aboutsummaryrefslogtreecommitdiff
path: root/arch/i386/lib/mmx.c
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@ppc970.osdl.org>2005-04-16 15:20:36 -0700
committerLinus Torvalds <torvalds@ppc970.osdl.org>2005-04-16 15:20:36 -0700
commit1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 (patch)
tree0bba044c4ce775e45a88a51686b5d9f90697ea9d /arch/i386/lib/mmx.c
downloadlinux-stericsson-2.6.12-rc2.tar.gz
Linux-2.6.12-rc2v2.6.12-rc2
Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
Diffstat (limited to 'arch/i386/lib/mmx.c')
-rw-r--r--arch/i386/lib/mmx.c399
1 files changed, 399 insertions, 0 deletions
diff --git a/arch/i386/lib/mmx.c b/arch/i386/lib/mmx.c
new file mode 100644
index 000000000000..01f8b1a2cc84
--- /dev/null
+++ b/arch/i386/lib/mmx.c
@@ -0,0 +1,399 @@
+#include <linux/config.h>
+#include <linux/types.h>
+#include <linux/string.h>
+#include <linux/sched.h>
+#include <linux/hardirq.h>
+
+#include <asm/i387.h>
+
+
+/*
+ * MMX 3DNow! library helper functions
+ *
+ * To do:
+ * We can use MMX just for prefetch in IRQ's. This may be a win.
+ * (reported so on K6-III)
+ * We should use a better code neutral filler for the short jump
+ * leal ebx. [ebx] is apparently best for K6-2, but Cyrix ??
+ * We also want to clobber the filler register so we don't get any
+ * register forwarding stalls on the filler.
+ *
+ * Add *user handling. Checksums are not a win with MMX on any CPU
+ * tested so far for any MMX solution figured.
+ *
+ * 22/09/2000 - Arjan van de Ven
+ * Improved for non-egineering-sample Athlons
+ *
+ */
+
+void *_mmx_memcpy(void *to, const void *from, size_t len)
+{
+ void *p;
+ int i;
+
+ if (unlikely(in_interrupt()))
+ return __memcpy(to, from, len);
+
+ p = to;
+ i = len >> 6; /* len/64 */
+
+ kernel_fpu_begin();
+
+ __asm__ __volatile__ (
+ "1: prefetch (%0)\n" /* This set is 28 bytes */
+ " prefetch 64(%0)\n"
+ " prefetch 128(%0)\n"
+ " prefetch 192(%0)\n"
+ " prefetch 256(%0)\n"
+ "2: \n"
+ ".section .fixup, \"ax\"\n"
+ "3: movw $0x1AEB, 1b\n" /* jmp on 26 bytes */
+ " jmp 2b\n"
+ ".previous\n"
+ ".section __ex_table,\"a\"\n"
+ " .align 4\n"
+ " .long 1b, 3b\n"
+ ".previous"
+ : : "r" (from) );
+
+
+ for(; i>5; i--)
+ {
+ __asm__ __volatile__ (
+ "1: prefetch 320(%0)\n"
+ "2: movq (%0), %%mm0\n"
+ " movq 8(%0), %%mm1\n"
+ " movq 16(%0), %%mm2\n"
+ " movq 24(%0), %%mm3\n"
+ " movq %%mm0, (%1)\n"
+ " movq %%mm1, 8(%1)\n"
+ " movq %%mm2, 16(%1)\n"
+ " movq %%mm3, 24(%1)\n"
+ " movq 32(%0), %%mm0\n"
+ " movq 40(%0), %%mm1\n"
+ " movq 48(%0), %%mm2\n"
+ " movq 56(%0), %%mm3\n"
+ " movq %%mm0, 32(%1)\n"
+ " movq %%mm1, 40(%1)\n"
+ " movq %%mm2, 48(%1)\n"
+ " movq %%mm3, 56(%1)\n"
+ ".section .fixup, \"ax\"\n"
+ "3: movw $0x05EB, 1b\n" /* jmp on 5 bytes */
+ " jmp 2b\n"
+ ".previous\n"
+ ".section __ex_table,\"a\"\n"
+ " .align 4\n"
+ " .long 1b, 3b\n"
+ ".previous"
+ : : "r" (from), "r" (to) : "memory");
+ from+=64;
+ to+=64;
+ }
+
+ for(; i>0; i--)
+ {
+ __asm__ __volatile__ (
+ " movq (%0), %%mm0\n"
+ " movq 8(%0), %%mm1\n"
+ " movq 16(%0), %%mm2\n"
+ " movq 24(%0), %%mm3\n"
+ " movq %%mm0, (%1)\n"
+ " movq %%mm1, 8(%1)\n"
+ " movq %%mm2, 16(%1)\n"
+ " movq %%mm3, 24(%1)\n"
+ " movq 32(%0), %%mm0\n"
+ " movq 40(%0), %%mm1\n"
+ " movq 48(%0), %%mm2\n"
+ " movq 56(%0), %%mm3\n"
+ " movq %%mm0, 32(%1)\n"
+ " movq %%mm1, 40(%1)\n"
+ " movq %%mm2, 48(%1)\n"
+ " movq %%mm3, 56(%1)\n"
+ : : "r" (from), "r" (to) : "memory");
+ from+=64;
+ to+=64;
+ }
+ /*
+ * Now do the tail of the block
+ */
+ __memcpy(to, from, len&63);
+ kernel_fpu_end();
+ return p;
+}
+
+#ifdef CONFIG_MK7
+
+/*
+ * The K7 has streaming cache bypass load/store. The Cyrix III, K6 and
+ * other MMX using processors do not.
