Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

It's too valuable for debugging to be disabled.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Also fix a few printf-style formats, to get rid of the following compiler
warnings when DEBUG is enabled:

arch/m68k/kernel/traps.c: In function ‘access_error060’:
arch/m68k/kernel/traps.c:166: warning: format ‘%d’ expects type ‘int’, but argument 2 has type ‘long unsigned int’
arch/m68k/kernel/traps.c: In function ‘bus_error030’:
arch/m68k/kernel/traps.c:568: warning: format ‘%#lx’ expects type ‘long unsigned int’, but argument 2 has type ‘void *’
arch/m68k/kernel/traps.c:682: warning: format ‘%#lx’ expects type ‘long unsigned int’, but argument 2 has type ‘void *’
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

When DEBUG is enabled, do_page_fault() may dereference a NULL pointer,
causing recursive bus errors.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Cfr. commit 7934779a ("Driver-Core: disable
/sbin/hotplug by default").
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Rename RTC_PORT() to ATARI_RTC_PORT(), as the rtc-cmos RTC driver uses the
presence of this macro to enable support for the second NVRAM bank, which
Atari doesn't have ("Unable to handle kernel access at virtual address
00ff8965").
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Since commit d6713b40 ("m68k: early
parameter support"), the user can specify multiple debug consoles using the
"debug=" kernel command line parameter.
However, as there's only a single struct console object, which is reused,
it would actually register the same console object multiple times, causing
the following warning:

WARNING: CPU: 0 PID: 0 at kernel/printk/printk.c:2233 register_console+0x36/
console 'debug0' already registered

Make sure to register the console object only once, to avoid the warning.

Note that still only one console (the one corresponding to the last
"debug=" parameter) will be active at the same time, as the .write() method
of the already registered console object is overwritten by a subsequent
"debug=" parameter.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

This reduces the kernel image size by ca. 160 KiB.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Add optional support to export the bootinfo used to boot the kernel in a
"bootinfo" file in procfs.  This is useful with kexec.

This is based on the similar feature for ATAGS on ARM.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Use the beam position registers, which provide at least 17 bits of data
changing at 1.79 MHz.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

On m68k, get_cycles() (the default implementation for random_get_entropy())
always returns zero, providing no entropy for the random driver.

Add a hook where platforms can provide their own implementation, and wire
it up in the infrastructure provided by commit
61875f30 ("random: allow architectures to
optionally define random_get_entropy()").
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

nf_init() uses virt_to_phys(), which depends on m68k_memoffset being set and
module_fixup() having been called, but this is only done in paging_init().
Hence call paging_init() before nf_init().

This went unnoticed, as virt_to_phys() is a no-op on Atari, unless you start
fiddling with the memory blocks in the bootinfo manually.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Move generic definitions used by bootstraps to uapi/asm/bootinfo.h:
  - Machine types,
  - CPU, FPU, and MMU types,
  - struct mem_info.

Keep a copy of struct mem_info for in-kernel use, and rename it to struct
m68k_mem_info, as the exported one will be modified later.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Export the bootinfo definitions that are used by bootstrap loaders, and
split them up in generic and platform-specific parts.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

struct mac_booter_data is no longer part of the bootinfo API, hence move it
from <asm/bootinfo.h> to <asm/macintosh.h>, dropping all unused fields in
the process.

Also remove the no longer used mac_booter_data pointer from head.S.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Drop remainings and API for backwards compatibility with bootinfo interface
version 1.0.  This was used when booting a 2.1.x or newer kernel on Amiga,
Atari, or Mac using a bootstrap for kernel 2.0.x.

Everybody upgraded his bootstrap a long time ago, so this can go.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Since the introduction of init sections (which are located after BSS), the
bootinfo is no longer located right after the BSS, but after all kernel
sections.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Fix member definitions for non-native userspace handling:
  - All multi-byte values are big-endian, hence use __be*,
  - All pointers are 32-bit pointers under AmigaOS, but unused (except for
    cd_BoardAddr) under Linux, hence use __be32.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

ZTWO_VADDR() converts from physical to virtual I/O addresses, so it should
return "void __iomem *" instead of "unsigned long".

This allows to drop several casts, but requires adding a few casts to
accomodate legacy driver frameworks that store "unsigned long" I/O
addresses.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Currently the array of Zorro devices is allocated statically, wasting
up to 4.5 KiB when running an Amiga or multi-platform kernel on a machine
with no or a handful of Zorro expansion cards. Convert it to conditional
dynamic memory allocation to fix this.

amiga_parse_bootinfo() still needs to store some information about the
detected Zorro devices, at a time even the bootmem allocator is not yet
available.  This is now handled using a much smaller array (typically less
than 0.5 KiB), which is __initdata and thus freed later.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

This reduces the kernel image size by ca. 100 KiB, while still allowing
NFS root.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

This is no longer done from iop_init()
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

As Sun 3 kernels cannot be multi-platform due to the different Sun 3 MMU
type, it made sense to statically allocate the table to track IOMMU use.

However, Sun 3x kernels can be multi-platform. Furthermore, Sun 3x uses
a larger table than Sun 3 (8192 bytes instead of 512 bytes).

Hence switch to dynamic allocation of this table using the bootmem
allocator to avoid wasting 8192 bytes when not running on a Sun 3x.
As this allocator returns zeroed memory, there's no need to explicitly
initialize the table to zeroes.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

Some functions that are only called (indirectly) from setup_arch() lack
__init annotations.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>

There are two code paths how page with pmd page table can be freed:
pmd_free() and pmd_free_tlb().

I've missed the second one and didn't add page table destructor call
there.  It leads to leak of page->ptl for pmd page tables, if
dynamically allocated page->ptl is in use.

The patch adds the missed destructor and modifies documentation
accordingly.
Signed-off-by: NKirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reported-by: NAndrey Vagin <avagin@openvz.org>
Tested-by: NAndrey Vagin <avagin@openvz.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The VSX MSR bit in the user context indicates if the context contains VSX
state.  Currently we set this when the process has touched VSX at any stage.

Unfortunately, if the user has not provided enough space to save the VSX state,
we can't save it but we currently still set the MSR VSX bit.

This patch changes this to clear the MSR VSX bit when the user doesn't provide
enough space.  This indicates that there is no valid VSX state in the user
context.

This is needed to support get/set/make/swapcontext for applications that use
VSX but only provide a small context.  For example, getcontext in glibc
provides a smaller context since the VSX registers don't need to be saved over
the glibc function call.  But since the program calling getcontext may have
used VSX, the kernel currently says the VSX state is valid when it's not.  If
the returned context is then used in setcontext (ie. a small context without
VSX but with MSR VSX set), the kernel will refuse the context.  This situation
has been reported by the glibc community.

Based on patch from Carlos O'Donell.
Tested-by: NHaren Myneni <haren@linux.vnet.ibm.com>
Signed-off-by: NMichael Neuling <mikey@neuling.org>
Cc: stable@vger.kernel.org
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

In commit a489043f "Implement arch_get_random_long() based on H_RANDOM" I
broke the SMP=n build. We were getting plpar_wrappers.h via spinlock.h
which breaks when SMP=n.
Signed-off-by: NMichael Ellerman <mpe@ellerman.id.au>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>