This is an attempt at cleaning up a bit the way we handle execute
permission on powerpc. _PAGE_HWEXEC is gone, _PAGE_EXEC is now only
defined by CPUs that can do something with it, and the myriad of
#ifdef's in the I$/D$ coherency code is reduced to 2 cases that
hopefully should cover everything.
The logic on BookE is a little bit different than what it was though
not by much. Since now, _PAGE_EXEC will be set by the generic code
for executable pages, we need to filter out if they are unclean and
recover it. However, I don't expect the code to be more bloated than
it already was in that area due to that change.
I could boast that this brings proper enforcing of per-page execute
permissions to all BookE and 40x but in fact, we've had that now for
some time as a side effect of my previous rework in that area (and
I didn't even know it :-) We would only enable execute permission if
the page was cache clean and we would only cache clean it if we took
and exec fault. Since we now enforce that the later only work if
VM_EXEC is part of the VMA flags, we de-fact already enforce per-page
execute permissions... Unless I missed something
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The MMUCSR is now defined as part of the Book-3E architecture so we
can move it into mmu-book3e.h and add some of the additional bits
defined by the architecture specs.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
hardirq.h on powerpc defines a __last_jiffy_stamp field, but it's not
actually used anywhere.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
As <asm/iommu.h> doesn't contain any other hardware specific definitions
but only interfaces.
Reported-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The mask used to encode the page table cache number in the
batch when freeing page tables was too small for the new
possible values of MMU page sizes. This increases it along
with a comment explaining the constraints.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
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>
The base TLB support didn't include support for SPARSEMEM_VMEMMAP, though
we did carve out some virtual space for it, the necessary support code
wasn't there. This implements it by using 16M pages for now, though the
page size could easily be changed at runtime if necessary.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This adds the TLB miss handler assembly, the low level TLB flush routines
along with the necessary hook for dealing with our virtual page tables
or indirect TLB entries that need to be flushes when PTE pages are freed.
There is currently no support for hugetlbfs
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This adds various fields in the PACA that are for use specifically
by Book3E processors, such as exception save areas, current pgd
pointer, special exceptions kernel stacks etc...
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This adds various definitions and macros used by the exception and TLB
miss handling on 64-bit BookE
It also adds the definitions of the SPRGs used for various exception types
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This adds the PTE and pgtable format definitions, along with changes
to the kernel memory map and other definitions related to implementing
support for 64-bit Book3E. This also shields some asm-offset bits that
are currently only relevant on 32-bit
We also move the definition of the "linux" page size constants to
the common mmu.h file and add a few sizes that are relevant to
embedded processors.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This adds various SPRs defined on 64-bit BookE, along with changes
to the definition of the base MSR values to add the values needed
for 64-bit Book3E.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
That patch used to just add a hook to page table flushing but
pulling that string brought out a whole bunch of issues, so it
now does that and more:
- We now make the RCU batching of page freeing SMP only, as I
believe it was intended initially. We make a few more things compile
to nothing on !CONFIG_SMP
- Some macros are turned into functions, though that forced me to
out of line a few stuffs due to unsolvable include depenencies,
however it's probably better that way anyway, it's not -that-
critical code path.
- 32-bit didn't call pte_free_finish() on tlb_flush() which means
that it wouldn't push out the batch to RCU for delayed freeing when
a bunch of page tables have been freed, they would just stay in there
until the batch gets full.
64-bit BookE will use that hook to maintain the virtually linear
page tables or the indirect entries in the TLB when using the
HW loader.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Those definitions are currently declared extern in the .c file where
they are used, move them to a header file instead.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Our 64-bit hash context handling has no init function, but 64-bit Book3E
will use the common mmu_context_nohash.c code which does, so define an
empty inline mmu_context_init() for 64-bit server and call it from
our 64-bit setup_arch()
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Kumar Gala <galak@kernel.crashing.org>
We need to pass down whether the page is direct or indirect and we'll
need to pass the page size to _tlbil_va and _tlbivax_bcast
We also add a new low level _tlbil_pid_noind() which does a TLB flush
by PID but avoids flushing indirect entries if possible
This implements those new prototypes but defines them with inlines
or macros so that no additional arguments are actually passed on current
processors.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The way I intend to use tophys/tovirt on 64-bit BookE is different
from the "trick" that we currently play for 32-bit BookE so change
the condition of definition of these macros to make it so.
