Without the update, we'll only see the new dmesg buffer after the
reboot, but previously we could see it right away. Making an oops
visible in pstore filesystem before reboot is a somewhat dubious
feature, but removing it wasn't an intentional change, so let's
restore it.

For this we have to make persistent_ram_save_old() safe for calling
multiple times, and also extern it.
Signed-off-by: NAnton Vorontsov <anton.vorontsov@linaro.org>
Acked-by: NKees Cook <keescook@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This is a first step for adding ECC support for pstore RAM backend: we
will use the persistent_ram routines, kindly provided by Google.

Basically, persistent_ram is a set of helper routines to deal with the
[optionally] ECC-protected persistent ram regions.

A bit of Makefile, Kconfig and header files adjustments were needed
because of the move.
Signed-off-by: NAnton Vorontsov <anton.vorontsov@linaro.org>
Acked-by: NKees Cook <keescook@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This is a transition patch to keep things bisectable, just moves
some routines under '#ifndef MODULE'. The code inside the #ifndef
will go away soon, but so far we must support pstore and ram_console.

So, we are about to use persistent_ram with pstore, with the ability
to compile persistent_ram routines as modules. Some parts of
persistent_ram uses memblock_reserve() routine, which is should be
only used built-in code, and thus it is not exported.

These persistent_ram bits are only used by Android's ram_console,
which is always built-in.

Without this patch, we won't able to compile persistent_ram as a
module:

ERROR: "memblock_reserve" [fs/pstore/ramoops.ko] undefined!
make[1]: *** [__modpost] Error 1
make: *** [modules] Error 2

(As alternative, we could export memblock_reserve, but the thing
is: we won't need it later.)
Signed-off-by: NAnton Vorontsov <anton.vorontsov@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

A corresponding function to persistent_ram_new().
Signed-off-by: NAnton Vorontsov <anton.vorontsov@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This includes devices' memory (e.g. framebuffers or memory mapped
EEPROMs on a local bus), as well as the normal RAM that we don't use
for the main memory.

For the normal (but unused) ram we could use kmaps, but this assumes
highmem support, so we don't bother and just use the memory via
ioremap.

As a side effect, the following hack is possible: when used together
with pstore_ram (new ramoops) module, we can limit the normal RAM region
with mem= and then point ramoops to use the rest of the memory, e.g.

	mem=128M ramoops.mem_address=0x8000000

Sure, we could just reserve the region with memblock_reserve() early in
the arch/ code, and then register a pstore_ram platform device pointing
to the reserved region. It's still a viable option if platform wants
to do so.

Also, we might want to use IO accessors in case of a real device,
but for now we don't bother (the old ramoops wasn't using it either, so
at least we don't make things worse).
Signed-off-by: NAnton Vorontsov <anton.vorontsov@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Factor out vmap logic out of persistent_ram_buffer_map(), this will
make the code a bit more understandable when we'll add support for
non-bootmem memory.
Signed-off-by: NAnton Vorontsov <anton.vorontsov@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The routine just creates a persistent ram zone at a specified address.

For persistent_ram_init_ringbuffer() we'd need to add a
'struct persistent_ram' to the global list, and associate it with a
device. We don't need all this complexity in pstore_ram, so we introduce
the simple function.
Signed-off-by: NAnton Vorontsov <anton.vorontsov@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Factor post init logic out of __persistent_ram_init(), we'll need
it for the new persistent_ram_new() routine.
Signed-off-by: NAnton Vorontsov <anton.vorontsov@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This is a longstanding bug, almost unnoticeable when calling
persistent_ram_write() for small buffers.

But when called for large data buffers, the write routine behaves
incorrectly, as the size may never update: instead of clamping
the size to the maximum buffer size, buffer_size_add_clamp() returns
an error (which is never checked by the write routine, btw).

To fix this, we now use buffer_size_add() that actually clamps the
size to the max value.

Also remove buffer_size_add_clamp(), it is no longer needed.
Signed-off-by: NAnton Vorontsov <anton.vorontsov@linaro.org>
Acked-by: NColin Cross <ccross@android.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The 'node' struct member is unused, so remove it.
Signed-off-by: NAnton Vorontsov <anton.vorontsov@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

If, in __persistent_ram_init(), the call to
persistent_ram_buffer_init() fails or the call to
persistent_ram_init_ecc() fails then we fail to free the memory we
allocated to 'prz' with kzalloc() - thus leaking it.

