If we are using 1TB segments and we are allowed to randomise the heap, we can
put it above 1TB so it is backed by a 1TB segment. Otherwise the heap will be
in the bottom 1TB which always uses 256MB segments and this may result in a
performance penalty.
This functionality is disabled when heap randomisation is turned off:
echo 1 > /proc/sys/kernel/randomize_va_space
which may be useful when trying to allocate the maximum amount of 16M or 16G
pages.
On a microbenchmark that repeatedly touches 32GB of memory with a stride of
256MB + 4kB (designed to stress 256MB segments while still mapping nicely into
the L1 cache), we see the improvement:
Force malloc to use heap all the time:
# export MALLOC_MMAP_MAX_=0 MALLOC_TRIM_THRESHOLD_=-1
Disable heap randomization:
# echo 1 > /proc/sys/kernel/randomize_va_space
# time ./test
12.51s
Enable heap randomization:
# echo 2 > /proc/sys/kernel/randomize_va_space
# time ./test
1.70s
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Currently, a single ifdef covers SLB related bits and more generic ppc64
related bits, split this in two separate ifdef's since 64-bit BookE will
need one but not the other.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
For some reason we've had an explicit KERN_INFO for GPR dumps. With
recent changes we get output like:
<6>GPR00: 00000000 ef855eb0 ef858000 00000001 000000d0 f1000000 ffbc8000 ffffffff
The KERN_INFO is causing the <6>. Don't see any reason to keep it
around.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This is a random collection of added ifdef's around portions of
code that only mak sense on server processors. Using either
CONFIG_PPC_STD_MMU_64 or CONFIG_PPC_BOOK3S as seems appropriate.
This is meant to make the future merging of Book3E 64-bit support
easier.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Randomise ELF_ET_DYN_BASE, which is used when loading position independent
executables.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Randomise the lower bits of the stack address. More randomisation is good for
security but the scatter can also help with SMT threads that share an L1. A
quick test case shows this working:
int main()
{
int sp;
printf("%x\n", (unsigned long)&sp & 4095);
}
before:
80
80
80
80
80
after:
610
490
300
6b0
d80
Signed-off-by: Anton Blanchard <anton@samba.org>
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>
The cpu time spent by the idle process actually doing something is
currently accounted as idle time. This is plain wrong, the architectures
that support VIRT_CPU_ACCOUNTING=y can do better: distinguish between the
time spent doing nothing and the time spent by idle doing work. The first
is accounted with account_idle_time and the second with account_system_time.
The architectures that use the account_xxx_time interface directly and not
the account_xxx_ticks interface now need to do the check for the idle
process in their arch code. In particular to improve the system vs true
idle time accounting the arch code needs to measure the true idle time
instead of just testing for the idle process.
To improve the tick based accounting as well we would need an architecture
primitive that can tell us if the pt_regs of the interrupted context
points to the magic instruction that halts the cpu.
In addition idle time is no more added to the stime of the idle process.
This field now contains the system time of the idle process as it should
be. On systems without VIRT_CPU_ACCOUNTING this will always be zero as
every tick that occurs while idle is running will be accounted as idle
time.
This patch contains the necessary common code changes to be able to
distinguish idle system time and true idle time. The architectures with
support for VIRT_CPU_ACCOUNTING need some changes to exploit this.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
On my screen, when something crashes, I only have space for maybe 16
functions of the stack trace before the information above it scrolls
off the screen. It's easy to hack the kernel to print out only that
much, but it's harder to remember to do it. This introduces a config
option for it so that I can keep the setting in my config.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Now that arch/ppc is gone and CONFIG_PPC_MERGE is always set, remove
the dead code associated with !CONFIG_PPC_MERGE from arch/powerpc
and include/asm-powerpc.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
* CONFIG_BOOKE is selected by CONFIG_44x so we dont need both
* Fixed a few comments
* Go back to only using DBCR0_IDM to determine if we are using
debug resources.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This patch implements support for HW based watchpoint via the
DBSR_DAC (Data Address Compare) facility of the BookE processors.
It does so by interfacing with the existing DABR breakpoint code
and adding the necessary bits and pieces for the new bits to
be properly set or cleared
Signed-off-by: Luis Machado <luisgpm@br.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
giveup_vsx didn't save the FPU and VMX regsiters. Change it to be
like giveup_fpr/altivec which save these registers.
Also update call sites where FPU and VMX are already saved to use the
original giveup_vsx (renamed to __giveup_vsx).
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This changes the oops and backtrace code to use the new %pS
printk extension to print out symbols rather than manually
calling print_symbol.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Since Roland's ptrace cleanup starting with commit
f65255e8d5 ("[POWERPC] Use user_regset
accessors for FP regs"), the dump_task_* functions are no longer being
used.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This correctly hooks the VSX dump into Roland McGrath core file
infrastructure. It adds the VSX dump information as an additional elf
note in the core file (after talking more to the tool chain/gdb guys).