Also, make sure we only use rfid and mtmsrd instead of rfi and mtmsr
for 64-bit server processors, not all 64-bit processors.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Kumar Gala <galak@kernel.crashing.org>
This adds various additional bit definitions for various MMU related
SPRs used on Book3E.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This adds the opcode definitions to ppc-opcode.h for the two instructions
tlbivax and tlbsrx. as defined by Book3E 2.06
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The current "no hash" MMU context management code is written with
the assumption that one CPU == one TLB. This is not the case on
implementations that support HW multithreading, where several
linux CPUs can share the same TLB.
This adds some basic support for this to our context management
and our TLB flushing code.
It also cleans up the optional debugging output a bit
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
A misplaced #endif causes more definitions than intended to be
protected by #ifndef __ASSEMBLY__. This breaks upcoming 64-bit
BookE support patch when using 64k pages.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The truncate syscall has a signed long parameter, so when using a 32-
bit userspace with a 64-bit kernel the argument is zero-extended
instead of sign-extended. Adding the compat_sys_truncate function
fixes the issue.
This was noticed during an LSB truncate test failure. The test was
checking for the correct error number set when truncate is called with
a length of -1. The test can be found at:
http://bzr.linuxfoundation.org/lsb/devel/runtime-test?cmd=inventory;rev=stewb%40linux-foundation.org-20090626205411-sfb23cc0tjj7jzgm;path=modules/vsx-pcts/tset/POSIX.os/files/truncate/
BenH: Added compat_sys_ftruncate() as well, same issue.
Signed-off-by: Chase Douglas <cndougla@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This change the SPRG used to store the PACA on ppc64 from
SPRG3 to SPRG1. SPRG3 is user readable on most processors
and we want to use it for other things. We change the scratch
SPRG used by exception vectors from SRPG1 to SPRG2.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The STAB code used on Power3 and RS/64 uses a second scratch SPRG to
save a GPR in order to decide whether to go to do_stab_bolted_* or
to handle a normal data access exception.
This prevents our scheme of freeing SPRG3 which is user visible for
user uses since we cannot use SPRG0 which, on RS/64, seems to be
read-only for supervisor mode (like POWER4).
This reworks the STAB exception entry to use the PACA as temporary
storage instead.
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>
The file include/asm/exception.h contains definitions
that are specific to exception handling on 64-bit server
type processors.
This renames the file to exception-64s.h to reflect that
fact and avoid confusion.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
On 64bit applications the VDSO is the only thing in segment 0. Since the VDSO
is position independent we can remove the hint and let get_unmapped_area pick
an area. This will mean the vdso will be near other mmaps and will share
an SLB entry:
10000000-10001000 r-xp 00000000 08:06 5778459 /root/context_switch_64
10010000-10011000 r--p 00000000 08:06 5778459 /root/context_switch_64
10011000-10012000 rw-p 00001000 08:06 5778459 /root/context_switch_64
fffa92ae000-fffa92b0000 rw-p 00000000 00:00 0
fffa92b0000-fffa9453000 r-xp 00000000 08:06 4334051 /lib64/power6/libc-2.9.so
fffa9453000-fffa9462000 ---p 001a3000 08:06 4334051 /lib64/power6/libc-2.9.so
fffa9462000-fffa9466000 r--p 001a2000 08:06 4334051 /lib64/power6/libc-2.9.so
fffa9466000-fffa947c000 rw-p 001a6000 08:06 4334051 /lib64/power6/libc-2.9.so
fffa947c000-fffa9480000 rw-p 00000000 00:00 0
fffa9480000-fffa94a8000 r-xp 00000000 08:06 4333852 /lib64/ld-2.9.so
fffa94b3000-fffa94b4000 rw-p 00000000 00:00 0
fffa94b4000-fffa94b7000 r-xp 00000000 00:00 0 [vdso] <----- here I am
fffa94b7000-fffa94b8000 r--p 00027000 08:06 4333852 /lib64/ld-2.9.so
fffa94b8000-fffa94bb000 rw-p 00028000 08:06 4333852 /lib64/ld-2.9.so
fffa94bb000-fffa94bc000 rw-p 00000000 00:00 0
fffe4c10000-fffe4c25000 rw-p 00000000 00:00 0 [stack]
On a microbenchmark that bounces a token between two 64bit processes over pipes
and calls gettimeofday each iteration (to access the VDSO), our context switch
rate goes from 268k to 277k ctx switches/sec (tested on a 4GHz POWER6).