+ */
+
+static void fast_clear_page(void *page)
+{
+ int i;
+
+ kernel_fpu_begin();
+
+ __asm__ __volatile__ (
+ " pxor %%mm0, %%mm0\n" : :
+ );
+
+ for(i=0;i<4096/64;i++)
+ {
+ __asm__ __volatile__ (
+ " movntq %%mm0, (%0)\n"
+ " movntq %%mm0, 8(%0)\n"
+ " movntq %%mm0, 16(%0)\n"
+ " movntq %%mm0, 24(%0)\n"
+ " movntq %%mm0, 32(%0)\n"
+ " movntq %%mm0, 40(%0)\n"
+ " movntq %%mm0, 48(%0)\n"
+ " movntq %%mm0, 56(%0)\n"
+ : : "r" (page) : "memory");
+ page+=64;
+ }
+ /* since movntq is weakly-ordered, a "sfence" is needed to become
+ * ordered again.
+ */
+ __asm__ __volatile__ (
+ " sfence \n" : :
+ );
+ kernel_fpu_end();
+}
+
+static void fast_copy_page(void *to, void *from)
+{
+ int i;
+
+ kernel_fpu_begin();
+
+ /* maybe the prefetch stuff can go before the expensive fnsave...
+ * but that is for later. -AV
+ */
+ __asm__ __volatile__ (
+ "1: prefetch (%0)\n"
+ " prefetch 64(%0)\n"
+ " prefetch 128(%0)\n"
+ " prefetch 192(%0)\n"
+ " prefetch 256(%0)\n"
+ "2: \n"
+ ".section .fixup, \"ax\"\n"
+ "3: movw $0x1AEB, 1b\n" /* jmp on 26 bytes */
+ " jmp 2b\n"
+ ".previous\n"
+ ".section __ex_table,\"a\"\n"
+ " .align 4\n"
+ " .long 1b, 3b\n"
+ ".previous"
+ : : "r" (from) );
+
+ for(i=0; i<(4096-320)/64; i++)
+ {
+ __asm__ __volatile__ (
+ "1: prefetch 320(%0)\n"
+ "2: movq (%0), %%mm0\n"
+ " movntq %%mm0, (%1)\n"
+ " movq 8(%0), %%mm1\n"
+ " movntq %%mm1, 8(%1)\n"
+ " movq 16(%0), %%mm2\n"
+ " movntq %%mm2, 16(%1)\n"
+ " movq 24(%0), %%mm3\n"
+ " movntq %%mm3, 24(%1)\n"
+ " movq 32(%0), %%mm4\n"
+ " movntq %%mm4, 32(%1)\n"
+ " movq 40(%0), %%mm5\n"
+ " movntq %%mm5, 40(%1)\n"
+ " movq 48(%0), %%mm6\n"
+ " movntq %%mm6, 48(%1)\n"
+ " movq 56(%0), %%mm7\n"
+ " movntq %%mm7, 56(%1)\n"
+ ".section .fixup, \"ax\"\n"
+ "3: movw $0x05EB, 1b\n" /* jmp on 5 bytes */
+ " jmp 2b\n"
+ ".previous\n"
+ ".section __ex_table,\"a\"\n"
+ " .align 4\n"
+ " .long 1b, 3b\n"
+ ".previous"
+ : : "r" (from), "r" (to) : "memory");
+ from+=64;
+ to+=64;
+ }
+ for(i=(4096-320)/64; i<4096/64; i++)
+ {
+ __asm__ __volatile__ (
+ "2: movq (%0), %%mm0\n"
+ " movntq %%mm0, (%1)\n"
+ " movq 8(%0), %%mm1\n"
+ " movntq %%mm1, 8(%1)\n"
+ " movq 16(%0), %%mm2\n"
+ " movntq %%mm2, 16(%1)\n"
+ " movq 24(%0), %%mm3\n"
+ " movntq %%mm3, 24(%1)\n"
+ " movq 32(%0), %%mm4\n"
+ " movntq %%mm4, 32(%1)\n"
+ " movq 40(%0), %%mm5\n"
+ " movntq %%mm5, 40(%1)\n"
+ " movq 48(%0), %%mm6\n"
+ " movntq %%mm6, 48(%1)\n"
+ " movq 56(%0), %%mm7\n"
+ " movntq %%mm7, 56(%1)\n"
+ : : "r" (from), "r" (to) : "memory");
+ from+=64;
+ to+=64;
+ }
+ /* since movntq is weakly-ordered, a "sfence" is needed to become
+ * ordered again.