This also ensures the formats are consistent between signals, ptrace
and core files.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.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>
We are going to change where the floating point registers are stored
in the thread_struct, so in preparation add some macros to access the
floating point registers. Update all code to use these new macros.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This makes it possible to use separate stacks for hard and soft IRQs
on 32-bit powerpc as well as on 64-bit. The code for 32-bit is just
the 32-bit analog of the 64-bit code.
* Added allocation and initialization of the irq stacks. We limit the
stacks to be in lowmem for ppc32.
* Implemented ppc32 versions of call_do_softirq() and call_handle_irq()
to switch the stack pointers
* Reworked how we do stack overflow detection. We now keep around the
limit of the stack in the thread_struct and compare against the limit
to see if we've overflowed. We can now use this on ppc64 if desired.
[ paulus@samba.org: Fixed bug on 6xx where we need to reload r9 with the
thread_info pointer. ]
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The powerpc kernel stacks need to be naturally aligned, as they
contain the thread info at the bottom, which is obtained by
clearing the low bits of the stack pointer.
However, when using 64K pages, the stack is smaller than a page,
so we use kmalloc to allocate it, but that doesn't provide the
alignment guarantee we need.
It appeared to work so far... until one enables SLUB debugging
which then returns unaligned pointers. Ooops...
This fixes it by using a slab cache with enforced alignment. It
relies on my previous patch that adds a thread_info_cache_init()
callback.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Andrew Morton <akpm@linux-foundation.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>
There is a bug in the powerpc DABR (data access breakpoint) handling,
which can result in us missing breakpoints if several threads are trying
to break on the same address.
The circumstances are that do_page_fault() calls do_dabr(), this clears
the DABR (sets it to 0) and sets up the signal which will report to
userspace that the DABR was hit. The do_signal() code will restore the DABR
value on the way out to userspace.
If we reschedule before calling do_signal(), __switch_to() will check the
cached DABR value and compare it to the new thread's value, if they match
we don't set the DABR in hardware.
So if two threads have the same DABR value, and we schedule from one to
the other after taking the interrupt for the first thread hitting the DABR,
the second thread will run without the DABR set in hardware.
The cleanest fix is to move the cache update into set_dabr(), that way we
can't forget to do it.
Reported-by: Jan Kratochvil <jan.kratochvil@redhat.com>
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The PT_DTRACE flag is meaningless and obsolete.
Don't touch it.
Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Since the PMU is an NMI now, it can come at any time we are only soft
disabled. We must hard disable around the two places we allow the kernel
stack SLB and r1 to go out of sync. Otherwise the PMU exception can
force a kernel stack SLB into another slot, which can lead to it
getting evicted, which can lead to a nasty unrecoverable SLB miss
in the exception entry code.
Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The powerpc show_regs prints CPU using smp_processor_id: change that to
raw_smp_processor_id, so that when it's showing a WARN_ON backtrace without
preemption disabled, DEBUG_PREEMPT doesn't mess up that warning with its own.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The commit fa13a5a1f2 (sched: restore
deterministic CPU accounting on powerpc), unconditionally calls
update_process_tick() in system context. In the deterministic
accounting case this is the correct thing to do. However, in the
non-deterministic accounting case we need to not do this, since doing
this results in the time accounted as hardware irq time being
artificially elevated.
Also this collapses 2 consecutive '#ifdef CONFIG_VIRT_CPU_ACCOUNTING'
checks in time.h into one for neatness.
Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Since powerpc started using CONFIG_GENERIC_CLOCKEVENTS, the
deterministic CPU accounting (CONFIG_VIRT_CPU_ACCOUNTING) has been
broken on powerpc, because we end up counting user time twice: once in
timer_interrupt() and once in update_process_times().
This fixes the problem by pulling the code in update_process_times
that updates utime and stime into a separate function called
account_process_tick. If CONFIG_VIRT_CPU_ACCOUNTING is not defined,
there is a version of account_process_tick in kernel/timer.c that
simply accounts a whole tick to either utime or stime as before. If
CONFIG_VIRT_CPU_ACCOUNTING is defined, then arch code gets to
implement account_process_tick.
This also lets us simplify the s390 code a bit; it means that the s390
timer interrupt can now call update_process_times even when
CONFIG_VIRT_CPU_ACCOUNTING is turned on, and can just implement a
suitable account_process_tick().
account_process_tick() now takes the task_struct * as an argument.
Tested both with and without CONFIG_VIRT_CPU_ACCOUNTING.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
One of the easiest things to isolate is the pid printed in kernel log.
There was a patch, that made this for arch-independent code, this one makes
so for arch/xxx files.
It took some time to cross-compile it, but hopefully these are all the
printks in arch code.
Signed-off-by: Alexey Dobriyan <adobriyan@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Cc: <linux-arch@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Update dump_task_altivec() (which has so far never been put to use) so that
it dumps the Altivec/VMX registers (VR[0] - VR[31], VSCR and VRSAVE) in the
same format as the ptrace get_vrregs(), and add the appropriate glue
typedef and #defines to make it work.