To prevent the leaks I consolidated all error exits from the function
at a 'err:' label at the end and made all error cases jump to that
label where we can then make sure we always free 'prz'. This is safe
since all the situations where the code bails out happen before 'prz'
has been stored anywhere and although we'll do a redundant kfree(NULL)
call in the case of kzalloc() itself failing that's OK since kfree()
deals gracefully with NULL pointers and I felt it was more important
to keep all error exits at a single location than to avoid that one
harmless/redundant kfree() on a error path.
Signed-off-by: NJesper Juhl <jj@chaosbits.net>
Acked-by: NColin Cross <ccross@android.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix printk format warnings in android/persistent_ram.c:

drivers/staging/android/persistent_ram.c:426:4: warning: format '%ld' expects type 'long int', but argument 2 has type 'size_t'
drivers/staging/android/persistent_ram.c:426:4: warning: format '%ld' expects type 'long int', but argument 3 has type 'size_t'
drivers/staging/android/persistent_ram.c:430:4: warning: format '%ld' expects type 'long int', but argument 2 has type 'size_t'
drivers/staging/android/persistent_ram.c:430:4: warning: format '%ld' expects type 'long int', but argument 3 has type 'size_t'
Signed-off-by: NRandy Dunlap <rdunlap@xenotime.net>
Cc: Brian Swetland <swetland@google.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Add the notrace attribute to persistent_ram_write and the
non-ecc functions that it calls to allow persistent_ram to
be used for ftracing (only when ecc is disabled).

CC: Greg KH <gregkh@linuxfoundation.org>
CC: Android Kernel Team <kernel-team@android.com>
Signed-off-by: NColin Cross <ccross@android.com>
Signed-off-by: NJohn Stultz <john.stultz@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Allow persistent_ram_write to be called on multiple cpus at
the same time, as long as ecc is not in use.  Uses atomics
for the buffer->start and buffer->size counters.

[jstultz: Fix up some pr_info casting issues on 64bit]
CC: Greg KH <gregkh@linuxfoundation.org>
CC: Android Kernel Team <kernel-team@android.com>
Signed-off-by: NColin Cross <ccross@android.com>
Signed-off-by: NJohn Stultz <john.stultz@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Replace the ioremapped memory passed in from the drivers with
a memblock_reserve and vmap.  Adds a new function,
persistent_ram_early_init, designed to be called from the machine
init_early callback, that calls memblock_remove and saves the
provided persistent ram area layout.

Drivers only pass in their struct device * and ecc settings.
Locating and mapping the memory is now handled entirely within
persistent_ram.

Also, convert ram_console to the new persistent_ram_init
parameters that only take a struct device * and ecc settings.

[jstultz: Fix pr_info casting issues on 64bit, folded two
patches as the build breaks if they are apart. Also replaced
phys_to_page() w/ pfn_to_page(addr>>PAGE_SHIFT), as phys_to_page
is only on a few arches.]
CC: Greg KH <gregkh@linuxfoundation.org>
CC: Android Kernel Team <kernel-team@android.com>
Signed-off-by: NColin Cross <ccross@android.com>
Signed-off-by: NJohn Stultz <john.stultz@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove CONFIG_ANDROID_RAM_CONSOLE_ERROR_CORRECTION and related
 #ifdefs.  Also allow persistent ram zones without ecc enabled.
For some use cases, like the data portion of the upcoming
persistent_vars patches, or a persistent ftrace ringbuffer,
ecc on every update is too expensive.

CC: Greg KH <gregkh@linuxfoundation.org>
CC: Android Kernel Team <kernel-team@android.com>
Signed-off-by: NColin Cross <ccross@android.com>
Signed-off-by: NJohn Stultz <john.stultz@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Split ram_console into two halves.

persistent_ram is a set of apis that handle a block of memory
that does not get erased across a reboot.  It provides functions
to fill it as a single buffer or a ring buffer, and to extract
the old data after a reboot.  It handles ecc on the data to
correct bit errors introduced during reboot.

ram_console is now a small wrapper around persistent_ram that
feeds console data into the ringbuffer, and exports the old
data to /proc/last_kmsg after a reboot.

[jstultz: Moved persistent_ram.h to staging dir]
CC: Greg KH <gregkh@linuxfoundation.org>
CC: Android Kernel Team <kernel-team@android.com>
Signed-off-by: NColin Cross <ccross@android.com>
Signed-off-by: NJohn Stultz <john.stultz@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>