Fix and improve D scripts for new message formats

[bnaecker]

No description provided.

[bnaecker]

Fixes #170

[bnaecker]

So during this, @ahl and I discovered what appears to be DTrace bug. I initially tried to write a few if-else statements to stringify the SpResponse variant ID into a name. Something like this:

xcvr-ctl$target:::packet-received
{
    this->peer = json(copyinstr(arg0), "ok");
    this->n_bytes = arg1;

    /* Pointer to the packet payload itself. */
    this->buf = (char*) copyin(arg2, this->n_bytes);
    this->vers = this->buf[0];
    this->message_id = *(uint64_t*) (this->buf + HEADER_MESSAGE_ID_OFFSET);
    this->message_kind = this->buf[HEADER_MESSAGE_KIND_OFFSET];
    printf("Recv payload from: %s\n", this->peer);
    printf("  Raw packet:\n");
    tracemem(this->buf, 128, this->n_bytes);
    printf("  n_bytes: %d\n", this->n_bytes);
    printf("  version: %d\n", this->vers);
    printf("  msg id: %d\n", this->message_id);

    this->message_kind_str = "Unknown";
    if (this->message_kind == MESSAGE_KIND_ERROR) {
        this->message_kind_str = "Error";
    } else if (this->message_kind == MESSAGE_KIND_HOST_REQUEST) {
        this->message_kind_str = "HostRequest";
    } else if (this->message_kind == MESSAGE_KIND_SP_RESPONSE) {
        this->message_kind_str = "SpResponse";
    }
    printf("MessageBody: %s (%d)\n", this->message_kind_str, this->message_kind);
    tracemem(this->buf, 128);
}

However, I noticed that the tracemem output at the beginning of the action and end were different, despite nothing touching the printed variable this->buf. Here is the output of running on one of the tests in this repo:

bnaecker@shale : ~/transceiver-control $ pfexec ./dtrace/trace-packets.d -c './target/debug/deps/transceiver_controller-1fd5c472f0e52c63 --test-threads 1 -- --nocapture test_enable_power'

running 1 test
test controller::tests::test_enable_power ... ok

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 30 filtered out; finished in 0.00s

Sent payload to ::1
  n_bytes: 21
  version: 1
  msg id: 0
  msg kind: HostRequest (1)
Recv payload from: ::1
  Raw packet:

             0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f  0123456789abcdef
         0: 01 00 00 00 00 00 00 00 00 02 03 02 03 03 00 00  ................
        10: 00 00 00 00 00 00 00 00 00 00 00 00 00           .............
  n_bytes: 29
  version: 1
  msg id: 0
MessageBody: SpResponse (2)

             0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f  0123456789abcdef
         0: 02 00 00 00 00 00 00 00 02 00 00 00 00 00 00 00  ................
        10: 02 00 00 00 00 00 00 00 02 00 00 00 00 00 00 00  ................
        20: 02 00 00 00 00 00 00 00 02 00 00 00 00 00 00 00  ................
        30: 02 00 00 00 00 00 00 00 02 00 00 00 00 00 00 00  ................
        40: 02 00 00 00 00 00 00 00 02 00 00 00 00 00 00 00  ................
        50: 02 00 00 00 00 00 00 00 02 00 00 00 00 00 00 00  ................
        60: 02 00 00 00 00 00 00 00 02 00 00 00 00 00 00 00  ................
        70: 02 00 00 00 00 00 00 00 02 00 00 00 00 00 00 00  ................
Sent payload to ::1
  n_bytes: 21
  version: 1
  msg id: 1
  msg kind: HostRequest (1)
Recv payload from: ::1
  Raw packet:

             0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f  0123456789abcdef
         0: 01 01 00 00 00 00 00 00 00 02 03 02 03 02 00 00  ................
        10: 00 00 00 00 00 01 00 00 00 00 00 00 00 02        ..............
  n_bytes: 30
  version: 1
  msg id: 1
MessageBody: SpResponse (2)

