bpf: make test py_test_tools_smoke pass on arm64

Changes include:
  . Add PT_REGS_FP to access base(FP) register in x64
  . Use macros, intead of directly ctx-><reg_name>
    in a few places
  . Let userspace fill in the value of PAGE_SIZE.
    Otherwise, arm64 needs additional headers to
    get this value for kernel.
  . In tools/wakeuptime.py, arm64 and x86_64 have
    the same stack walker mechanism. But they
    have different symbol/macro to represent
    kernel start address.
With these changes, the test py_test_tools_smoke
can pass on arm64.
Signed-off-by: Yonghong Song <yhs@fb.com>

src/cc/export/helpers.h

@@ -566,6 +566,7 @@ int bpf_usdt_readarg_p(int argc, struct pt_regs *ctx, void *buf, u64 len) asm("l
 #define PT_REGS_PARM4(ctx)	((ctx)->cx)
 #define PT_REGS_PARM5(ctx)	((ctx)->r8)
 #define PT_REGS_PARM6(ctx)	((ctx)->r9)
+#define PT_REGS_FP(ctx)         ((ctx)->bp) /* Works only with CONFIG_FRAME_POINTER */
 #define PT_REGS_RC(ctx)		((ctx)->ax)
 #define PT_REGS_IP(ctx)		((ctx)->ip)
 #define PT_REGS_SP(ctx)		((ctx)->sp)

tools/lib/ucalls.py

@@ -185,7 +185,7 @@ int trace_return(struct pt_regs *ctx) {
 #ifdef SYSCALLS
 int syscall_entry(struct pt_regs *ctx) {
     u64 pid = bpf_get_current_pid_tgid();
-    u64 *valp, ip = ctx->ip, val = 0;
+    u64 *valp, ip = PT_REGS_IP(ctx), val = 0;
     PID_FILTER
 #ifdef LATENCY
     struct syscall_entry_t data = {};

tools/memleak.py

@@ -14,6 +14,7 @@
 from bcc import BPF
 from time import sleep
 from datetime import datetime
+import resource
 import argparse
 import subprocess
 import os
@@ -366,6 +367,7 @@ if kernel_trace:
 
 bpf_source = bpf_source.replace("SHOULD_PRINT", "1" if trace_all else "0")
 bpf_source = bpf_source.replace("SAMPLE_EVERY_N", str(sample_every_n))
+bpf_source = bpf_source.replace("PAGE_SIZE", str(resource.getpagesize()))
 
 size_filter = ""
 if min_size is not None and max_size is not None:

tools/wakeuptime.py

@@ -76,13 +76,19 @@ BPF_HASH(start, u32);
 
 static u64 get_frame(u64 *bp) {
     if (*bp) {
-        // The following stack walker is x86_64 specific
+        // The following stack walker is x86_64/arm64 specific
         u64 ret = 0;
         if (bpf_probe_read(&ret, sizeof(ret), (void *)(*bp+8)))
             return 0;
         if (bpf_probe_read(bp, sizeof(*bp), (void *)*bp))
             return 0;
+#ifdef __x86_64__
         if (ret < __START_KERNEL_map)
+#elif __aarch64__
+        if (ret < VA_START)
+#else
+#error "Unsupported architecture for stack walker"
+#endif
             return 0;
         return ret;
     }
@@ -121,7 +127,7 @@ int waker(struct pt_regs *ctx, struct task_struct *p) {
 
     bpf_probe_read(&key.target, sizeof(key.target), p->comm);
     bpf_get_current_comm(&key.waker, sizeof(key.waker));
-    bp = ctx->bp;
+    bp = PT_REGS_FP(ctx);
 
     // unrolled loop (MAXDEPTH):
     if (!(key.ret[depth++] = get_frame(&bp))) goto out;