The count register is volatile so we don't need to preserve it.
Store zero to the entry in the exception frame.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The XER is a volatile register so there is no need to save and restore
it over a system call - zero it out in the exception stack frame
instead.
This should fix a 5 cycle stall of the mfxer/std seen on POWER7.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
syscall_dotrace_cont and syscall_error_cont tend to complicate perf
output so make them local.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The current implementation of lazy interrupts handling has some
issues that this tries to address.
We don't do the various workarounds we need to do when re-enabling
interrupts in some cases such as when returning from an interrupt
and thus we may still lose or get delayed decrementer or doorbell
interrupts.
The current scheme also makes it much harder to handle the external
"edge" interrupts provided by some BookE processors when using the
EPR facility (External Proxy) and the Freescale Hypervisor.
Additionally, we tend to keep interrupts hard disabled in a number
of cases, such as decrementer interrupts, external interrupts, or
when a masked decrementer interrupt is pending. This is sub-optimal.
This is an attempt at fixing it all in one go by reworking the way
we do the lazy interrupt disabling from the ground up.
The base idea is to replace the "hard_enabled" field with a
"irq_happened" field in which we store a bit mask of what interrupt
occurred while soft-disabled.
When re-enabling, either via arch_local_irq_restore() or when returning
from an interrupt, we can now decide what to do by testing bits in that
field.
We then implement replaying of the missed interrupts either by
re-using the existing exception frame (in exception exit case) or via
the creation of a new one from an assembly trampoline (in the
arch_local_irq_enable case).
This removes the need to play with the decrementer to try to create
fake interrupts, among others.
In addition, this adds a few refinements:
- We no longer hard disable decrementer interrupts that occur
while soft-disabled. We now simply bump the decrementer back to max
(on BookS) or leave it stopped (on BookE) and continue with hard interrupts
enabled, which means that we'll potentially get better sample quality from
performance monitor interrupts.
- Timer, decrementer and doorbell interrupts now hard-enable
shortly after removing the source of the interrupt, which means
they no longer run entirely hard disabled. Again, this will improve
perf sample quality.
- On Book3E 64-bit, we now make the performance monitor interrupt
act as an NMI like Book3S (the necessary C code for that to work
appear to already be present in the FSL perf code, notably calling
nmi_enter instead of irq_enter). (This also fixes a bug where BookE
perfmon interrupts could clobber r14 ... oops)
- We could make "masked" decrementer interrupts act as NMIs when doing
timer-based perf sampling to improve the sample quality.
Signed-off-by-yet: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
v2:
- Add hard-enable to decrementer, timer and doorbells
- Fix CR clobber in masked irq handling on BookE
- Make embedded perf interrupt act as an NMI
- Add a PACA_HAPPENED_EE_EDGE for use by FSL if they want
to retrigger an interrupt without preventing hard-enable
v3:
- Fix or vs. ori bug on Book3E
- Fix enabling of interrupts for some exceptions on Book3E
v4:
- Fix resend of doorbells on return from interrupt on Book3E
v5:
- Rebased on top of my latest series, which involves some significant
rework of some aspects of the patch.
v6:
- 32-bit compile fix
- more compile fixes with various .config combos
- factor out the asm code to soft-disable interrupts
- remove the C wrapper around preempt_schedule_irq
v7:
- Fix a bug with hard irq state tracking on native power7
On 64-bit, the mfmsr instruction can be quite slow, slower
than loading a field from the cache-hot PACA, which happens
to already contain the value we want in most cases.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We unconditionally hard enable interrupts. This is unnecessary as
syscalls are expected to always be called with interrupts enabled.
While at it, we add a WARN_ON if that is not the case and
CONFIG_TRACE_IRQFLAGS is enabled (we don't want to add overhead
to the fast path when this is not set though).
Thus let's remove the enabling (and associated irq tracing) from
the syscall entry path. Also on Book3S, replace a few mfmsr
instructions with loads of PACAMSR from the PACA, which should be
faster & schedule better.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This removes the various bits of assembly in the kernel entry,
exception handling and SLB management code that were specific
to running under the legacy iSeries hypervisor which is no
longer supported.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We have a few problems when returning to userspace. This is a
quick set of fixes for 3.3, I'll look into a more comprehensive
rework for 3.4. This fixes:
- We kept interrupts soft-disabled when schedule'ing or calling
do_signal when returning to userspace as a result of a hardware
interrupt.
- Rename do_signal to do_notify_resume like all other archs (and
do_signal_pending back to do_signal, which it was before Roland
changed it).
- Add the missing call to key_replace_session_keyring() to
do_notify_resume().
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
Some of the 64bit PPC CPU features are MMU-related, so this patch moves
them to MMU_FTR_ bits. All cpu_has_feature()-style tests are moved to
mmu_has_feature(), and seven feature bits are freed as a result.
Signed-off-by: Matt Evans <matt@ozlabs.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The DSCR (aka Data Stream Control Register) is supported on some
server PowerPC chips and allow some control over the prefetch
of data streams.
