summaryrefslogtreecommitdiffstats
path: root/arch/arm/nwfpe/softfloat.h
diff options
context:
space:
mode:
authorLennert Buytenhek <buytenh@wantstofly.org>2005-11-07 22:12:07 +0100
committerRussell King <rmk+kernel@arm.linux.org.uk>2005-11-07 22:12:07 +0100
commit06c03cac9487555478c7d80065ebf7818bf6fd06 (patch)
tree1a03ff9ae347f40c11d0e3eee790a1b88ff29cc1 /arch/arm/nwfpe/softfloat.h
parent[ARM SMP] Add Realview MPcore SMP support (diff)
downloadlinux-06c03cac9487555478c7d80065ebf7818bf6fd06.tar.xz
linux-06c03cac9487555478c7d80065ebf7818bf6fd06.zip
[ARM] 3117/1: nwfpe kernel memory info leak
Patch from Lennert Buytenhek The routine that nwfpe uses for converting floats/doubles to extended precision fails to zero two bytes of kernel stack. This is not immediately obvious, as the floatx80 structure has 16 bits of implicit padding (by design.) These two bytes are copied to userspace when an stfe is emulated, causing a possible info leak. Make the padding explicit and zero it out in the relevant places. Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Diffstat (limited to 'arch/arm/nwfpe/softfloat.h')
-rw-r--r--arch/arm/nwfpe/softfloat.h1
1 files changed, 1 insertions, 0 deletions
diff --git a/arch/arm/nwfpe/softfloat.h b/arch/arm/nwfpe/softfloat.h
index 14151700b6b2..1301d97e037f 100644
--- a/arch/arm/nwfpe/softfloat.h
+++ b/arch/arm/nwfpe/softfloat.h
@@ -55,6 +55,7 @@ typedef unsigned long int float32;
typedef unsigned long long float64;
typedef struct {
unsigned short high;
+ unsigned short __padding;
unsigned long long low;
} floatx80;