A new note type of NT_PPC_VMX was chosen to be 0x100 (arbitrarily) because
it allows the low range values to be used for more generic purposes and
0x100 seems an adequate starting point for PowerPC extensions.
Signed-off-by: Mark Nelson <markn@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Andi Kleen <ak@suse.de>
Cc: <linux-arch@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.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>
On non-book-E, exceptions execute in real mode. If a fault happens
that leads to a register dump, the kernel currently prints XXXXXXXX
because it doesn't realize that PC is a physical address.
This patch checks whether instruction address translation is turned
on, and if not converts PC into a virtual address.
Signed-off-by: Scott Wood <scottwood@freescale.com>
Acked-by: Kumar Gala <galak@kernel.crashing.org>
Acked-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
When PTRACE_O_TRACEEXEC is used, a ptrace call to fetch the registers at
the PTRACE_EVENT_EXEC stop (PTRACE_PEEKUSR) will oops in CHECK_FULL_REGS.
With recent versions, "gdb --args /bin/sh -c 'exec /bin/true'" and "run" at
the (gdb) prompt is sufficient to produce this. I also have written an
isolated test case, see https://bugzilla.redhat.com/show_bug.cgi?id=301791#c15.
This change fixes the problem by clearing the low bit of pt_regs.trap in
start_thread so that FULL_REGS is true again. This is correct since all of
the GPRs that "full" refers to are cleared in start_thread.
Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Make it so that SPE support can be determined at runtime. This is similiar
to how we handle AltiVec. This allows us to have SPE support built in and
work on processors with and without SPE.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
When we flush register state for FP, Altivec, or SPE in flush_*_to_thread
we need to respect the task_struct that the caller has passed to us.
Most cases we are called with current, however sometimes (ptrace) we may
be passed a different task_struct.
This showed up when using gdbserver debugging a simple program that used
floating point. When gdb tried to show the FP regs they all showed up as
0, because the child's FP registers were never properly flushed to memory.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
In a show_regs() message The DEAR and ESR were reported as
DAR and DSISR which only exist on classic parts.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
This patch removes the #ifdef CONFIG_PPC64 around setting the DABR.
The actual setting of the SPR inside of the set_dabr() function is dependent
on CONFIG_PPC64 || CONFIG_6xx but you can always provide a ppc_md hook to
override that. We should improve support for different HW breakpoints
facilities but this is a first step.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Remove includes of <linux/smp_lock.h> where it is not used/needed.
Suggested by Al Viro.
Builds cleanly on x86_64, i386, alpha, ia64, powerpc, sparc,
sparc64, and arm (all 59 defconfigs).
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The current tlb flush code on powerpc 64 bits has a subtle race since we
lost the page table lock due to the possible faulting in of new PTEs
after a previous one has been removed but before the corresponding hash
entry has been evicted, which can leads to all sort of fatal problems.
This patch reworks the batch code completely. It doesn't use the mmu_gather
stuff anymore. Instead, we use the lazy mmu hooks that were added by the
paravirt code. They have the nice property that the enter/leave lazy mmu
mode pair is always fully contained by the PTE lock for a given range
of PTEs. Thus we can guarantee that all batches are flushed on a given
CPU before it drops that lock.
We also generalize batching for any PTE update that require a flush.
Batching is now enabled on a CPU by arch_enter_lazy_mmu_mode() and
disabled by arch_leave_lazy_mmu_mode(). The code epects that this is
always contained within a PTE lock section so no preemption can happen
and no PTE insertion in that range from another CPU. When batching
is enabled on a CPU, every PTE updates that need a hash flush will
use the batch for that flush.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Use lowercase for hex printouts in oops messages. The number of times I have
tried to copy and paste from an oops into an objdump search...
Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Remove last_syscall from 32bit powerpc, its been gone in 64bit for years.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Fix atomicity of TIF update in flush_thread() for powerpc
Fixes it correctly with *_ti_thread_flag.
Race :
parent process executing :
sys_ptrace()
(lock_kernel())
(ptrace_get_task_struct(pid))
arch_ptrace()
ptrace_detach()
ptrace_disable(child);
clear_singlestep(child);
clear_tsk_thread_flag(child, TIF_SINGLESTEP);
(which clears the TIF_SINGLESTEP flag atomically from a different
process)
(put_task_struct(child))
(unlock_kernel())
And at the same time, in the child process :
sys_execve()
do_execve()
search_binary_handler()
load_elf_binary()
flush_old_exec()
flush_thread()
doing a non-atomic thread flag update
Applies on 2.6.20.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Paul Mackerras <paulus@samba.org>
If something has overflowed or corrupted the stack and causes an oops,
and we try to print a stack trace, that will call validate_sp, which
can itself cause an oops if the cpu field of the thread_info struct at
the bottom of the stack has been corrupted (if CONFIG_IRQSTACKS is
set). This makes debugging harder.
To avoid the second oops, this adds a check to make sure that the cpu
number is reasonable before using it to check whether the stack is on
the softirq or hardirq stack.
Signed-off-by: Paul Mackerras <paulus@samba.org>