This patch allows the value to be specified per thread by emulating
the corresponding mfspr and mtspr instructions. Children of such
threads inherit the value. Other threads use a default value that
can be specified in sysfs - /sys/devices/system/cpu/dscr_default.
If a thread starts with non default value in the sysfs entry,
all children threads inherit this non default value even if
the sysfs value is changed later.
Signed-off-by: Alexey Kardashevskiy <aik@au1.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
When running in Hypervisor mode (arch 2.06 or later), we store the PACA
in HSPRG0 instead of SPRG1. The architecture specifies that SPRGs may be
lost during a "nap" power management operation (though they aren't
currently on POWER7) and this enables use of SPRG1 by KVM guests.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Currently, when CONFIG_VIRT_CPU_ACCOUNTING is enabled, we use the
PURR register for measuring the user and system time used by
processes, as well as other related times such as hardirq and
softirq times. This turns out to be quite confusing for users
because it means that a program will often be measured as taking
less time when run on a multi-threaded processor (SMT2 or SMT4 mode)
than it does when run on a single-threaded processor (ST mode), even
though the program takes longer to finish. The discrepancy is
accounted for as stolen time, which is also confusing, particularly
when there are no other partitions running.
This changes the accounting to use the timebase instead, meaning that
the reported user and system times are the actual number of real-time
seconds that the program was executing on the processor thread,
regardless of which SMT mode the processor is in. Thus a program will
generally show greater user and system times when run on a
multi-threaded processor than on a single-threaded processor.
On pSeries systems on POWER5 or later processors, we measure the
stolen time (time when this partition wasn't running) using the
hypervisor dispatch trace log. We check for new entries in the
log on every entry from user mode and on every transition from
kernel process context to soft or hard IRQ context (i.e. when
account_system_vtime() gets called). So that we can correctly
distinguish time stolen from user time and time stolen from system
time, without having to check the log on every exit to user mode,
we store separate timestamps for exit to user mode and entry from
user mode.
On systems that have a SPURR (POWER6 and POWER7), we read the SPURR
in account_system_vtime() (as before), and then apportion the SPURR
ticks since the last time we read it between scaled user time and
scaled system time according to the relative proportions of user
time and system time over the same interval. This avoids having to
read the SPURR on every kernel entry and exit. On systems that have
PURR but not SPURR (i.e., POWER5), we do the same using the PURR
rather than the SPURR.
This disables the DTL user interface in /sys/debug/kernel/powerpc/dtl
for now since it conflicts with the use of the dispatch trace log
by the time accounting code.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The POWER architecture does not require stcx to check that it is operating
on the same address as the larx. This means it is possible for an
an exception handler to execute a larx, get a reservation, decide
not to do the stcx and then return back with an active reservation. If the
interrupted code was in the middle of a larx/stcx sequence the stcx could
incorrectly succeed.
All recent POWER CPUs check the address before letting the stcx succeed
so we can create a CPU feature and nop it out. As Ben suggested, we can
only do this in our syscall path because there is a remote possibility
some kernel code gets interrupted by an exception that ends up operating
on the same cacheline.
Thanks to Paul Mackerras and Derek Williams for the idea.
To test this I used a very simple null syscall (actually getppid) testcase
at http://ozlabs.org/~anton/junkcode/null_syscall.c
I tested against 2.6.35-git10 with the following changes against the
pseries_defconfig:
CONFIG_VIRT_CPU_ACCOUNTING=n
CONFIG_AUDIT=n
CONFIG_PPC_4K_PAGES=n
CONFIG_PPC_64K_PAGES=y
CONFIG_FORCE_MAX_ZONEORDER=9
CONFIG_PPC_SUBPAGE_PROT=n
CONFIG_FUNCTION_TRACER=n
CONFIG_FUNCTION_GRAPH_TRACER=n
CONFIG_IRQSOFF_TRACER=n
CONFIG_STACK_TRACER=n
to remove the overhead of virtual CPU accounting, syscall auditing and
the ftrace mcount tracers. 64kB pages were enabled to minimise TLB misses.
POWER6: +8.2%
POWER7: +7.0%
Another suggestion was to use a larx to something in the L1 instead of a stcx.
This was almost as fast as removing the larx on POWER6, but only 3.5% faster
on POWER7. We can use this to speed up the reservation clear in our
exception exit code.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Anton Blanchard found that large POWER systems would occasionally
crash in the exception exit path when profiling with perf_events.
The symptom was that an interrupt would occur late in the exit path
when the MSR[RI] (recoverable interrupt) bit was clear. Interrupts
should be hard-disabled at this point but they were enabled. Because
the interrupt was not recoverable the system panicked.
The reason is that the exception exit path was calling
perf_event_do_pending after hard-disabling interrupts, and
perf_event_do_pending will re-enable interrupts.
The simplest and cleanest fix for this is to use the same mechanism
that 32-bit powerpc does, namely to cause a self-IPI by setting the
decrementer to 1. This means we can remove the tests in the exception
exit path and raw_local_irq_restore.
This also makes sure that the call to perf_event_do_pending from
timer_interrupt() happens within irq_enter/irq_exit. (Note that
calling perf_event_do_pending from timer_interrupt does not mean that
there is a possible 1/HZ latency; setting the decrementer to 1 ensures
that the timer interrupt will happen immediately, i.e. within one
timebase tick, which is a few nanoseconds or 10s of nanoseconds.)
Signed-off-by: Paul Mackerras <paulus@samba.org>
Cc: stable@kernel.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
RTAS should never cause an exception but if it does (for example accessing
outside our RMO) then we might go a long way through the kernel before
oopsing. If we unset MSR_RI we should at least stop things on exception
exit.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
If CONFIG_PPC_ISERIES isn't defined we end up with
iseries_check_pending_irqs and do_work at the same address.
perf ends up picking iseries_check_pending_irqs which creates
confusing backtraces. Hide it.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Based on an original patch by Valentine Barshak <vbarshak@ru.mvista.com>
Use preempt_schedule_irq to prevent infinite irq-entry and
eventual stack overflow problems with fast-paced IRQ sources.
This kind of problems has been observed on the PASemi Electra IDE
controller. We have to make sure we are soft-disabled before calling
preempt_schedule_irq and hard disable interrupts after that
to avoid unrecoverable exceptions.
This patch also moves the "clrrdi r9,r1,THREAD_SHIFT" out of
the #ifdef CONFIG_PPC_BOOK3E scope, since r9 is clobbered
and has to be restored in both cases.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The mod_return_to_handler needs to switch to the kernel TOC before
jumping to a the kernel code. It currently does this by looking
at the kernel function data and retrieves the TOC that way.
Not only is this inefficient, it also breaks with a relocatable kernel.
The PACA contains the kernel TOC and we can easily retrieve it that
way.
Reported-by: Sachin Sant <sachinp@in.ibm.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Bye-bye Performance Counters, welcome Performance Events!
In the past few months the perfcounters subsystem has grown out its
initial role of counting hardware events, and has become (and is
becoming) a much broader generic event enumeration, reporting, logging,
monitoring, analysis facility.
Naming its core object 'perf_counter' and naming the subsystem
'perfcounters' has become more and more of a misnomer. With pending
code like hw-breakpoints support the 'counter' name is less and
less appropriate.
All in one, we've decided to rename the subsystem to 'performance
events' and to propagate this rename through all fields, variables
and API names. (in an ABI compatible fashion)
The word 'event' is also a bit shorter than 'counter' - which makes
it slightly more convenient to write/handle as well.
Thanks goes to Stephane Eranian who first observed this misnomer and
suggested a rename.
User-space tooling and ABI compatibility is not affected - this patch
should be function-invariant. (Also, defconfigs were not touched to
keep the size down.)
This patch has been generated via the following script:
FILES=$(find * -type f | grep -vE 'oprofile|[^K]config')
sed -i \
-e 's/PERF_EVENT_/PERF_RECORD_/g' \
-e 's/PERF_COUNTER/PERF_EVENT/g' \
-e 's/perf_counter/perf_event/g' \
-e 's/nb_counters/nb_events/g' \
-e 's/swcounter/swevent/g' \
-e 's/tpcounter_event/tp_event/g' \
$FILES
for N in $(find . -name perf_counter.[ch]); do
M=$(echo $N | sed 's/perf_counter/perf_event/g')
mv $N $M
done
FILES=$(find . -name perf_event.*)
sed -i \
-e 's/COUNTER_MASK/REG_MASK/g' \
-e 's/COUNTER/EVENT/g' \
-e 's/\<event\>/event_id/g' \
-e 's/counter/event/g' \
-e 's/Counter/Event/g' \
$FILES
... to keep it as correct as possible. This script can also be
used by anyone who has pending perfcounters patches - it converts
a Linux kernel tree over to the new naming. We tried to time this
change to the point in time where the amount of pending patches
is the smallest: the end of the merge window.
Namespace clashes were fixed up in a preparatory patch - and some
stylistic fallout will be fixed up in a subsequent patch.
( NOTE: 'counters' are still the proper terminology when we deal
with hardware registers - and these sed scripts are a bit
over-eager in renaming them. I've undone some of that, but
in case there's something left where 'counter' would be
better than 'event' we can undo that on an individual basis
instead of touching an otherwise nicely automated patch. )
Suggested-by: Stephane Eranian <eranian@google.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Reviewed-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: <linux-arch@vger.kernel.org>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
This contains all the bits that didn't fit in previous patches :-) This
includes the actual exception handlers assembly, the changes to the
kernel entry, other misc bits and wiring it all up in Kconfig.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
enter_prom() used to save and restore registers such as CTR, XER etc..
which are volatile, or SRR0,1... which we don't care about. This
removes a bunch of useless code and while at it turns an mtmsrd into
an MTMSRD macro which will be useful to Book3E.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The kernel uses SPRG registers for various purposes, typically in
low level assembly code as scratch registers or to hold per-cpu
global infos such as the PACA or the current thread_info pointer.
