vm/vmimpl: show BBLog entries for all locked tcpcbs

[tuexen]

---

Before sending a pull request, please review Contribution Guidelines:
https://github.com/google/syzkaller/blob/master/docs/contributing.md

---

[markjdb]

Why move it? Is the console report getting truncated? Maybe it makes sense to raise the limit, or stop showing uma and malloc, since most of that output is not useful.

[tuexen]

Why move it? Is the console report getting truncated? Maybe it makes sense to raise the limit, or stop showing uma and malloc, since most of that output is not useful.

I looked around and found this report. It seems that the output is truncated. I assume that if a tcpcb is locked, the problem is TCP related. In this case the output of show all tcpcbs/bl might show the information needed, but I don't think the other output is of any help. At least for me. That is why I moved it almost to the top. If the problem is not TCP related, I do not expect any tcpcb to be locked and therefore the output of show all tcpcbs/bl is only one line.
Since the tail of the output of show all tcpcbs/bl is more important than the head, I limit the number of events to 1000 in #6059 hoping that we can see 1000 lines of output. We will see, when the bug is triggered after D50629 is committed to the FreeBSD tree and #6059 and #6060 are merged to syzkaller.

[tuexen]

OK, let us just add the logging of the BBLog entries in this commit. I will remove

show uma
show malloc

in #6095. Then we can see if this is good enough in combination with #6059.
If that does not work out, I will try to find out where the truncation happens. But my focus is on fixing the TCP bug syzkaller (and right now only syzkaller) is triggering.