Firmware PowerPC processor binding,
the firmware is free to overwrite the system interrupt handlers during
OF calls, restoring the OS handlers on exit. On single CPU systems, this
process is invisible to the operating system. On multiple CPU systems,
taking any exception on a secondary CPU while an OF call is in progress
ends with that exception vectored into OF, resulting in a slow movement
of the entire system into firmware context and a machine hang.
MFC after: 3 days
here, make the style of assertion used by pmap_enter() consistent
across all architectures.
On entry to pmap_remove_write(), assert that the page is neither
unmanaged nor fictitious, since we cannot remove write access to
either kind of page.
With the push down of the page queues lock, pmap_remove_write() cannot
condition its behavior on the state of the PG_WRITEABLE flag if the
page is busy. Assert that the object containing the page is locked.
This allows us to know that the page will neither become busy nor will
PG_WRITEABLE be set on it while pmap_remove_write() is running.
Correct a long-standing bug in vm_page_cowsetup(). We cannot possibly
do copy-on-write-based zero-copy transmit on unmanaged or fictitious
pages, so don't even try. Previously, the call to pmap_remove_write()
would have failed silently.
firmware in order to take over control of the SMU. Without doing this,
the firmware background process doing fan control will run amok as we
take over the system and crash the management chip.
This is limited to these two machines because our kernel is heavily
dependent on firmware accesses, and so quiescing firmware can cause
nasty problems.
vm_page_try_to_free(). Consequently, push down the page queues lock into
pmap_enter_quick(), pmap_page_wired_mapped(), pmap_remove_all(), and
pmap_remove_write().
Push down the page queues lock into Xen's pmap_page_is_mapped(). (I
overlooked the Xen pmap in r207702.)
Switch to a per-processor counter for the total number of pages cached.
architecture from page queue lock to a hashed array of page locks
(based on a patch by Jeff Roberson), I've implemented page lock
support in the MI code and have only moved vm_page's hold_count
out from under page queue mutex to page lock. This changes
pmap_extract_and_hold on all pmaps.
Supported by: Bitgravity Inc.
Discussed with: alc, jeffr, and kib
pmap_ts_referenced() is not always appropriate for checking whether or
not pages have been referenced because it clears any reference bits
that it encounters. For example, in mincore(), clearing the reference
bits has two negative consequences. First, it throws off the activity
count calculations performed by the page daemon. Specifically, a page
on which mincore() has called pmap_ts_referenced() looks less active
to the page daemon than it should. Consequently, the page could be
deactivated prematurely by the page daemon. Arguably, this problem
could be fixed by having mincore() duplicate the activity count
calculation on the page. However, there is a second problem for which
that is not a solution. In order to clear a reference on a 4KB page,
it may be necessary to demote a 2/4MB page mapping. Thus, a mincore()
by one process can have the side effect of demoting a superpage
mapping within another process!
to the image_params struct instead of several members of that struct
individually. This makes it easier to expand its arguments in the future
without touching all platforms.
Reviewed by: jhb
in Open Firmware was Apple-specific, and we have complete coverage of Apple
system controllers, so move RTC responsibilities into the system controller
drivers. This avoids interesting problems from manipulating these devices
through Open Firmware behind the backs of their drivers.
Obtained from: NetBSD
MFC after: 2 weeks
generally protected by the VM page queue mutex. Instead of extending the
table lock to cover the PVO heads, add some asserts that the page queue
mutex is in fact held. This fixes several LORs and possible deadlocks.
This also adds an optimization to moea64_kextract() useful for
direct-mapped quantities, like UMA buffers. Being able to use this from
inside UMA removes an additional LOR.
access, and reflects this by autonomously writing LPTE_M into PTE entries.
As such, we should not panic if LPTE_M changes by itself. While here,
fix a harmless typo in moea64_sync_icache().
counting in incrementing the interrupt nesting level. This fixes a number
of bugs in which the interrupt thread could be preempted by an IPI,
indefinitely delaying acknowledgement of the interrupt to the PIC, causing
interrupt starvation and hangs.
Reported by: linimon
Reviewed by: marcel, jhb
MFC after: 1 week
its PVO to map physical address 0 instead of kernelstart. This fixes a
situation in which a user process could attempt to return this address
via KVM, have it fault while being modified, and then panic the kernel
because (a) it is supposed to map a valid address and (b) it lies in the
no-fault region between VM_MIN_KERNEL_ADDRESS and virtual_avail.