We want to be able to easily shuffle the usage of those registers
as some implementations have specific constraints realted to some
of them, for example, some have userspace readable aliases, etc..
and the current choice isn't always the best.
This patch should not change any code generation, and replaces the
usage of SPRN_SPRGn everywhere in the kernel with a named replacement
and adds documentation next to the definition of the names as to
what those are used for on each processor family.
The only parts that still use the original numbers are bits of KVM
or suspend/resume code that just blindly needs to save/restore all
the SPRGs.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Merge reason: we have gathered quite a few conflicts, need to merge upstream
Conflicts:
arch/powerpc/kernel/Makefile
arch/x86/ia32/ia32entry.S
arch/x86/include/asm/hardirq.h
arch/x86/include/asm/unistd_32.h
arch/x86/include/asm/unistd_64.h
arch/x86/kernel/cpu/common.c
arch/x86/kernel/irq.c
arch/x86/kernel/syscall_table_32.S
arch/x86/mm/iomap_32.c
include/linux/sched.h
kernel/Makefile
Signed-off-by: Ingo Molnar <mingo@elte.hu>
The TOCS used by modules are different than the one used by
the core kernel code. The function graph tracer must save and
restore the TOC whenever it traces a module call. But this
is an added overhead to burden the majority of core kernel
code being traced.
Benjamin Herrenschmidt suggested in testing the entry of
the call to tell if it is a core kernel function or a module.
He recommended using the REGION_ID() macro to perform this test.
This patch implements Benjamin's idea, and uses a different
return_to_handler routine dependent on if the entry is a core
kernel function or not. The module version saves the TOC, where as
the core kernel version does not.
Geoff Lavand tested on PS3.
Tested-by: Geoff Levand <geoffrey.levand@am.sony.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This is the port of the function graph tracer to PowerPC with
dynamic tracing.
Geoff Lavand tested on PS3.
Tested-by: Geoff Levand <geoffrey.levand@am.sony.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This is a port of the function graph tracer that was written by
Frederic Weisbecker for the x86.
This only works for PPC64 at the moment and only for static tracing.
PPC32 and dynamic function graph tracing support will come later.
The trace produces a visual calling of functions:
# tracer: function_graph
#
# CPU DURATION FUNCTION CALLS
# | | | | | | |
0) 2.224 us | }
0) ! 271.024 us | }
0) ! 320.080 us | }
0) ! 324.656 us | }
0) ! 329.136 us | }
0) | .put_prev_task_fair() {
0) | .update_curr() {
0) 2.240 us | .update_min_vruntime();
0) 6.512 us | }
0) 2.528 us | .__enqueue_entity();
0) + 15.536 us | }
0) | .pick_next_task_fair() {
0) 2.032 us | .__pick_next_entity();
0) 2.064 us | .__clear_buddies();
0) | .set_next_entity() {
0) 2.672 us | .__dequeue_entity();
0) 6.864 us | }
Geoff Lavand tested on PS3.
Tested-by: Geoff Levand <geoffrey.levand@am.sony.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Because 64-bit powerpc uses lazy (soft) interrupt disabling, it is
possible for a performance monitor exception to come in when the
kernel thinks interrupts are disabled (i.e. when they are
soft-disabled but hard-enabled). In such a situation the performance
monitor exception handler might have some processing to do (such as
process wakeups) which can't be done in what is effectively an NMI
handler.
This provides a way to defer that work until interrupts get enabled,
either in raw_local_irq_restore() or by returning from an interrupt
handler to code that had interrupts enabled. We have a per-processor
flag that indicates that there is work pending to do when interrupts
subsequently get re-enabled. This flag is checked in the interrupt
return path and in raw_local_irq_restore(), and if it is set,
perf_counter_do_pending() is called to do the pending work.
Signed-off-by: Paul Mackerras <paulus@samba.org>
It turns out that on Cell, on a kernel with CONFIG_VIRT_CPU_ACCOUNTING
= y, if a program sets the SO (summary overflow) bit in the XER and
then does a system call, the SO bit in CR0 will be set on return
regardless of whether the system call detected an error. Since CR0.SO
is used as the error indication from the system call, this means that
all system calls appear to fail.
The reason is that the workaround for the timebase bug on Cell uses a
compare instruction. With CONFIG_VIRT_CPU_ACCOUNTING = y, the
ACCOUNT_CPU_USER_ENTRY macro reads the timebase, so we end up doing a
compare instruction, which copies XER.SO to CR0.SO. Since we were
doing this in the system call entry patch after clearing CR0.SO but
before saving the CR, this meant that the saved CR image had CR0.SO
set if XER.SO was set on entry.
This fixes it by moving the clearing of CR0.SO to after the
ACCOUNT_CPU_USER_ENTRY call in the system call entry path.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Impact: quicken mcount calls that are not replaced by dyn ftrace
Dynamic ftrace no longer does on the fly recording of mcount locations.
