The tricky part here is that task_pt_regs() on m68k works *only* for
process inside do_signal().  However, we need something much simpler -
pt_regs of a process inside do_signal() may be at different offsets
from the stack bottom, depending on the way we'd entered the kernel,
but for a task inside sys_execve() it *is* at constant offset.
Moreover, for a kernel thread about to become a userland process the
same location is also fine - setting sp to that will leave the kernel
stack pointer at the very bottom of the kernel stack when we finally
switch to userland.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

In the old times, the whole idle task was considered
as an RCU quiescent state. But as RCU became more and
more successful overtime, some RCU read side critical
section have been added even in the code of some
architectures idle tasks, for tracing for example.

So nowadays, rcu_idle_enter() and rcu_idle_exit() must
be called by the architecture to tell RCU about the part
in the idle loop that doesn't make use of rcu read side
critical sections, typically the part that puts the CPU
in low power mode.

This is necessary for RCU to find the quiescent states in
idle in order to complete grace periods.

Add this missing pair of calls in the m68k's idle loop.
Reported-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: NFrederic Weisbecker <fweisbec@gmail.com>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Cc: m68k <linux-m68k@lists.linux-m68k.org>
Cc: <stable@vger.kernel.org> # 3.3+
Signed-off-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: NJosh Triplett <josh@joshtriplett.org>

Most architectures implement this in exactly the same way. Instead of
having each architecture duplicate this function, provide a single
implementation in the core and make it a weak symbol so that it can be
overridden on architectures where it is required.
Signed-off-by: NThierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

The semantic patch that makes this change is available
in scripts/coccinelle/api/ptr_ret.cocci.

More information about semantic patching is available at
http://coccinelle.lip6.fr/Signed-off-by: NThomas Meyer <thomas@m3y3r.de>
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

User space access must always go through uaccess accessors, since on
classic m68k user space and kernel space are completely separate.
Signed-off-by: NAndreas Schwab <schwab@linux-m68k.org>
Tested-by: NThorsten Glaser <tg@debian.org>
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Cc: stable@vger.kernel.org

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Basic set of definitions and support code required to turn on CONFIG_PCI
for the m68k architecture. Nothing specific to any PCI implementation in
any m68k class CPU hardware yet.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

The dma cache support functions do not currently support the direction flag
DMA_BIDIRECTIONAL. If a driver passes this direction to dma_map_single or
friends you will get console output like this:

  dma_sync_single_for_device: unsupported dir 0

For example when using the Intel e100 ethernet driver on a ColdFire platform
with PCI bus. You will get a stream of these messages coming out.

Modify the dma cache support code adding support for DMA_BIDIRECTIONAL. It is
actioned by doing a cache push operation.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>

On all ColdFire platforms (whether MMU enabled or not) we want to use the
simple page based dma_alloc_coherent. We don't want the virtual mapping
version that is used on classic m68k setups. So modify the conditionals
to use the existing simpler dma_alloc_coherent on all ColdFire and non-MMU
builds.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>

Some of the entry.S code is common to both MMU and non-MMU builds.
So merge the entry_no.S and entry_mm.S files back into a single file.
With a little code movement we only need a single #ifdef.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

There is a few places that the m68k entry code uses the bsrl instruction
to call other functions. That instruction is only supported on 68020 and
higher CPU types. If we use jbsr instead the code will be clean for all
68k and ColdFire CPU types.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

The ret_from_excption code is referenced by its function name, or by a label
set at the start of its code. The non-MMU code can share some of this code
if we make direct calls to ret_from_exception instead of the associated label.
The effected function paths are: buserr, trap and ret_from_fork. So change
these to branch directly to ret_from_exception.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

BSS_SECTION() provides the __bss_{start,stop} symbols, so there's no need
to wrap our own _[se]bss around it.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Greg Ungerer<gerg@uclinux.org>

The assembler entry code calls directly to the syscall_trace_enter() and
syscall_trace_leave() functions. But currently they are conditionaly
compiled out for the non-MMU classic m68k CPU types (so 68328 for example),
resulting in a link error:

      LD      vmlinux
    arch/m68k/platform/68328/built-in.o: In function `do_trace':
    (.text+0x1c): undefined reference to `syscall_trace_enter'
    arch/m68k/platform/68328/built-in.o: In function `do_trace':
    (.text+0x4c): undefined reference to `syscall_trace_leave'

Change the conditional check that includes these functions to be true for
the !defined(CONFIG_MMU) case as well.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

When building for non-MMU based classic 68k CPU types (like the 68328 for
example) you get a compilation error:

  CC      arch/m68k/kernel/time.o
arch/m68k/kernel/time.c:91:5: error: redefinition of ‘arch_gettimeoffset’
include/linux/time.h:145:19: note: previous definition of ‘arch_gettimeoffset’ was here

The arch_gettimeoffset() code is included when building for these CPU types,
but it shouldn't be. Those machine types do not have
CONFIG_ARCH_USES_GETTIMEOFFSET set.

The fix is simply to conditionally include the arch_gettimeoffset() code on
that same config setting that specifies its use or not.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

  - Remove casts and unneeded address-of ('&') operators,
  - Use %p to format pointers, %lx to format unsigned longs.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Acked-by: NGreg Ungerer <gerg@uclinux.org>

Does block_sigmask() + tracehook_signal_handler();  called when
sigframe has been successfully built.  All architectures converted
to it; block_sigmask() itself is gone now (merged into this one).