While here, move msgbuf and dpcpu make into regular KVA space for
consistency with other implementations.
physical address is changed, there is a brief window during which its PTE
is invalid. Since moea64_set_scratchpage_pa() does not and cannot hold
the page table lock, it was possible for another CPU to insert a new PTE
into the scratch page's PTEG slot during this interval, corrupting both
mappings.
Solve this by creating a new flag, LPTE_LOCKED, such that
moea64_pte_insert will avoid claiming locked PTEG slots even if they
are invalid. This change also incorporates some additional paranoia
added to solve things I thought might be this bug.
Reported by: linimon
UMA segments at their physical addresses instead of into KVA. This emulates
the direct mapping behavior of OEA32 in an ad-hoc way. To make this work
properly required sharing the entire kernel PMAP with Open Firmware, so
ofw_pmap is transformed into a stub on 64-bit CPUs.
Also implement some more tweaks to get more mileage out of our limited
amount of KVA, principally by extending KVA into segment 16 until the
beginning of the first OFW mapping.
Reported by: linimon
PVOs, and so the modified state of the page can no longer be communicated
to the VM layer, causing pages not to be flushed to swap when needed, in
turn causing memory corruption. Also make several correctness adjustments
to I-Cache synchronization and TLB invalidation for 64-bit Book-S CPUs.
Obtained from: projects/ppc64
Discussed with: grehan
MFC after: 2 weeks
more. This provides three new sysctls to user space:
hw.cpu_features - A bitmask of available CPU features
hw.floatingpoint - Whether or not there is hardware FP support
hw.altivec - Whether or not Altivec is available
PR: powerpc/139154
MFC after: 10 days
from CD on 64-bit hardware to replace existing band-aids. This occurred
when the preloaded mdroot required too many mappings for the static
buffer.
Since we only use the translations buffer once, allocate a dynamic
buffer on the stack. This early in the boot process, the call chain
is quite short and we can be assured of having sufficient stack space.
Reviewed by: grehan
that use many translation regions in firmware, and add bounds checking
to prevent buffer overflows in case even the new value is exceeded.
Reported by: Jacob Lambert
MFC after: 3 days
the current value of its argument before atomically replacing it, which
could occasionally return the wrong value on an SMP system. This resulted
in user mutex operations hanging when using threaded applications.
while in kernel mode, and later changing signal mask to block the
signal, was fixed for sigprocmask(2) and ptread_exit(3). The same race
exists for sigreturn(2), setcontext(2) and swapcontext(2) syscalls.
Use kern_sigprocmask() instead of direct manipulation of td_sigmask to
reschedule newly blocked signals, closing the race.
Reviewed by: davidxu
Tested by: pho
MFC after: 1 month
only used in real mode and keeping them mapped only serves to make NULL
a valid address, which results in silent NULL pointer deferences.
Suggested by: Patrick Kerharo
Obtained from: projects/ppc64
at least on my Xserve, getting the decrementer and timebase on APs to tick
requires setting up a clock chip over I2C, which is not yet done.
While here, correct the 64-bit tlbie function to set the CPU to 64-bit
mode correctly.
Hardware donated by: grehan
the memory or D-cache, depending on the semantics of the platform.
vm_sync_icache() is basically a wrapper around pmap_sync_icache(),
that translates the vm_map_t argumument to pmap_t.
o Introduce pmap_sync_icache() to all PMAP implementation. For powerpc
it replaces the pmap_page_executable() function, added to solve
the I-cache problem in uiomove_fromphys().
o In proc_rwmem() call vm_sync_icache() when writing to a page that
has execute permissions. This assures that when breakpoints are
written, the I-cache will be coherent and the process will actually
hit the breakpoint.
o This also fixes the Book-E PMAP implementation that was missing
necessary locking while trying to deal with the I-cache coherency
in pmap_enter() (read: mmu_booke_enter_locked).
The key property of this change is that the I-cache is made coherent
*after* writes have been done. Doing it in the PMAP layer when adding
or changing a mapping means that the I-cache is made coherent *before*
any writes happen. The difference is key when the I-cache prefetches.
bandaid to prevent exhaustion of the primary and secondary hash groups
in the event of extreme stress on the PMAP layer (e.g. a forkbomb). This
wastes memory, and should be revised to properly handle PTEG spills instead.
Suggested by: grehan
Approved by: re (kensmith)