The mcount locations are now found at compile time. The mcount
function no longer needs to store registers and call a stub function.
It can now just simply return.
Since there are some functions that do not get converted to a nop
(.init sections and other code that may disappear), this patch should
help speed up that code.
Also, the stub for mcount on PowerPC 32 can not be a simple branch
link register like it is on PowerPC 64. According to the ABI specification:
"The _mcount routine is required to restore the link register from
the stack so that the profiling code can be inserted transparently,
whether or not the profiled function saves the link register itself."
This means that we must restore the link register that was used
to make the call to mcount. The minimal mcount function for PPC32
ends up being:
mcount:
mflr r0
mtctr r0
lwz r0, 4(r1)
mtlr r0
bctr
Where we move the link register used to call mcount into the
ctr register, and then restore the link register from the stack.
Then we use the ctr register to jump back to the mcount caller.
The r0 register is free for us to use.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Due to confusion between the ftrace infrastructure and the gcc profiling
tracer "ftrace", this patch renames the config options from FTRACE to
FUNCTION_TRACER. The other two names that are offspring from FTRACE
DYNAMIC_FTRACE and FTRACE_MCOUNT_RECORD will stay the same.
This patch was generated mostly by script, and partially by hand.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Using LOAD_REG_IMMEDIATE to get the address of kernel symbols
generates 5 instructions where LOAD_REG_ADDR can do it in one,
and will generate R_PPC64_ADDR16_* relocations in the output when
we get to making the kernel as a position-independent executable,
which we'd rather not have to handle. This changes various bits
of assembly code to use LOAD_REG_ADDR when we need to get the
address of a symbol, or to use suitable position-independent code
for cases where we can't access the TOC for various reasons, or
if we're not running at the address we were linked at.
It also cleans up a few minor things; there's no reason to save and
restore SRR0/1 around RTAS calls, __mmu_off can get the return
address from LR more conveniently than the caller can supply it in
R4 (and we already assume elsewhere that EA == RA if the MMU is on
in early boot), and enable_64b_mode was using 5 instructions where
2 would do.
Signed-off-by: Paul Mackerras <paulus@samba.org>
There is a small passage of code in ret_from_except_lite which is
only required on iSeries. For a multi-platform kernel on non-iSeries
machines this means we end up executing ~15 nops in ret_from_except_lite.
It would be nicer if non-iSeries could skip the code entirely, and on
iSeries we can jump out of line to execute the code.
I have no performance numbers to justify this, other than the assertion
that executing 15 nops takes longer than executing 0.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This adds TIF_NOTIFY_RESUME support for powerpc. When set,
we call tracehook_notify_resume() on the way to user mode.
This overloads do_signal() to do the work, but changes its
arguments to it has the TIF_* bits handy in a register and
drops the useless first argument that was always zero.
Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This changes powerpc syscall tracing to use the new tracehook.h entry
points. There is no change, only cleanup.
In addition, the assembly changes allow do_syscall_trace_enter() to
abort the syscall without losing the information about the original
r0 value.
Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This patch extends the floating point save and restore code to use the
VSX load/stores when VSX is available. This will make FP context
save/restore marginally slower on FP only code, when VSX is available,
as it has to load/store 128bits rather than just 64bits.
Mixing FP, VMX and VSX code will get constant architected state.
The signals interface is extended to enable access to VSR 0-31
doubleword 1 after discussions with tool chain maintainers. Backward
compatibility is maintained.
The ptrace interface is also extended to allow access to VSR 0-31 full
registers.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Use an alternative feature section in _switch. There are three cases
handled here, either we don't have an SLB, in which case we jump over the
entire code section, or if we do we either do or don't have 1TB segments.
Boot tested on Power3, Power5 and Power5+.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Record the address of the mcount call-site. Currently all archs except sparc64
record the address of the instruction following the mcount call-site. Some
general cleanups are entailed. Storing mcount addresses in rec->ip enables
looking them up in the kprobe hash table later on to check if they're kprobe'd.
Signed-off-by: Abhishek Sagar <sagar.abhishek@gmail.com>
Cc: davem@davemloft.net
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
This patch adds full support for ftrace for PowerPC (both 64 and 32 bit).
This includes dynamic tracing and function filtering.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This adds the low level irq tracing hooks to the powerpc architecture
needed to enable full lockdep functionality.
This is partly based on Johannes Berg's initial version. I removed
the asm trampoline that isn't needed (thus improving performance) and
modified all sorts of bits and pieces, reworking most of the assembly,
etc...
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This moves various definitions used all over the place to parse stack
frames to ptrace.h so only one definition is needed.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
PA6T has a bug where the slbie instruction does not honor the large
segment bit. As a result, we have to always use slbia when switching
context.
We don't have to worry about changing the slbie's during fault processing,
since they should never be replacing one VSID with another using the
same ESID. I.e. there's no risk for inserting duplicate entries due to a
failed slbie of the old entry. So as long as we clear it out on context
switch we should be fine.
Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This makes the kernel use 1TB segments for all kernel mappings and for
user addresses of 1TB and above, on machines which support them
(currently POWER5+, POWER6 and PA6T).
We detect that the machine supports 1TB segments by looking at the
ibm,processor-segment-sizes property in the device tree.
We don't currently use 1TB segments for user addresses < 1T, since
that would effectively prevent 32-bit processes from using huge pages
unless we also had a way to revert to using 256MB segments. That
would be possible but would involve extra complications (such as
keeping track of which segment size was used when HPTEs were inserted)
and is not addressed here.
Parts of this patch were originally written by Ben Herrenschmidt.
Signed-off-by: Paul Mackerras <paulus@samba.org>
After talking to an IBM POWER hypervisor (PHYP) design and development
guy, there seems to be no need for memory barriers when updating the SLB
shadow buffer provided we only update it from the current CPU, which we
do.
Also, these guys see no need in the future for these barriers.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
On a machine with hardware 64kB pages and a kernel configured for a
64kB base page size, we need to change the vmalloc segment from 64kB
pages to 4kB pages if some driver creates a non-cacheable mapping in
the vmalloc area. However, we never updated with SLB shadow buffer.
This fixes it. Thanks to paulus for finding this.
Also added some write barriers to ensure the shadow buffer contents
are always consistent.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Remove unused TIF_NOTIFY_RESUME flag for all processor architectures. The
flag was not used excecpt on IA-64 where the patch replaces it with
TIF_PERFMON_WORK.
Signed-off-by: stephane eranian <eranian@hpl.hp.com>
Cc: <linux-arch@vger.kernel.org>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Some instruction tracing tools use the RI (recoverable interrupt) bit
in the MSR to indicate when it's safe to single-step. Currently we
clear RI after restoring r13 when returning to userspace. However,
if we single-step past the point where r13 is restored, we'll corrupt
r13 in the exception entry code and not restore it. This moves the
clearing of RI to just before r13 is restored so this doesn't happen.
Signed-off-by: Paul Mackerras <paulus@samba.org>
When we switched over to the generic BUG mechanism we forgot to change
the assembly code which open-codes a WARN_ON() in enter_rtas(), so the
bug table got corrupted.
This patch provides an EMIT_BUG_ENTRY macro for use in assembly code,
and uses it in entry_64.S. Tested with CONFIG_DEBUG_BUGVERBOSE on ppc64
but not without -- I tried to turn it off but it wouldn't go away; I
suspect Aunt Tillie probably needed it.
This version gets __FILE__ and __LINE__ right in the assembly version --
rather than saying include/asm-powerpc/bug.h line 21 every time which is
a little suboptimal.
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Commit 3ccfc65c50 missed the same fixes for
legacy iSeries specific code, so make some more symbols no longer global.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Since iSeries is merged to MULTIPLATFORM, there is no way to build a 64bit
kernel without MULTIPLATFORM, so PPC_MULTIPLATFORM can be removed in
64bit-only files.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The lazy IRQ disable patch missed a couple of places where the
interrupt enable flags need to be restored correctly. First, we
weren't restoring the paca->hard_enabled flag on interrupt exit.
Instead of saving it on entry, we compute it from the MSR_EE bit
in the MSR we are restoring at exit. Secondly, the MMU hash miss
code was clearing both paca->soft_enabled and paca->hard_enabled
but not restoring them in the case where hash_page was able to
resolve the miss from the Linux page tables.
Signed-off-by: Paul Mackerras <paulus@samba.org>
This implements a lazy strategy for disabling interrupts. This means
that local_irq_disable() et al. just clear the 'interrupts are
enabled' flag in the paca. If an interrupt comes along, the interrupt
entry code notices that interrupts are supposed to be disabled, and
clears the EE bit in SRR1, clears the 'interrupts are hard-enabled'
flag in the paca, and returns. This means that interrupts only
actually get disabled in the processor when an interrupt comes along.
When interrupts are enabled by local_irq_enable() et al., the code
sets the interrupts-enabled flag in the paca, and then checks whether
interrupts got hard-disabled. If so, it also sets the EE bit in the
MSR to hard-enable the interrupts.
This has the potential to improve performance, and also makes it
easier to make a kernel that can boot on iSeries and on other 64-bit
machines, since this lazy-disable strategy is very similar to the
soft-disable strategy that iSeries already uses.
This version renames paca->proc_enabled to paca->soft_enabled, and
changes a couple of soft-disables in the kexec code to hard-disables,
which should fix the crash that Michael Ellerman saw. This doesn't
yet use a reserved CR field for the soft_enabled and hard_enabled
flags. This applies on top of Stephen Rothwell's patches to make it
possible to build a combined iSeries/other kernel.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Cleanup some of the #define magic as suggested by Milton.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This adds a shadow buffer for the SLBs and regsiters it with PHYP.
