pstore-ram: Allow optional mapping with pgprot_noncached

On some ARMs the memory can be mapped pgprot_noncached() and still
be working for atomic operations. As pointed out by Colin Cross
<[email protected]>, in some cases you do want to use
pgprot_noncached() if the SoC supports it to see a debug printk
just before a write hanging the system.

On ARMs, the atomic operations on strongly ordered memory are
implementation defined. So let's provide an optional kernel parameter
for configuring pgprot_noncached(), and use pgprot_writecombine() by
default.

Cc: Arnd Bergmann <[email protected]>
Cc: Rob Herring <[email protected]>
Cc: Randy Dunlap <[email protected]>
Cc: Anton Vorontsov <[email protected]>
Cc: Colin Cross <[email protected]>
Cc: Olof Johansson <[email protected]>
Cc: Russell King <[email protected]>
Cc: [email protected]
Acked-by: Kees Cook <[email protected]>
Signed-off-by: Tony Lindgren <[email protected]>
Signed-off-by: Tony Luck <[email protected]>
This commit is contained in:
Tony Lindgren 2014-09-16 13:50:01 -07:00 • committed by Tony Luck
commit 027bc8b082
4 changed files with 47 additions and 14 deletions

View file

@ -14,11 +14,19 @@ survive after a restart.
1. Ramoops concepts
Ramoops uses a predefined memory area to store the dump. The start and size of
the memory area are set using two variables:
Ramoops uses a predefined memory area to store the dump. The start and size
and type of the memory area are set using three variables:
* "mem_address" for the start
* "mem_size" for the size. The memory size will be rounded down to a
power of two.
* "mem_type" to specifiy if the memory type (default is pgprot_writecombine).
Typically the default value of mem_type=0 should be used as that sets the pstore
mapping to pgprot_writecombine. Setting mem_type=1 attempts to use
pgprot_noncached, which only works on some platforms. This is because pstore
depends on atomic operations. At least on ARM, pgprot_noncached causes the
memory to be mapped strongly ordered, and atomic operations on strongly ordered
memory are implementation defined, and won't work on many ARMs such as omaps.
The memory area is divided into "record_size" chunks (also rounded down to
power of two) and each oops/panic writes a "record_size" chunk of
@ -55,6 +63,7 @@ Setting the ramoops parameters can be done in 2 different manners:
static struct ramoops_platform_data ramoops_data = {
.mem_size = <...>,
.mem_address = <...>,
.mem_type = <...>,
.record_size = <...>,
.dump_oops = <...>,
.ecc = <...>,