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The bitops.h functions that operate on a single bit in a bitfield are
implemented by operating on the corresponding word location. In all
cases the inner logic is valid if the mask being applied has more than
one bit set, so this patch exposes those inner operations. Indeed,
set_bits() was already available, but it duplicated code from
set_bit() (rather than making the latter a wrapper) - it was also
missing the PPC405_ERR77() workaround and the "volatile" address
qualifier present in other APIs. This corrects that, and exposes the
other multi-bit equivalents.
One advantage of these multi-bit forms is that they allow word-sized
variables to essentially be their own spinlocks, eg. very useful for
state machines where an atomic "flags" variable can obviate the need
for any additional locking.
Signed-off-by: Geoff Thorpe <geoff@geoffthorpe.net>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
On 32-bit systems with 64-bit PTEs, the PTEs have to be written in two
32-bit halves. On SMP we write the higher-order half and then the
lower-order half, with a write barrier between the two halves, but on
UP there was no particular ordering of the writes to the two halves.
This extends the ordering that we already do on SMP to the UP case as
well. The reason is that with the perf_counter subsystem potentially
accessing user memory at interrupt time to get stack traces, we have
to be careful not to create an incorrect but apparently valid PTE even
on UP.
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Eliminates this compiler warning:
arch/powerpc/kvm/../../../virt/kvm/kvm_main.c:1178: error: integer overflow in expression
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Avi Kivity <avi@redhat.com>
This adds two functions, phys_to_dma() and dma_to_phys() to x86, IA64
and powerpc. swiotlb uses them. phys_to_dma() converts a physical
address to a dma address. dma_to_phys() does the opposite.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Acked-by: Becky Bruce <beckyb@kernel.crashing.org>
dma_capable() eventually replaces is_buffer_dma_capable(), which tells
if a memory area is dma-capable or not. The problem of
is_buffer_dma_capable() is that it doesn't take a pointer to struct
device so it doesn't work for POWERPC.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Acked-by: Becky Bruce <beckyb@kernel.crashing.org>
mm: Pass virtual address to [__]p{te,ud,md}_free_tlb()
Upcoming paches to support the new 64-bit "BookE" powerpc architecture
will need to have the virtual address corresponding to PTE page when
freeing it, due to the way the HW table walker works.
Basically, the TLB can be loaded with "large" pages that cover the whole
virtual space (well, sort-of, half of it actually) represented by a PTE
page, and which contain an "indirect" bit indicating that this TLB entry
RPN points to an array of PTEs from which the TLB can then create direct
entries. Thus, in order to invalidate those when PTE pages are deleted,
we need the virtual address to pass to tlbilx or tlbivax instructions.
The old trick of sticking it somewhere in the PTE page struct page sucks
too much, the address is almost readily available in all call sites and
almost everybody implemets these as macros, so we may as well add the
argument everywhere. I added it to the pmd and pud variants for consistency.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: David Howells <dhowells@redhat.com> [MN10300 & FRV]
Acked-by: Nick Piggin <npiggin@suse.de>
Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com> [s390]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull the initial preempt_count value into a single
definition site.
Maintainers for: alpha, ia64 and m68k, please have a look,
your arch code is funny.
The header magic is a bit odd, but similar to the KERNEL_DS
one, CPP waits with expanding these macros until the
INIT_THREAD_INFO macro itself is expanded, which is in
arch/*/kernel/init_task.c where we've already included
sched.h so we're good.
Cc: tony.luck@intel.com
Cc: rth@twiddle.net
Cc: geert@linux-m68k.org
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Matt Mackall <mpm@selenic.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Current implementation of spin_event_timeout can be interrupted by an
IRQ or context switch after testing the condition, but before checking
the timeout. This can cause the loop to report a timeout when the
condition actually became true in the middle.
This patch adds one final check of the condition upon exit of the loop
if the last test of the condition was still false.
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
* 'perfcounters-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (47 commits)
perf report: Add --symbols parameter
perf report: Add --comms parameter
perf report: Add --dsos parameter
perf_counter tools: Adjust only prelinked symbol's addresses
perf_counter: Provide a way to enable counters on exec
perf_counter tools: Reduce perf stat measurement overhead/skew
perf stat: Use percentages for scaling output
perf_counter, x86: Update x86_pmu after WARN()
perf stat: Micro-optimize the code: memcpy is only required if no event is selected and !null_run
perf stat: Improve output
perf stat: Fix multi-run stats
perf stat: Add -n/--null option to run without counters
perf_counter tools: Remove dead code
perf_counter: Complete counter swap
perf report: Print sorted callchains per histogram entries
perf_counter tools: Prepare a small callchain framework
perf record: Fix unhandled io return value
perf_counter tools: Add alias for 'l1d' and 'l1i'
perf-report: Add bare minimum PERF_EVENT_READ parsing
perf-report: Add modes for inherited stats and no-samples
...
