[RFC PATCH] x86: let 'reservetop' functioning right

From: Liang Li
Date: Tue Apr 06 2010 - 11:57:24 EST

When specify 'reservetop=0xbadc0de' kernel parameter, the kernel will
stop booting due to a early_ioremap bug that relate to commit 8827247ff.

The root cause of boot failure problem is the value of 'slot_virt[i]'
was initialized in setup_arch->early_ioremap_init. But later in
setup_arch, the function 'parse_early_param' will modify 'FIXADDR_TOP'
when 'reservetop=0xbadc0de' being specified.

The simplest fix might be use __fix_to_virt(idx0) to get updated value
of 'FIXADDR_TOP' in '__early_ioremap' instead of reference old value
from slot_virt[slot] directly.

Signed-off-by: Liang Li <liang.li@xxxxxxxxxxxxx>

I am not sure if kernel command line option 'reservetop' will block boot
is the normal situation or not. Could someone tell me if we should fix
it or just leave it as is or someone is doing something to replace
'reservetop' with some other stuff?

arch/x86/mm/ioremap.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c
index 5eb1ba7..ea82ef0 100644
--- a/arch/x86/mm/ioremap.c
+++ b/arch/x86/mm/ioremap.c
@@ -537,9 +537,9 @@ __early_ioremap(resource_size_t phys_addr, unsigned long size, pgprot_t prot)
if (early_ioremap_debug)
- printk(KERN_CONT "%08lx + %08lx\n", offset, slot_virt[slot]);
+ printk(KERN_CONT "%08lx + %08lx\n", offset, __fix_to_virt(idx0));

- prev_map[slot] = (void __iomem *)(offset + slot_virt[slot]);
+ prev_map[slot] = (void __iomem *)(offset + __fix_to_virt(idx0));
return prev_map[slot];


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/