[PATCH] powerpc/85xx: Fix SMP when "cpu-release-addr" is in lowmem

Peter Tyser ptyser at xes-inc.com
Sat Dec 19 09:50:37 EST 2009

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

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]
  Modules linked in:
  <... eventual kernel panic>

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

Signed-off-by: Peter Tyser <ptyser at xes-inc.com>
Signed-off-by: Nate Case <ncase at xes-inc.com>
Reported-by: Dipen Dudhat <B09055 at freescale.com>
Tested-by: Dipen Dudhat <B09055 at freescale.com>
 arch/powerpc/platforms/85xx/smp.c |   21 +++++++++++++++++++--
 1 files changed, 19 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/platforms/85xx/smp.c b/arch/powerpc/platforms/85xx/smp.c
index 04160a4..a15f582 100644
--- a/arch/powerpc/platforms/85xx/smp.c
+++ b/arch/powerpc/platforms/85xx/smp.c
@@ -46,6 +46,7 @@ smp_85xx_kick_cpu(int nr)
 	__iomem u32 *bptr_vaddr;
 	struct device_node *np;
 	int n = 0;
+	int ioremappable;
 	WARN_ON (nr < 0 || nr >= NR_CPUS);
@@ -59,21 +60,37 @@ smp_85xx_kick_cpu(int nr)
+	/*
+	 * A secondary core could be in a spinloop in the bootpage
+	 * (0xfffff000), somewhere in highmem, or somewhere in lowmem.
+	 * The bootpage and highmem can be accessed via ioremap(), but
+	 * we need to directly access the spinloop if its in lowmem.
+	 */
+	ioremappable = *cpu_rel_addr > virt_to_phys(high_memory);
 	/* Map the spin table */
-	bptr_vaddr = ioremap(*cpu_rel_addr, SIZE_BOOT_ENTRY);
+	if (ioremappable)
+		bptr_vaddr = ioremap(*cpu_rel_addr, SIZE_BOOT_ENTRY);
+	else
+		bptr_vaddr = phys_to_virt(*cpu_rel_addr);
 	out_be32(bptr_vaddr + BOOT_ENTRY_PIR, nr);
 	out_be32(bptr_vaddr + BOOT_ENTRY_ADDR_LOWER, __pa(__early_start));
+	if (!ioremappable)
+		flush_dcache_range((ulong)bptr_vaddr,
+				(ulong)(bptr_vaddr + SIZE_BOOT_ENTRY));
 	/* Wait a bit for the CPU to ack. */
 	while ((__secondary_hold_acknowledge != nr) && (++n < 1000))
-	iounmap(bptr_vaddr);
+	if (ioremappable)
+		iounmap(bptr_vaddr);
 	pr_debug("waited %d msecs for CPU #%d.\n", n, nr);

More information about the Linuxppc-dev mailing list