I'm still not too happy with the signature, but that's a separate
story (IMO we need a structure that would contain signal number +
siginfo + k_sigaction, so that get_signal_to_deliver() would fill one,
signal_delivered(), handle_signal() and probably setup...frame() -
take one).
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Only 3 out of 63 do not.  Renamed the current variant to __set_current_blocked(),
added set_current_blocked() that will exclude unblockable signals, switched
open-coded instances to it.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

replace boilerplate "should we use ->saved_sigmask or ->blocked?"
with calls of obvious inlined helper...
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

first fruits of ..._restore_sigmask() helpers: now we can take
boilerplate "signal didn't have a handler, clear RESTORE_SIGMASK
and restore the blocked mask from ->saved_mask" into a common
helper.  Open-coded instances switched...
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

TIF_NOTIFY_RESUME added (as bit 5).  That way nommu glue needs no changes at
all; mmu one needs just to replace jmi do_signal_return to jne do_signal_return
There we have flags shifted up, until bit 6 (SIGPENDING) is in MSBit; instead
of checking that MSBit is set (jmi) we check that MSBit or something below it
is set (jne); bits 0..4 are never set, so that's precisely "bit 6 or bit 5 is
set".

Usual handling of NOTIFY_RESUME/SIGPENDING is done in do_notify_resume(); glue
calls it instead of do_signal().
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

As described in e6fa16ab ("signal: sigprocmask() should do
retarget_shared_pending()") the modification of current->blocked is
incorrect as we need to check whether the signal we're about to block is
pending in the shared queue.

Also, use the new helper function introduced in commit 5e6292c0
("signal: add block_sigmask() for adding sigmask to current->blocked")
which centralises the code for updating current->blocked after
successfully delivering a signal and reduces the amount of duplicate code
across architectures.  In the past some architectures got this code wrong,
so using this helper function should stop that from happening again.
Acked-by: NOleg Nesterov <oleg@redhat.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: NGreg Ungerer <gerg@uclinux.org>
Signed-off-by: NMatt Fleming <matt.fleming@intel.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

guts of saved_sigmask-based sigsuspend/rt_sigsuspend.  Takes
kernel sigset_t *.

Open-coded instances replaced with calling it.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

The majority of the m68k architecture dma code is the same, so merge the
current separated files dma_no.c and dma_mm.c back into a single dma.c

The main alloc and free routines are a little different, so we keep a
single #ifdef based on CONFIG_MMU for them. All the other support functions
are now identical.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

The MMU (signal_mm.c) and non-MMU (signal_no.c) versions of the m68k
architecture signal handling code are very similar. Most of their code is
the same.

Merge the two back into a single signal.c, and move some of the code around
inside the file to minimize the number of #ifdefs required. Specificially
we can group out the CONFIG_FPU and the CONFIG_MMU code. We end up needing
a few other "#ifdef CONFIG_MMU" as well, but not too many.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Same code. Use the generic version.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Link: http://lkml.kernel.org/r/20120503085034.816195285@linutronix.de

Disintegrate asm/system.h for M68K.
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Acked-by: NGreg Ungerer <gerg@uclinux.org>
cc: linux-m68k@lists.linux-m68k.org

The MMU and non-MMU varients of the m68k arch process.c code are pretty
much the same. Only a few minor details differ between the two. The
majority of the difference is to deal with having or wanting hardware FPU
support. So merge them back into a single process.c file.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Most of the code in the non-mmu ptrace_no.c file is the same as the mmu
version ptrace_mm.c. So merge them back into a single file.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

The set_rtc_mmss() function is defined "static inline" but is never used
in this file. Remove it.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

There is only trivial differences between the mmu time_mm.c and non-mmu
time_no.c files. Merge them back into a single time.c.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>

The CONFIG_GENERIC_CMOS_UPDATE switch is always enabled for the non-MMU
m68k case. But the underlying code to support it, update_persistent_clock(),
doesn't end up doing anything on the currently supported non-MMU platforms.
No platforms supply the necessary function support for writing back the RTC.

So lets remove this option and support code. This also brings m68knommu
in line with the m68k, which doesn't enabled this switch either.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>

With a few small changes we can make the m68knommu timer init code the
same as the m68k code. By using the mach_sched_init function pointer
and reworking the current timer initializers to keep track of the common
m68k timer_interrupt() handler we end up with almost identical code for
m68knommu.

This will allow us to more easily merge the mmu and non-mmu m68k time.c
in future patches.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>

The read_persistent_clock() code is different on m68knommu, for really no
reason. With a few changes to support function names and some code
re-organization the code can be made the same.

This will make it easier to merge the arch/m68k/kernel/time.c for m68k and
m68knommu in a future patch.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>

There is a lot of years of collected cruft in the m68knommu linker script.
Clean it all up and use the well defined linker script support macros.

Support is maintained for building both ROM/FLASH based and RAM based setups.
No major changes to section layouts, though the rodata section is now lumped
in with the read/write data section.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>

Coccinelle based conversion.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/n/tip-24swm5zut3h9c4a6s46x8rws@git.kernel.orgSigned-off-by: NIngo Molnar <mingo@elte.hu>

Passing the address of a variable as an operand to an asm statement
doesn't mark the value of this variable as used, so gcc may optimize its
initialisation away.  Fix this by using the "m" constraint instead.
Signed-off-by: NAndreas Schwab <schwab@linux-m68k.org>
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Cc: stable@vger.kernel.org