Make userland / assembly getting started more uniform / visible

Forward --gcc-which to ./run --tmux.

Use gdb-multiarch for --gcc-which host.

userland/arch/arm/vadd_scalar.S

@@ -3,16 +3,6 @@
 
 #include "common.h"
 
-.data;
-a1:
-    .float 0.0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5
-a2:
-    .float 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5
-sum:
-    .skip 32
-sum_expect:
-    .float 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0
-
 ENTRY
     /* Minimal single precision floating point example.
      * TODO: floating point representation constraints due to 4-byte instruction?
@@ -79,74 +69,4 @@ my_float_sum:
     vmov s1, s0
     vmov r1, s1
     ASSERT_EQ_REG(r0, r1)
-
-    /* Now a more complex test function. */
-    ldr r0, =sum
-    ldr r1, =a1
-    ldr r2, =a2
-    mov r3, 8
-    bl vec_sum
-    /* The assert works easily because all floats used
-     * have exact base-2 representation.
-     */
-    ASSERT_MEMCMP(sum, sum_expect, 0x20)
 EXIT
-
-/* void vec_sum(float *sum, float *a1, float *a2, int length) {
- *   int i;
- *   for (i=0; i < length; i++)
- *     *(sum+i) = *(a1+i) + *(a2+i);
- * }
- */
-vec_sum:
-    /* Setup */
-    push {r0, r1, r4, lr}
-    push {r0, r1}
-    mov r0, 1
-    mov r1, 8
-    bl reconfig
-    pop {r0, r1}
-    asr r3, 3
-
-    /* Do the sum. */
-1:
-    fldmias r1!, {s8-s15}
-    fldmias r2!, {s16-s23}
-    vadd.f32 s24, s8, s16
-    fstmias r0!, {s24-s31}
-    subs r3, r3, 1
-    bne 1b
-
-    /* Teardown. */
-    bl deconfig
-    pop {r0, r1, r4, pc}
-
-/* inputs:
- * r0: desired vector stride (1 or 2)
- * r1: desired vector length (min. 1, max. 8)
- * outputs: (none)
- * modified: r0, r1, FPSCR
- * notes:
- * r0 and r1 will be truncated before fitting into FPSCR
- */
-reconfig:
-    push {r0-r2}
-    and r0, r0, 3
-    eor r0, r0, 1
-    sub r1, r1, 1
-    and r1, r1, 7
-    mov r0, r0, lsl 20
-    orr r0, r0, r1, lsl 16
-    vmrs r2, fpscr
-    bic r2, 55*65536
-    orr r2, r2, r0
-    vmsr fpscr, r0
-    pop {r0-r2}
-    bx lr
-
-deconfig:
-    push {r0, r1, lr}
-    mov r0, 1
-    mov r1, 1
-    bl reconfig
-    pop {r0, r1, pc}