             0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f  0123456789abcdef
         0: 02 01 00 00 00 00 00 00 02 01 00 00 00 00 00 00  ................
        10: 02 01 00 00 00 00 00 00 02 01 00 00 00 00 00 00  ................
        20: 02 01 00 00 00 00 00 00 02 01 00 00 00 00 00 00  ................
        30: 02 01 00 00 00 00 00 00 02 01 00 00 00 00 00 00  ................
        40: 02 01 00 00 00 00 00 00 02 01 00 00 00 00 00 00  ................
        50: 02 01 00 00 00 00 00 00 02 01 00 00 00 00 00 00  ................
        60: 02 01 00 00 00 00 00 00 02 01 00 00 00 00 00 00  ................
        70: 02 01 00 00 00 00 00 00 02 01 00 00 00 00 00 00  ................

At the end there, you can see the first row of octets are different in those two outputs. (The latter prints more than it should, sorry.)

After I manually desugared the if-else stuff into a bunch of separate actions (the script in this PR), everything worked fine, and this->buf was the same in both printouts. Here is Adam's attempt to show the desugaring DTrace does to the broken action:

xcvr-ctl26649:::packet-received
{
        ((self->%error) = 0x0);
}

xcvr-ctl26649:::packet-received
{
        ((this->peer) = json(copyinstr(arg0), "ok"));
        ((this->n_bytes) = arg1);
        ((this->buf) = (/* bad node 160c220, kind 6 */
)copyin(arg2, (this->n_bytes)));
        ((this->vers) = ((this->buf)[0x0]));
        ((this->message_id) = *((/* bad node 142ebd0, kind 6 */
)((this->buf) + 0x1)));
        ((this->message_kind) = ((this->buf)[0x9]));
        printf("Recv payload from: %s\n", (this->peer));
        printf("  Raw packet:\n");
        tracemem((this->buf), 0x80, (this->n_bytes));
        printf("  n_bytes: %d\n", (this->n_bytes));
        printf("  version: %d\n", (this->vers));
        printf("  msg id: %d\n", (this->message_id));
        ((this->message_kind_str) = "Unknown");
}

xcvr-ctl26649:::packet-received
/!((self->%error))/
{
        ((this->%condition_1) = (0x1 && ((this->message_kind) == 0x4d2)));
}

xcvr-ctl26649:::packet-received
/(!((self->%error)) && (this->%condition_1))/
{
        ((this->message_kind_str) = "Error");
}

xcvr-ctl26649:::packet-received
/!((self->%error))/
{
        ((this->%condition_2) = (0x1 && !((this->%condition_1))));
}

xcvr-ctl26649:::packet-received
/!((self->%error))/
{
        ((this->%condition_3) = ((this->%condition_2) && ((this->message_kind) == 0x162e)));
}

xcvr-ctl26649:::packet-received
/(!((self->%error)) && (this->%condition_3))/
{
        ((this->message_kind_str) = "HostRequest");
}

xcvr-ctl26649:::packet-received
/!((self->%error))/
{
        ((this->%condition_4) = ((this->%condition_2) && !((this->%condition_3))));
}

xcvr-ctl26649:::packet-received
/!((self->%error))/
{
        ((this->%condition_5) = ((this->%condition_4) && ((this->message_kind) == 0x2)));
}

xcvr-ctl26649:::packet-received
/(!((self->%error)) && (this->%condition_5))/
{
        ((this->message_kind_str) = "SpResponse");
}

xcvr-ctl26649:::packet-received
/!((self->%error))/
{
        printf("MessageBody: %s (%d)\n", (this->message_kind_str), (this->message_kind));
        tracemem((this->buf), 0x80);
}

That looks pretty close to what I wrote by hand, so nothing is obviously broken. We noted that every line of the wrong tracemem output is identical, and Adam is curious if something is stomping on scratch space.