When using 64k page sizes, our PTE pages are split in two halves,
the second half containing the "extension" used to keep track of
individual 4k pages when not using HW 64k pages.
However, our page tables used for hugetlb have a slightly different
format and don't carry that "second half".
Our code that batched PTEs to be invalidated unconditionally reads
the "second half" (to put it into the batch), which means that when
called to invalidate hugetlb PTEs, it will access unrelated memory.
It breaks when CONFIG_DEBUG_PAGEALLOC is enabled.
This fixes it by only accessing the second half when the _PAGE_COMBO
bit is set in the first half, which indicates that we are dealing with
a "combo" page which represents 16x4k subpages. Anything else shouldn't
have this bit set and thus not require loading from the second half.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Several platforms use their own copy of what is essentially the same code,
using RTAS to synchronize the timebases when bringing up new CPUs. This
moves it all into a single common implementation and additionally
turns the spinlock into a raw spinlock since the former can rely on
the timebase not being frozen when spinlock debugging is enabled, and finally
masks interrupts while the timebase is disabled.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
RTAS currently uses a normal spinlock. However it can be called from
contexts where this is not necessarily a good idea. For example, it
can be called while syncing timebases, with the core timebase being
frozen. Unfortunately, that will deadlock in case of lock contention
when spinlock debugging is enabled as the spin lock debugging code
will try to use __delay() which ... relies on the timebase being
enabled.
Also RTAS can be used in some low level IRQ handling code path so it
may as well be a raw spinlock for -rt sake.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Based on initial work from: Dale Farnsworth <dale@farnsworth.org>
Add the low level irq tracing hooks for 32-bit powerpc needed
to enable full lockdep functionality.
The approach taken to deal with the code in entry_32.S is that
we don't trace all the transitions of MSR:EE when we just turn
it off to peek at TI_FLAGS without races. Only when we are
calling into C code or returning from exceptions with a state
that have changed from what lockdep thinks.
There's a little bugger though: If we take an exception that
keeps interrupts enabled (such as an alignment exception) while
interrupts are enabled, we will call trace_hardirqs_on() on the
way back spurriously. Not a big deal, but to get rid of it would
require remembering in pt_regs that the exception was one of the
type that kept interrupts enabled which we don't know at this
stage. (Well, we could test all cases for regs->trap but that
sucks too much).
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Tested-by: Kumar Gala <galak@kernel.crashing.org>
Those functions are way too big to be inline, besides, kmap_atomic()
wants to call debug_kmap_atomic() which isn't exported for modules
and causes module link failures.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Turning on SWIOTLB selects or enables PPC_NEED_DMA_SYNC_OPS, which means
we get the non empty versions of dma_sync_* in asm/dma-mapping.h
On my pseries machine the dma_ops have no such routines and we die with
a null pointer - this patch gets it booting, is there a more elegant way
to do it?
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Update the mmap control page with the needed information to
use the userspace RDPMC instruction for self monitoring.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
* 'perfcounters-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (49 commits)
perfcounter: Handle some IO return values
perf_counter: Push perf_sample_data through the swcounter code
perf_counter tools: Define and use our own u64, s64 etc. definitions
perf_counter: Close race in perf_lock_task_context()
perf_counter, x86: Improve interactions with fast-gup
perf_counter: Simplify and fix task migration counting
perf_counter tools: Add a data file header
perf_counter: Update userspace callchain sampling uses
perf_counter: Make callchain samples extensible
perf report: Filter to parent set by default
perf_counter tools: Handle lost events
perf_counter: Add event overlow handling
fs: Provide empty .set_page_dirty() aop for anon inodes
perf_counter: tools: Makefile tweaks for 64-bit powerpc
perf_counter: powerpc: Add processor back-end for MPC7450 family
perf_counter: powerpc: Make powerpc perf_counter code safe for 32-bit kernels
perf_counter: powerpc: Change how processor-specific back-ends get selected
perf_counter: powerpc: Use unsigned long for register and constraint values
perf_counter: powerpc: Enable use of software counters on 32-bit powerpc
perf_counter tools: Add and use isprint()
...