Only the bolted SLB entries (top 3) are shadowed.
The SLB shadow buffer tells the hypervisor what the kernel needs to
have in the SLB for the kernel to be able to function. The hypervisor
can use this information to speed up partition context switches.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
In the syscall path we currently have:
crclr so
mfcr r9
If we shift the crclr up we can avoid a stall on some CPUs.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
A bug in the RTAS services incorrectly interprets some bits in the CR
when called from the OS. Specifically, bits in CR4. The result could
be a firmware crash that also takes down the partition. A firmware
fix is in the works. We have seen this situation when performing DLPAR
operations. As a temporary workaround, clear the CR in enter_rtas().
Note that enter_rtas() will not set any bits in CR4 before calling RTAS.
Also note that the 32 bit version of enter_rtas() should have the same
work around even though the chances of hitting the bug are much smaller
due to the lack of DLPAR on 32 bit kernels. However, my assembly skills
are a bit rusty and the 32 bit code doesn't seem to follow the conventions
for where things should be saved. In addition, I don't have a system
to test 32 bit kernels. Help creating and at least touch testing the
same workaround for 32 bit would be appreciated.
Signed-off-by: Mike Kravetz <kravetz@us.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
A careful reading of the recent changes to the system call entry/exit
paths revealed several problems, plus some things that could be
simplified and improved:
* 32-bit wasn't testing the _TIF_NOERROR bit in the syscall fast exit
path, so it was only doing anything with it once it saw some other
bit being set. In other words, the noerror behaviour would apply to
the next system call where we had to reschedule or deliver a signal,
which is not necessarily the current system call.
* 32-bit wasn't doing the call to ptrace_notify in the syscall exit
path when the _TIF_SINGLESTEP bit was set.
* _TIF_RESTOREALL was in both _TIF_USER_WORK_MASK and
_TIF_PERSYSCALL_MASK, which is odd since _TIF_RESTOREALL is only set
by system calls. I took it out of _TIF_USER_WORK_MASK.
* On 64-bit, _TIF_RESTOREALL wasn't causing the non-volatile registers
to be restored (unless perhaps a signal was delivered or the syscall
was traced or single-stepped). Thus the non-volatile registers
weren't restored on exit from a signal handler. We probably got
away with it mostly because signal handlers written in C wouldn't
alter the non-volatile registers.
* On 32-bit I simplified the code and made it more like 64-bit by
making the syscall exit path jump to ret_from_except to handle
preemption and signal delivery.
* 32-bit was calling do_signal unnecessarily when _TIF_RESTOREALL was
set - but I think because of that 32-bit was actually restoring the
non-volatile registers on exit from a signal handler.
* I changed the order of enabling interrupts and saving the
non-volatile registers before calling do_syscall_trace_leave; now we
enable interrupts first.
Signed-off-by: Paul Mackerras <paulus@samba.org>
This implements accurate task and cpu time accounting for 64-bit
powerpc kernels. Instead of accounting a whole jiffy of time to a
task on a timer interrupt because that task happened to be running at
the time, we now account time in units of timebase ticks according to
the actual time spent by the task in user mode and kernel mode. We
also count the time spent processing hardware and software interrupts
accurately. This is conditional on CONFIG_VIRT_CPU_ACCOUNTING. If
that is not set, we do tick-based approximate accounting as before.
To get this accurate information, we read either the PURR (processor
utilization of resources register) on POWER5 machines, or the timebase
on other machines on
* each entry to the kernel from usermode
* each exit to usermode
* transitions between process context, hard irq context and soft irq
context in kernel mode
* context switches.
On POWER5 systems with shared-processor logical partitioning we also
read both the PURR and the timebase at each timer interrupt and
context switch in order to determine how much time has been taken by
the hypervisor to run other partitions ("steal" time). Unfortunately,
since we need values of the PURR on both threads at the same time to
accurately calculate the steal time, and since we can only calculate
steal time on a per-core basis, the apportioning of the steal time
between idle time (time which we ceded to the hypervisor in the idle
loop) and actual stolen time is somewhat approximate at the moment.
This is all based quite heavily on what s390 does, and it uses the
generic interfaces that were added by the s390 developers,
i.e. account_system_time(), account_user_time(), etc.
This patch doesn't add any new interfaces between the kernel and
userspace, and doesn't change the units in which time is reported to
userspace by things such as /proc/stat, /proc/<pid>/stat, getrusage(),
times(), etc. Internally the various task and cpu times are stored in
timebase units, but they are converted to USER_HZ units (1/100th of a
second) when reported to userspace. Some precision is therefore lost
but there should not be any accumulating error, since the internal
accumulation is at full precision.
Signed-off-by: Paul Mackerras <paulus@samba.org>
This patch removes all self references and fixes references to files
in the now defunct arch/ppc64 tree. I think this accomplises
everything wanted, though there might be a few references I missed.
Signed-off-by: Jon Mason <jdmason@us.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Implement the TIF_RESTORE_SIGMASK flag in the new arch/powerpc kernel, for
both 32-bit and 64-bit system call paths.