+ */
+ __asm__ __volatile__ (
+ " sfence \n" : :
+ );
+ kernel_fpu_end();
+}
+
+#else
+
+/*
+ * Generic MMX implementation without K7 specific streaming
+ */
+
+static void fast_clear_page(void *page)
+{
+ int i;
+
+ kernel_fpu_begin();
+
+ __asm__ __volatile__ (
+ " pxor %%mm0, %%mm0\n" : :
+ );
+
+ for(i=0;i<4096/128;i++)
+ {
+ __asm__ __volatile__ (
+ " movq %%mm0, (%0)\n"
+ " movq %%mm0, 8(%0)\n"
+ " movq %%mm0, 16(%0)\n"
+ " movq %%mm0, 24(%0)\n"
+ " movq %%mm0, 32(%0)\n"
+ " movq %%mm0, 40(%0)\n"
+ " movq %%mm0, 48(%0)\n"
+ " movq %%mm0, 56(%0)\n"
+ " movq %%mm0, 64(%0)\n"
+ " movq %%mm0, 72(%0)\n"
+ " movq %%mm0, 80(%0)\n"
+ " movq %%mm0, 88(%0)\n"
+ " movq %%mm0, 96(%0)\n"
+ " movq %%mm0, 104(%0)\n"
+ " movq %%mm0, 112(%0)\n"
+ " movq %%mm0, 120(%0)\n"
+ : : "r" (page) : "memory");
+ page+=128;
+ }
+
+ kernel_fpu_end();
+}
+
+static void fast_copy_page(void *to, void *from)
+{
+ int i;
+
+
+ kernel_fpu_begin();
+
+ __asm__ __volatile__ (
+ "1: prefetch (%0)\n"
+ " prefetch 64(%0)\n"
+ " prefetch 128(%0)\n"
+ " prefetch 192(%0)\n"
+ " prefetch 256(%0)\n"
+ "2: \n"
+ ".section .fixup, \"ax\"\n"
+ "3: movw $0x1AEB, 1b\n" /* jmp on 26 bytes */
+ " jmp 2b\n"
+ ".previous\n"
+ ".section __ex_table,\"a\"\n"
+ " .align 4\n"
+ " .long 1b, 3b\n"
+ ".previous"
+ : : "r" (from) );
+
+ for(i=0; i<4096/64; i++)
+ {
+ __asm__ __volatile__ (
+ "1: prefetch 320(%0)\n"
+ "2: movq (%0), %%mm0\n"
+ " movq 8(%0), %%mm1\n"
+ " movq 16(%0), %%mm2\n"
+ " movq 24(%0), %%mm3\n"
+ " movq %%mm0, (%1)\n"
+ " movq %%mm1, 8(%1)\n"
+ " movq %%mm2, 16(%1)\n"
+ " movq %%mm3, 24(%1)\n"
+ " movq 32(%0), %%mm0\n"
+ " movq 40(%0), %%mm1\n"
+ " movq 48(%0), %%mm2\n"
+ " movq 56(%0), %%mm3\n"
+ " movq %%mm0, 32(%1)\n"
+ " movq %%mm1, 40(%1)\n"
+ " movq %%mm2, 48(%1)\n"
+ " movq %%mm3, 56(%1)\n"
+ ".section .fixup, \"ax\"\n"
+ "3: movw $0x05EB, 1b\n" /* jmp on 5 bytes */
+ " jmp 2b\n"
+ ".previous\n"
+ ".section __ex_table,\"a\"\n"
+ " .align 4\n"
+ " .long 1b, 3b\n"
+ ".previous"
+ : : "r" (from), "r" (to) : "memory");
+ from+=64;
+ to+=64;
+ }
+ kernel_fpu_end();
+}
+
+
+#endif
+
+/*
+ * Favour MMX for page clear and copy.
+ */
+
+static void slow_zero_page(void * page)
+{
+ int d0, d1;
+ __asm__ __volatile__( \
+ "cld\n\t" \
+ "rep ; stosl" \
+ : "=&c" (d0), "=&D" (d1)
+ :"a" (0),"1" (page),"0" (1024)
+ :"memory");
+}
+
+void mmx_clear_page(void * page)
+{
+ if(unlikely(in_interrupt()))
+ slow_zero_page(page);
+ else
+ fast_clear_page(page);
+}
+
+static void slow_copy_page(void *to, void *from)
+{
+ int d0, d1, d2;
+ __asm__ __volatile__( \
+ "cld\n\t" \
+ "rep ; movsl" \
+ : "=&c" (d0), "=&D" (d1), "=&S" (d2) \
+ : "0" (1024),"1" ((long) to),"2" ((long) from) \
+ : "memory");
+}
+
+
+void mmx_copy_page(void *to, void *from)
+{
+ if(unlikely(in_interrupt()))
+ slow_copy_page(to, from);
+ else
+ fast_copy_page(to, from);
+}