> > I originally thought that patch to kernel is going to be big and
> > heavy. If it is not (2 lines seem reasonable to me), could you please
> > post proposed kernel modification?
[instructions skipped]
> Beside this all to do is to make sure all ports define FPUCW_DEFAULT
> and also use this value in the code (I don't know whether such a macro
> already exists; if not it's high time to add one). That's all.
I tried to find similar macro, and I failed. I tried to find what
value FPU is initialized to at i386, and I failed, too. Ok, at least
I've successfully applied your "patch" and turned it into real patch
so it is easier to handle. I made that change to kernel I'm running
at, so it is not _too_ toxic.
Someone with better knowledge of architectures and FPU initialization
should probably take it, change 0 into real value, and put it into
real kernel. (Dave?)
Pavel
--- linux/include/linux/elf.h.ulrich Thu May 28 22:10:34 1998
+++ linux/include/linux/elf.h Thu May 28 22:11:33 1998
@@ -130,5 +130,7 @@
#define AT_PLATFORM 15 /* string identifying cpu for optimizations */
#define AT_HWCAP 16 /* arch dependent hints at cpu capabilities */
+#define AT_LINUX_FPUCW 17 /* initial settings of FPU */
+
typedef struct dynamic{
Elf32_Sword d_tag;
union{
--- linux/fs/binfmt_elf.c.ulrich Thu May 28 22:11:56 1998
+++ linux/fs/binfmt_elf.c Thu May 28 22:24:56 1998
@@ -156,14 +156,14 @@
__put_user ((id), sp+(nr*2)); \
__put_user ((val), sp+(nr*2+1)); \
- sp -= 2;
+ sp -= 6;
NEW_AUX_ENT(0, AT_NULL, 0);
+ NEW_AUX_ENT(1, AT_HWCAP, hwcap);
+ NEW_AUX_ENT(2, AT_LINUX_FPUCW, 0 /* FPUCW_DEFAULT */ );
if (k_platform) {
sp -= 2;
NEW_AUX_ENT(0, AT_PLATFORM, (elf_addr_t)(unsigned long) u_platform);
}
- sp -= 2;
- NEW_AUX_ENT(0, AT_HWCAP, hwcap);
if (exec) {
sp -= 11*2;
-- I'm really pavel@atrey.karlin.mff.cuni.cz. Pavel Look at http://atrey.karlin.mff.cuni.cz/~pavel/ ;-).- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu