powerpc/85xx: Fix SMP when "cpu-release-addr" is in lowmem
authorPeter Tyser <ptyser@xes-inc.com>
Fri, 18 Dec 2009 12:50:37 +0000 (12:50 +0000)
committerBenjamin Herrenschmidt <benh@kernel.crashing.org>
Wed, 3 Feb 2010 06:39:49 +0000 (17:39 +1100)
commit7b62922a071aea362e879252d7482e448bd63d9c
tree5193d0011deef2ca331381f84afa2f41c4755da0
parent5be3492f972b73051ead7ecbac6fb9efd1e8e0ec
powerpc/85xx: Fix SMP when "cpu-release-addr" is in lowmem

Recent U-Boot commit 5ccd29c3679b3669b0bde5c501c1aa0f325a7acb caused
the "cpu-release-addr" device tree property to contain the physical RAM
location that secondary cores were spinning at.  Previously, the
"cpu-release-addr" property contained a value referencing the boot page
translation address range of 0xfffffxxx, which then indirectly accessed
RAM.

The "cpu-release-addr" is currently ioremapped and the secondary cores
kicked.  However, due to the recent change in "cpu-release-addr", it
sometimes points to a memory location in low memory that cannot be
ioremapped.  For example on a P2020-based board with 512MB of RAM the
following error occurs on bootup:

  <...>
  mpic: requesting IPIs ...
  __ioremap(): phys addr 0x1ffff000 is RAM lr c05df9a0
  Unable to handle kernel paging request for data at address 0x00000014
  Faulting instruction address: 0xc05df9b0
  Oops: Kernel access of bad area, sig: 11 [#1]
  SMP NR_CPUS=2 P2020 RDB
  Modules linked in:
  <... eventual kernel panic>

Adding logic to conditionally ioremap or access memory directly resolves
the issue.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Signed-off-by: Nate Case <ncase@xes-inc.com>
Reported-by: Dipen Dudhat <B09055@freescale.com>
Tested-by: Dipen Dudhat <B09055@freescale.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
arch/powerpc/platforms/85xx/smp.c