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
At present the lppaca - the structure shared with the iSeries
hypervisor and phyp - is contained within the PACA, our own low-level
per-cpu structure. This doesn't have to be so, the patch below
removes it, making a separate array of lppaca structures.
This saves approximately 500*NR_CPUS bytes of image size and kernel
memory, because we don't need aligning gap between the Linux and
hypervisor portions of every PACA. On the other hand it means an
extra level of dereference in many accesses to the lppaca.
The patch also gets rid of several places where we assign the paca
address to a local variable for no particular reason.
Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This patch consolidates the variety of macros used for loading 32 or
64-bit constants in assembler (LOADADDR, LOADBASE, SET_REG_TO_*). The
idea is to make the set of macros consistent across 32 and 64 bit and
to make it more obvious which is the appropriate one to use in a given
situation. The new macros and their semantics are described in the
comments in ppc_asm.h.
In the process, we change several places that were unnecessarily using
immediate loads on ppc64 to use the GOT/TOC. Likewise we cleanup a
couple of places where we were clumsily subtracting PAGE_OFFSET with
asm instructions to use assemble-time arithmetic or the toreal() macro
instead.
Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This patch separates usage of KERNELBASE and PAGE_OFFSET. I haven't
looked at any of the PPC32 code, if we ever want to support Kdump on
PPC we'll have to do another audit, ditto for iSeries.
This patch makes PAGE_OFFSET the constant, it'll always be 0xC * 1
gazillion for 64-bit.
To get a physical address from a virtual one you subtract PAGE_OFFSET,
_not_ KERNELBASE.
KERNELBASE is the virtual address of the start of the kernel, it's
often the same as PAGE_OFFSET, but _might not be_.
If you want to know something's offset from the start of the kernel
you should subtract KERNELBASE.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>
On Tue, 2005-11-15 at 18:52 +0000, David Woodhouse wrote:
> This cleanup patch speeds up the null syscall path on ppc64 by about 3%,
> and brings the ppc32 and ppc64 code slightly closer together.
Needs this unless your binutils, like mine, are clever enough to notice
my stupidity and fix it up automatically...
Spotted by Paul.
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This cleanup patch speeds up the null syscall path on ppc64 by about 3%,
and brings the ppc32 and ppc64 code slightly closer together.
The ppc64 code was checking current_thread_info()->flags twice in the
syscall exit path; once for TIF_SYSCALL_T_OR_A before disabling
interrupts, and then again for TIF_SIGPENDING|TIF_NEED_RESCHED etc after
disabling interrupts. Now we do the same as ppc32 -- check the flags
only once in the fast path, and re-enable interrupts if necessary in the
ptrace case.
The patch abolishes the 'syscall_noerror' member of struct thread_info
and replaces it with a TIF_NOERROR bit in the flags, which is handled in
the slow path. This shortens the syscall entry code, which no longer
needs to clear syscall_noerror.
The patch adds a TIF_SAVE_NVGPRS flag which causes the syscall exit slow
path to save the non-volatile GPRs into a signal frame. This removes the
need for the assembly wrappers around sys_sigsuspend(),
sys_rt_sigsuspend(), et al which existed solely to save those registers
in advance. It also means I don't have to add new wrappers for ppoll()
and pselect(), which is what I was supposed to be doing when I got
distracted into this...
Finally, it unifies the ppc64 and ppc32 methods of handling syscall exit
directly into a signal handler (as required by sigsuspend et al) by
introducing a TIF_RESTOREALL flag which causes _all_ the registers to be
reloaded from the pt_regs by taking the ret_from_exception path, instead
of the normal syscall exit path which stomps on the callee-saved GPRs.
It appears to pass an LTP test run on ppc64, and passes basic testing on
ppc32 too. Brief tests of ptrace functionality with strace and gdb also
appear OK. I wouldn't send it to Linus for 2.6.15 just yet though :)
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Since we don't restore the volatile registers in the syscall exit
path, we need to make sure we don't leak any potentially interesting
values from the kernel to userspace. This was already the case for
all except r11. This makes it use r11 for an MSR value, so r11 will
have an (uninteresting) MSR value in it on return to userspace.
Signed-off-by: Paul Mackerras <paulus@samba.org>
This section of code calls .audit_syscal_exit, but is inside CONFIG_AUDIT,
so it will fail to build if CONFIG_AUDITSYSCALL is not defined.
After discussion with David Woodhouse, change the ifdef to
CONFIG_AUDITSYSCALL
Signed-off-by: Horms <horms@verge.net.au>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This allows us to get rid of one type of entry in systbl.S.
In passing we remove the duplicate compat_sys_getdents and
compat_sys_utimes for which there are generic versions.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
The system call table has been consolidated into systbl.S. We have
separate 32-bit and 64-bit versions of entry.S and misc.S since the
code is mostly sufficiently different to be not worth merging.
There are some common bits that will be extracted in future.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Anton Blanchard