[PATCH] SH: catch negative denormal_subf1() retval in denormal_add()

[PATCH] SH: catch negative denormal_subf1() retval in denormal_add()

[Roel Kluin]

'ix' is unsigned but denormal_subf1() may return a negative int.

Signed-off-by: Roel Kluin <12o3l@tiscali.nl>
---
diff --git a/arch/sh/kernel/cpu/sh2a/fpu.c b/arch/sh/kernel/cpu/sh2a/fpu.c
index 5627c0b..6df2fb9 100644
--- a/arch/sh/kernel/cpu/sh2a/fpu.c
+++ b/arch/sh/kernel/cpu/sh2a/fpu.c
@@ -300,7 +300,7 @@ static int denormal_addf(int hx, int hy)
 		iy = hy & 0x7fffffff;
 		if (iy < 0x00800000) {
 			ix = denormal_subf1(ix, iy);
-			if (ix < 0) {
+			if ((int) ix < 0) {
 				ix = -ix;
 				sign ^= 0x80000000;
 			}
@@ -385,7 +385,7 @@ static long long denormal_addd(long long hx, long long hy)
 		iy = hy & 0x7fffffffffffffffLL;
 		if (iy < 0x0010000000000000LL) {
 			ix = denormal_subd1(ix, iy);
-			if (ix < 0) {
+			if ((int) ix < 0) {
 				ix = -ix;
 				sign ^= 0x8000000000000000LL;
 			}

Re: [PATCH] SH: catch negative denormal_subf1() retval in denormal_add()

[Paul Mundt]

On Thu, Apr 17, 2008 at 06:01:22AM +0200, Roel Kluin wrote:
> 'ix' is unsigned but denormal_subf1() may return a negative int.
> 
> Signed-off-by: Roel Kluin <12o3l@tiscali.nl>

I'm at a loss as to why you couldn't be bothered CC'ing the appropriate
list or anyone else on this, when you seem to have worked it out for
other patches.

Anyways, applied.