This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: sendmail function probe registration error with systemtap
- From: fche at redhat dot com (Frank Ch. Eigler)
- To: Xiaokui Shu <subbyte at gmail dot com>
- Cc: systemtap at sourceware dot org
- Date: Fri, 10 Jan 2014 13:46:46 -0500
- Subject: Re: sendmail function probe registration error with systemtap
- Authentication-results: sourceware.org; auth=none
- References: <CAMU3N_JxL8CFUOmiyuh0FMaVt_q3tF3b1ATSP+nW6HiS=dkV+g at mail dot gmail dot com>
Xiaokui Shu <subbyte@gmail.com> writes:
> [...] I start systemtap with a simple probe
> (shown below), I get two warnings in Pass 5. While running systemtap,
> [...]
> WARNING: probe process("/usr/sbin/sendmail.sendmail").function("mime_getchar_crlf.part.1")
> inode-offset 0005e620 registration error (rc -524)
> WARNING: probe process("/usr/sbin/sendmail.sendmail").function("frame_dummy")
> inode-offset 0001cc80 registration error (rc -524)
That's -ENOTSUPP; maybe there is some sort of instruction in there the
kernel doesn't like.
> I can send emails to the monitored host, but my mailbox on the host
> does not receive them. [...] Can anyone help me to figure out what
> is the problem brought by the tracing?
This sounds should not happen.
> probe process(@1).function("*")
> {
> printf ("%s:0x%x\n", probefunc (), ustack (1))
> }
You might try rerunning '# stap -DSTP_ALIBI foo.stp' to eliminate most
of the systemtap probe handler logic out of the script. If the
userspace sendmail process still misbehaves, this helps incriminate
uprobes -- and vice versa.
Could you collect kernel / stap version information, or the whole
`stap-report` shebang?
- FChE